abstract
stringlengths 383
2.62k
| TLDR
stringlengths 4
250
|
|---|---|
Adversarial training is a promising method to improve the robustness against adversarial attacks. To enhance its performance, recent methods impose high weights on the cross-entropy loss for important data points near the decision boundary. However, these importance-aware methods are vulnerable to sophisticated attacks, e.g., Auto-Attack. In this paper, we experimentally investigate the cause of their vulnerability via margins between logits for the true label and the other labels because they should be large enough to prevent the largest logit from being flipped by the attacks. Our experiments reveal that the histogram of the logit margins of naive adversarial training has two peaks. Thus, the levels of difficulty in increasing logit margins are roughly divided into two: difficult samples (small logit margins) and easy samples (large logit margins). On the other hand, only one peak near zero appears in the histogram of importance-aware methods, i.e., they reduce the logit margins of easy samples. To increase logit margins of difficult samples without reducing those of easy samples, we propose switching one-versus-the-rest loss (SOVR), which switches from cross-entropy to one-versus-the-rest loss (OVR) for difficult samples. We derive trajectories of logit margins for a simple problem and prove that OVR increases logit margins two times larger than the weighted cross-entropy loss. Thus, SOVR increases logit margins of difficult samples, unlike existing methods. We experimentally show that SOVR achieves better robustness against Auto-Attack than importance-aware methods. | We prove that one-versus-rest loss (OVR) increases logit margins two times greater than cross-entropy and propose switching between cross-entropy and OVR by the criterion of logit margins to improve adversarial robustness. |
By conditioning on natural language instructions, large language models (LLMs) have displayed impressive capabilities as general-purpose computers. However, task performance depends significantly on the quality of the prompt used to steer the model, and most effective prompts have been handcrafted by humans. Inspired by classical program synthesis and the human approach to prompt engineering, we propose Automatic Prompt Engineer (APE) for automatic instruction generation and selection. In our method, we treat the instruction as the "program," optimized by searching over a pool of instruction candidates proposed by an LLM in order to maximize a chosen score function. To evaluate the quality of the selected instruction, we evaluate the zero-shot performance of another LLM following the selected instruction. Experiments on 24 NLP tasks show that our automatically generated instructions outperform the prior LLM baseline by a large margin and achieve better or comparable performance to the instructions generated by human annotators on 21/24 tasks. We conduct extensive qualitative and quantitative analyses to explore the performance of APE. We show that APE-engineered prompts can be applied to steer models toward truthfulness and/or informativeness, as well as to improve few-shot learning performance by simply prepending them to standard in-context learning prompts. | We propose an algorithm for automatic instruction generation and selection for large language models with human level performance. |
Meta-learning aims to extract useful inductive biases from a set of related datasets. In Bayesian meta-learning, this is typically achieved by constructing a prior distribution over neural network parameters. However, specifying families of computationally viable prior distributions over the high-dimensional neural network parameters is difficult. As a result, existing approaches resort to meta-learning restrictive diagonal Gaussian priors, severely limiting their expressiveness and performance. To circumvent these issues, we approach meta-learning through the lens of functional Bayesian neural network inference which views the prior as a stochastic process and performs inference in the function space. Specifically, we view the meta-training tasks as samples from the data-generating process and formalize meta-learning as empirically estimating the law of this stochastic process. Our approach can seamlessly acquire and represent complex prior knowledge by meta-learning the score function of the data-generating process marginals instead of parameter space priors. In a comprehensive benchmark, we demonstrate that our method achieves state-of-the-art performance in terms of predictive accuracy and substantial improvements in the quality of uncertainty estimates. | Meta-learning in the function space by estimating the score function of the data-generating process marginals. |
Data-efficiency has always been an essential issue in pixel-based reinforcement learning (RL). As the agent not only learns decision-making but also meaningful representations from images. The line of reinforcement learning with data augmentation shows significant improvements in sample-efficiency. However, it is challenging to guarantee the optimality invariant transformation, that is, the augmented data are readily recognized as a completely different state by the agent. In the end, we propose a contrastive invariant transformation (CoIT), a simple yet promising learnable data augmentation combined with standard model-free algorithms to improve sample-efficiency. Concretely, the differentiable CoIT leverages original samples with augmented samples and hastens the state encoder for a contrastive invariant embedding. We evaluate our approach on DeepMind Control Suite and Atari100K. Empirical results verify advances using CoIT, enabling it to outperform the new state-of-the-art on various tasks. Source code is available at https://github.com/mooricAnna/CoIT. | CoIT is a learnable image transformation for sample-efficiency improvement. |
Autoregressive model is widely adopted to generate 3D molecules which can fit any protein binding pocket. Current autoregressive model suffers from two major drawbacks. First, it is hard to capture local geometric patterns as only one atom is generated at each step. Second, most of the autoregressive models generate atoms and chemical bonds in two separate processes, which causes a number of problems such as incorrect counts of rings, a bias distribution of bond lengths, and inaccurate 3D molecular structures. To tackle this problem, we designed a model, named FragDiff, to generate 3D molecules fragment-by-fragment for pockets. In each generation step, FragDiff places a molecular fragment around the pocket by using E(3)-equivariant diffusion generative models to simultaneously predict the atom types, atom coordinates and the chemical bonds of the fragment. Extensive experimental results confirm our assumption that unifying the atoms and bonds generations could significantly improve the quality of the sampled 3D molecules in terms of more accurate distributions of 2D subgraphs and 3D substructures. | Using fragment-based autoregressive diffusion model to generate 3D molecules for protein binding pockets |
To reduce the human annotation efforts, the programmatic weak supervision (PWS) paradigm abstracts weak supervision sources as labeling functions (LFs) and involves a label model to aggregate the output of multiple LFs to produce training labels. Most existing label models require a parameter learning step for each dataset. In this work, we present a hyper label model that (once learned) infers the ground-truth labels for each dataset in a single forward pass without dataset-specific parameter learning. The hyper label model approximates an optimal analytical (yet computationally intractable) solution of the ground-truth labels. We train the model on synthetic data generated in the way that ensures the model approximates the analytical optimal solution, and build the model upon Graph Neural Network (GNN) to ensure the model prediction being invariant (or equivariant) to the permutation of LFs (or data points). On 14 real-world datasets, our hyper label model outperforms the best existing methods in both accuracy (by 1.4 points on average) and efficiency (by six times on average). Our code is available at https://github.com/wurenzhi/hyper_label_model | A hyper label model to aggregate weak labels from multiple weak supervision sources to infer the ground-truth labels in a single forward pass |
Retrieval-augmented generation models offer many benefits over standalone language models: besides a textual answer to a given query they provide provenance items retrieved from an updateable knowledge base. However, they are also more complex systems and need to handle long inputs. In this work, we introduce FiD-Light to strongly increase the efficiency of the state-of-the-art retrieval-augmented FiD model, while maintaining the same level of effectiveness. Our FiD-Light model constrains the information flow from the encoder (which encodes passages separately) to the decoder (using concatenated encoded representations). Furthermore, we adapt FiD-Light with re-ranking capabilities through textual source pointers, to improve the top-ranked provenance precision. Our experiments on a diverse set of seven knowledge intensive tasks (KILT) show FiD-Light consistently improves the Pareto frontier between query latency and effectiveness. FiD-Light with source pointing sets substantial new state-of-the-art results on six KILT tasks for combined text generation and provenance retrieval evaluation, while maintaining reasonable efficiency. | We increase the efficiency of FID with FiD-Light including a source pointing workflow for effective retrieval augmented generation. |
Many metric learning tasks, such as triplet learning, nearest neighbor retrieval, and visualization, are treated primarily as embedding tasks where the ultimate metric is some variant of the Euclidean distance (e.g., cosine or Mahalanobis), and the algorithm must learn to embed points into the pre-chosen space. The study of non-Euclidean geometries is often not explored, which we believe is due to a lack of tools for learning non-Euclidean measures of distance. Recent work has shown that Bregman divergences can be learned from data, opening a promising approach to learning asymmetric distances. We propose a new approach to learning arbitrary Bergman divergences in a differentiable manner via input convex neural networks and show that it overcomes significant limitations of previous works. We also demonstrate that our method more faithfully learns divergences over a set of both new and previously studied tasks, including asymmetric regression, ranking, and clustering. Our tests further extend to known asymmetric, but non-Bregman tasks, where our method still performs competitively despite misspecification, showing the general utility of our approach for asymmetric learning. | We develop the first effective deep learning tooling for learning arbitrary Bregman divergences |
The discovery of novel therapeutics to cure genetic pathologies relies on the identification of the different genes involved in the underlying disease mechanism. With billions of potential hypotheses to test, an exhaustive exploration of the entire space of potential interventions is impossible in practice. Sample-efficient methods based on active learning or bayesian optimization bear the promise of identifying interesting targets using the least experiments possible. However, genomic perturbation experiments typically rely on proxy outcomes measured in biological model systems that may not completely correlate with the outcome of interventions in humans. In practical experiment design, one aims to find a set of interventions which maximally move a target phenotype via a diverse set of mechanisms in order to reduce the risk of failure in future stages of trials. To that end, we introduce DiscoBAX — a sample-efficient algorithm for the discovery of genetic interventions that maximize the movement of a phenotype in a direction of interest while covering a diverse set of underlying mechanisms. We provide theoretical guarantees on the optimality of the approach under standard assumptions, conduct extensive experiments in synthetic and real-world settings relevant to genomic discovery and demonstrate that DiscoBAX outperforms state-of-the-art active learning and Bayesian optimization methods in this task. Better methods for selecting effective and diverse perturbations in biological systems could enable researchers to potentially discover novel therapeutics for a range of genetically-driven diseases. | We introduce DiscoBAX — a sample-efficient algorithm for the discovery of genetic interventions that maximize the movement of a phenotype in a direction of interest while covering a diverse set of underlying mechanisms |
Mobile UI understanding is important for enabling various interaction tasks such as UI automation and accessibility. Previous mobile UI modeling often depends on the view hierarchy information of a screen, which directly provides the structural data of the UI, with the hope to bypass challenging tasks of visual modeling from screen pixels. However, view hierarchies are not always available, and are often corrupted with missing object descriptions or misaligned structure information. As a result, despite the use of view hierarchies could offer short-term gains, it may ultimately hinder the applicability and performance of the model. In this paper, we propose Spotlight, a vision-only approach for mobile UI understanding. Specifically, we enhance a vision-language model that only takes the screenshot of the UI and a region of interest on the screen---the focus---as the input. This general architecture of Spotlight is easily scalable and capable of performing a range of UI modeling tasks. Our experiments show that our model establishes SoTA results on several representative UI tasks and outperforms previous methods that use both screenshots and view hierarchies as inputs. Furthermore, we explore multi-task learning and few-shot prompting capacities of the proposed models, demonstrating promising results in the multi-task learning direction. | We propose an enhanced vision-language model for UI tasks that achieves SoTA on representative UI tasks and supports few-shot and multitask learning. |
Building systems that achieve a deeper understanding of language is one of the central goals of natural language processing (NLP). Towards this goal, recent works have begun to train language models on narrative datasets which require extracting the most critical information by integrating across long contexts. However, it is still an open question whether these models are learning a deeper understanding of the text, or if the models are simply learning a heuristic to complete the task. This work investigates this further by turning to the one language processing system that truly understands complex language: the human brain. We show that training language models for deeper narrative understanding results in richer representations that have improved alignment to human brain activity. We further find that the improvements in brain alignment are larger for character names than for other discourse features, which indicates that these models are learning important narrative elements. Taken together, these results suggest that this type of training can indeed lead to deeper language understanding. These findings have consequences both for cognitive neuroscience by revealing some of the significant factors behind brain-NLP alignment, and for NLP by highlighting that understanding of long-range context can be improved beyond language modeling. | We show that training language models for deeper narrative understanding (characters, emotions, relationships) results in richer representations that have improved alignment to human brain activity. |
We propose the Gradient-weighted Object Detector Activation Mapping (Grad-ODAM), a visualized explanation technique for interpreting the predictions of object detectors. Utilizing the gradients of detector targets flowing into the intermediate feature maps, Grad-ODAM produces heat maps that show the influence of regions on the detector's decision. Compared to previous classification activation mapping works, Grad-ODAM generates instance-specific explanations rather than class-specific ones. We show that Grad-ODAM is applicable to both one-stage detectors such as FCOS and two-stage detectors such as Faster R-CNN, and produces higher-quality visual explanations than the state-of-the-art both effectively and efficiently. We next propose a training scheme, ODAM-Train, to improve the explanation ability on object discrimination of the detector through encouraging consistency between explanations for detections on the same object, and distinct explanations for detections on different objects. Based on the heat maps produced by Grad-ODAM with ODAM-Train, we propose ODAM-NMS, which considers the information of the model's explanation for each prediction to distinguish the duplicate detected objects. We present a detailed analysis of the visualized explanations of detectors and carry out extensive experiments to validate the effectiveness of the proposed ODAM.
| ODAM: a gradient-based instance-specific explanation technique for object detectors; ODAM-Train: improve the explanation ability on object discrimination; ODAM-NMS: distinguish the duplicate detected objects with the help of ODAM. |
Deep Ensembles (DE) are a prominent approach for achieving excellent performance on key metrics such as accuracy, calibration, uncertainty estimation, and out-of-distribution detection. However, hardware limitations of real-world systems constrain to smaller ensembles and lower-capacity networks, significantly deteriorating their performance and properties. We introduce Packed-Ensembles (PE), a strategy to design and train lightweight structured ensembles by carefully modulating the dimension of their encoding space. We leverage grouped convolutions to parallelize the ensemble into a single shared backbone and forward pass to improve training and inference speeds. PE is designed to operate within the memory limits of a standard neural network. Our extensive research indicates that PE accurately preserves the properties of DE, such as diversity, and performs equally well in terms of accuracy, calibration, out-of-distribution detection, and robustness to distribution shift. We make our code available at https://github.com/ENSTA-U2IS/torch-uncertainty. | Packed-Ensembles leverage the width of DNNs and grouped convolutions to train subnetworks in parallel and form an efficient ensemble. |
Text-to-image models offer unprecedented freedom to guide creation through natural language. Yet, it is unclear how such freedom can be exercised to generate images of specific unique concepts, modify their appearance, or compose them in new roles and novel scenes.
In other words, we ask: how can we use language-guided models to turn *our* cat into a painting, or imagine a new product based on *our* favorite toy?
Here we present a simple approach that allows such creative freedom. Using only $3$-$5$ images of a user-provided concept, like an object or a style, we learn to represent it through new ``words" in the embedding space of a frozen text-to-image model.
These ``words" can be composed into natural language sentences, guiding *personalized* creation in an intuitive way.
Notably, we find evidence that a *single* word embedding is sufficient for capturing unique and varied concepts.
We compare our approach to a wide range of baselines, and demonstrate that it can more faithfully portray the concepts across a range of applications and tasks. Our code, data and new words will be available. | We present the task of personalized text-to-image generation, and introduce an inversion-based method that allows us to synthesize novel scenes of user-provided visual concepts, guided by natural language instructions. |
DADAO is a novel decentralized asynchronous stochastic first order algorithm to minimize a sum of $L$-smooth and $\mu$-strongly convex functions distributed over a time-varying connectivity network of size $n$. We model the local gradient updates and gossip communication procedures with separate independent Poisson Point Processes, decoupling the computation and communication steps in addition to making the whole approach completely asynchronous. Our method employs primal gradients and does not use a multi-consensus inner loop nor other ad-hoc mechanisms such as Error Feedback, Gradient Tracking, or a Proximal operator. By relating the inverse of the smallest positive eigenvalue $\chi^*_1$ and the effective resistance $\chi_2^*$ of our graph to a necessary minimal communication rate between nodes of the network, we show that our algorithm requires $\mathcal{O}(n\sqrt{\frac{L}{\mu}}\log \epsilon)$ local gradients and only $\mathcal{O}(n\sqrt{\chi_1^*\chi_2^*}\sqrt{\frac{L}{\mu}}\log \epsilon)$ communications to reach a precision $\epsilon$. If SGD with uniform noise $\sigma^2$ is used, we reach a precision $\epsilon$ with same speed, up to a bias term in $\mathcal{O}(\frac{\sigma^2}{\sqrt{\mu L}})$. This improves upon the bounds obtained with current state-of-the-art approaches, our simulations validating the strength of our relatively unconstrained method. | We introduce a novel decentralized asynchronous accelerated stochastic first order algorithm to minimize a sum of smooth and strongly convex functions over a time-varying connectivity network. |
Endowing deep neural networks with the ability to reason about cause and effect would be an important step to make them more robust and interpretable. In this work we propose a variational framework that allows deep networks to learn latent variables and their causal relationships from unstructured data, with no supervision, or labeled interventions. Starting from an abstract Structural Equation Model (SEM), we show that maximizing its posterior probability yields a similar construction to a Variational Auto-Encoder (VAE), but with a structured prior coupled by non-linear equations. This prior represents an interpretable SEM with learnable parameters (such as a physical model or dependence structure), which can be fitted to data while simultaneously learning the latent variables. Unfortunately, computing KL-divergences with this non-linear prior is intractable. We show how linearizing arbitrary SEMs via back-propagation produces local non-isotropic Gaussian priors, for which the KL-divergences can be computed efficiently and differentiably. We propose two versions, one for IID data (such as images) which detects related causal variables within a sample, and one for non-IID data (such as video) which detects variables that are also related over time. Our proposal is complementary to causal discovery techniques, which assume given variables, and instead discovers both variables and their causal relationships. We experiment with recovering causal models from images, and learning temporal relations based on the Super Mario Bros videogame. | We propose a modification to the VAE that learns variables and causal relationships between them in an unsupervised way. |
Deep reinforcement learning (DRL) has achieved groundbreaking successes in various applications, including robotics. A natural consequence is the adoption of this paradigm for safety-critical tasks, where human safety and expensive hardware can be involved. In this context, it is crucial to optimize the performance of DRL-based agents while providing guarantees about their behavior. This paper presents a novel technique for incorporating domain-expert knowledge into a constrained DRL training loop. Our technique exploits the scenario-based programming paradigm, designed to specify such knowledge in a simple and intuitive way. While our approach can be considered general purpose, we validated our method by performing experiments on a synthetic set of benchmark environments, and the popular robotic mapless navigation problem, in simulation and on the actual platform. Our results demonstrate that using our approach to leverage expert knowledge dramatically improves the safety and performance of the agent. | A novel technique for incorporating domain-expert knowledge to train a constrained DRL agent, based on scenario-based programming paradigm, we validated our method on the popular robotic mapless navigation problem, both physically and in simulation. |
The task of building general agents that perform well over a wide range of tasks has been an important goal in reinforcement learning since its inception. The problem has been subject of research of a large body of work, with performance frequently measured by observing scores over the wide range of environments contained in the Atari 57 benchmark. Agent57 was the first agent to surpass the human benchmark on all 57 games, but this came at the cost of poor data-efficiency, requiring nearly 80 billion frames of experience to achieve. Taking Agent57 as a starting point, we employ a diverse set of strategies to achieve a 200-fold reduction of experience needed to outperform the human baseline, within our novel agent MEME. We investigate a range of instabilities and bottlenecks we encountered while reducing the data regime, and propose effective solutions to build a more robust and efficient agent. We also demonstrate competitive performance with high-performing methods such as Muesli and MuZero. Our contributions aim to achieve faster propagation of learning signals related to rare events, stabilize learning under differing value scales, improve the neural network architecture, and make updates more robust under a rapidly-changing policy. | We propose an RL agent 'MEME' that achieves human-level performance on all 57 Atari games within 390M environment frames, only 1/200 of the experience required by Agent57. |
Boundary conditions (BCs) are important groups of physics-enforced constraints that are necessary for solutions of Partial Differential Equations (PDEs) to satisfy at specific spatial locations. These constraints carry important physical meaning, and guarantee the existence and the uniqueness of the PDE solution. Current neural-network based approaches that aim to solve PDEs rely only on training data to help the model learn BCs implicitly, however, there is no guarantee of BC satisfaction by these models during evaluation. In this work, we propose Boundary enforcing Operator Network (BOON) that enables the BC satisfaction of neural operators by making structural changes to the operator kernel. We provide our refinement procedure, and demonstrate the satisfaction of physics-based BCs such as Dirichlet, Neumann, and periodic by the solutions obtained by BOON. Numerical experiments based on multiple PDEs with a wide variety of applications indicate that the proposed approach ensures satisfaction of BCs, and leads to more accurate solutions over the whole domain. The proposed method exhibits a (2X-20X) improvement in accuracy (0.000084 relative $L^2$ error for Burgers' equation). Code available at: https://github.com/amazon-science/boon. | We propose novel kernel correction mechanisms for neural operators to satisfy physical boundary constraints which are effective in improving the overall performance. |
In this work, we propose to tackle the problem of domain generalization in the context of insufficient samples. Instead of extracting latent feature embeddings based on deterministic models, we propose to learn a domain-invariant representation based on the probabilistic framework by mapping each data point into probabilistic embeddings. Specifically, we first extend empirical maximum mean discrepancy (MMD) to a novel probabilistic MMD that can measure the discrepancy between mixture distributions (i.e., source domains) consisted of a serial of latent distributions rather than latent points. Moreover, instead of imposing the contrastive semantic alignment (CSA) loss based on pairs of latent points, a novel probabilistic CSA loss encourages positive probabilistic embedding pairs to be closer while pulling other negative ones apart. Benefiting from the learned representation captured by probabilistic models, our proposed method can marriage the measurement on the distribution over distributions (i.e., the global perspective alignment) and the distribution-based contrastive semantic alignment (i.e., the local perspective alignment). Extensive experimental results on three challenging medical datasets show the effectiveness of our proposed method in the context of insufficient data compared with state-of-the-art baseline methods. | A novel domain generalization method in the context of insufficient data is proposed in this work |
In recent years, gradient based multi-agent reinforcement learning is growing in success. One contributing factor is the use of shared parameters for learning policy networks. While this approach scales well with the number of agents during execution it lacks this ambiguity for training as the number of produced samples grows linearly with the number of agents. For a very large number of agents, this could lead to an inefficient use of the circumstantial amount of produced samples. Moreover in single-agent reinforcement learning policy search with evolutionary algorithms showed viable success when sampling can be parallelized on a larger scale. The here proposed method does not only consider sampling in concurrent environments but further investigates sampling diverse parameters from the population in co-evolution in joint environments during training. This co-evolutionary policy search has shown to be capable of training a large number of agents. Beyond that, it has been shown to produce competitive results in smaller environments in comparison to gradient descent based methods. This surprising result make evolutionary algorithms a promising candidate for further research in the context of multi-agent reinforcement learning. | Evolutionary Algorithms can be competitively used for policy search in multi-agent reinforcement and can scale to a high number of agents. |
Graph Neural Networks (GNNs) have achieved state-of-the-art results on a variety of graph learning tasks, however, it has been demonstrated that they are vulnerable to adversarial attacks, raising serious security concerns. A lot of studies have been developed to train GNNs in a noisy environment and increase their robustness against adversarial attacks. However, existing methods have not uncovered a principled difficulty: the convoluted mixture distribution between clean and attacked data samples, which leads to sub-optimal model design and limits their frameworks’ robustness. In this work, we first begin by identifying the root cause of mixture distribution, then, for tackling it, we propose a novel method GAME - Graph Adversarial Mixture of Experts to enlarge the model capacity and enrich the representation diversity of adversarial samples, from three perspectives of model, training, and optimization. Specifically, we first propose a plug-and- play GAME layer that can be easily incorporated into any GNNs and enhance their adversarial learning capabilities. Second, we design a decoupling-based graph adversarial training in which the component of the model used to generate adversarial graphs is separated from the component used to update weights. Third, we introduce a graph diversity regularization that enables the model to learn diverse representation and further improves model performance. Extensive experiments demonstrate the effectiveness and advantages of GAME over the state-of-the-art adversarial training methods across various datasets given different attacks. | We identify a fundamental issue in graph adversarial learning and then propose a novel method to enlarge the model capacity and enrich the representation diversity of adversarial samples. |
Communication is important in many multi-agent reinforcement learning (MARL) problems for agents to share information and make good decisions. However, when deploying trained communicative agents in a real-world application where noise and potential attackers exist, the safety of communication-based policies becomes a severe issue that is underexplored. Specifically, if communication messages are manipulated by malicious attackers, agents relying on untrustworthy communication may take unsafe actions that lead to catastrophic consequences. Therefore, it is crucial to ensure that agents will not be misled by corrupted communication, while still benefiting from benign communication. In this work, we consider an environment with $N$ agents, where the attacker may arbitrarily change the communication from any $C<\frac{N-1}{2}$ agents to a victim agent. For this strong threat model, we propose a certifiable defense by constructing a message-ensemble policy that aggregates multiple randomly ablated message sets. Theoretical analysis shows that this message-ensemble policy can utilize benign communication while being certifiably robust to adversarial communication, regardless of the attacking algorithm. Experiments in multiple environments verify that our defense significantly improves the robustness of trained policies against various types of attacks. | We propose a defense method such that an agent receiving communication in an multi-agent system can be certifiably robust when a subset of communication messages get (arbitrarily) perturbed. |
Given a trained model and a data sample, membership-inference (MI) attacks predict whether the sample was in the model’s training set. A common counter- measure against MI attacks is to utilize differential privacy (DP) during model training to mask the presence of individual examples. While this use of DP is a principled approach to limit the efficacy of MI attacks, there is a gap between the bounds provided by DP and the empirical performance of MI attacks. In this paper, we derive bounds for the advantage of an adversary mounting a MI attack, and demonstrate tightness for the widely-used Gaussian mechanism. Our analysis answers an open problem in the field of differential privacy, namely the fact that membership inference is not 100% successful even for relatively high budgets ($\epsilon> 10$). Finally, using our analysis, we provide MI metrics for models trained on CIFAR10 dataset. To the best of our knowledge, our analysis provides the state-of-the-art membership inference bounds. | We prove optimal membership inference bounds for DP-SGD, beating previously known upper bounds on membership inference. |
Noting the importance of factorizing (or disentangling) the latent space, we propose a novel, non-probabilistic disentangling framework for autoencoders, based on the principles of symmetry transformations that are independent of one another. To the best of our knowledge, this is the first deterministic model that is aiming to achieve disentanglement based on autoencoders without pairs of images or labels, by explicitly introducing inductive biases into a model architecture through Euler encoding. The proposed model is then compared with a number of state-of-the-art models, relevant to disentanglement, including symmetry-based and generative models based on autoencoders. Our evaluation using six different disentanglement metrics, including the unsupervised disentanglement metric we propose here in this paper, shows that the proposed model can offer better disentanglement, especially when variances of the features are different, where other methods may struggle. We believe that this model opens several opportunities for linear disentangled representation learning based on deterministic autoencoders. | We propose the first deterministic model that is aiming to achieve disentanglement based on autoencoders without a pair of images or labels by explicitly introducing inductive biases into a model architecture through Euler encoding. |
We study the use of amortized optimization to predict optimal transport (OT) maps from the input measures, which we call Meta OT. This helps repeatedly solve similar OT problems between different measures by leveraging the knowledge and information present from past problems to rapidly predict and solve new problems. Otherwise, standard methods ignore the knowledge of the past solutions and suboptimally re-solve each problem from scratch. We instantiate Meta OT models in discrete and continuous (Wasserstein-2) settings between images, spherical data, and color palettes and use them to improve the computational time of standard OT solvers by multiple orders of magnitude. | We learn to predict the solution to optimal transport problems |
Explainable recommender systems have attracted much interest in recent years as they can explain their recommendation decisions, enhancing user trust in the systems. Most explainable recommender systems rely on human-generated rationales or annotated aspect features from user reviews to train models for rational generation or extraction. The rationales produced are often confined to a single review. To avoid the expensive human annotation process and to generate explanations beyond individual reviews, we propose an explainable recommender system trained on user reviews by developing a transferable Geometric Information Bottleneck (GIANT), which leverages the prior knowledge acquired through clustering on a user-item graph built on user-item rating interactions, since graph nodes in the same cluster tend to share common characteristics or preferences. We then feed user reviews and item reviews into a variational network to learn latent topic distributions which are regularised by the distributions of user/item estimated based on their distances to various cluster centroids of the user-item graph. By iteratively refining the instance-level review latent topics with GIANT, our method learns a robust latent space from the text for rating prediction and explanation generation. Experimental results on three e-commerce datasets show that our model significantly improves the interpretability of a variational recommender using a standard Gaussian prior, in terms of coherence, diversity and faithfulness, while achieving performance comparable to existing content-based recommender systems in terms of rating prediction accuracy. | To consider user-item interactions for an interpretable recommender system, we propose to incorporate the geometric regularisation derived from user-item interaction graphs to learn the latent factors of review text in a variational network. |
A visual system has to learn both which features to extract from images and how to group locations into (proto-)objects. Those two aspects are usually dealt with separately, although predictability is discussed as a cue for both. To incorporate features and boundaries into the same model, we model a layer of feature maps with a pairwise Markov random field model in which each factor is paired with an additional binary variable, which switches the factor on or off. Using one of two contrastive learning objectives, we can learn both the features and the parameters of the Markov random field factors from images without further supervision signals. The features learned by shallow neural networks based on this loss are local averages, opponent colors, and Gabor-like stripe patterns. Furthermore, we can infer connectivity between locations by inferring the switch variables. Contours inferred from this connectivity perform quite well on the Berkeley segmentation database (BSDS500) without any training on contours. Thus, computing predictions across space aids both segmentation and feature learning, and models trained to optimize these predictions show similarities to the human visual system. We speculate that retinotopic visual cortex might implement such predictions over space through lateral connections. | Using a contrastive loss for a local prediction, we learn a representation of both features and segmentation, which is similar to the human visual system. |
Increasing the size of overparameterized neural networks has been shown to improve their generalization performance. However, real-world datasets often contain a significant fraction of noisy labels, which can drastically harm the performance of the models trained on them. In this work, we study how neural networks' test loss changes with model size when the training set contains noisy labels. We show that under a sufficiently large noise-to-sample size ratio, generalization error eventually increases with model size. First, we provide a theoretical analysis on random feature regression and show that this phenomenon occurs as the variance of the generalization loss experiences a second ascent under large noise-to-sample size ratio. Then, we present extensive empirical evidence confirming that our theoretical results hold for neural networks. Furthermore, we empirically observe that the adverse effect of network size is more pronounced when robust training methods are employed to learn from noisy-labeled data. Our results have important practical implications: First, larger models should be employed with extra care, particularly when trained on smaller dataset or using robust learning methods. Second, a large sample size can alleviate the effect of noisy labels and allow larger models to achieve a superior performance even under noise. | When noise-to-sample-size ratio is sufficiently large, increasing the width or density of the model beyond a certain point only hurts the generalization performance. |
Retrieving candidate items with low latency and computational cost is important for large-scale advertising systems. Negative sampling is a general approach to model million-scale items with rich features in the retrieval. The training-inference inconsistency of data distribution brought from sampling negatives is a key challenge. In this work, we propose a novel negative sampling strategy Consistent Data Distribution Sampling (CDDS) to solve such an issue. Specifically, we employ a relative large-scale of uniform training negatives and batch negatives to adequately train long-tail and hot items respectively, and employ high divergence negatives to improve the learning convergence. To make the above training samples approximate the serving item data distribution, we introduce an auxiliary loss based on an asynchronous item embedding matrix over the entire item pool. Offline experiments on real datasets achieve SOTA performance. Online experiments with multiple advertising scenarios show that our method has achieved significant increases in GMV. The source code will be released in the future. | A novel negative sampling strategy to tackle training-inference inconsistency of data distribution for large-scale retrieval. |
Pre-training serves as a foundation of recent NLP models, where language modeling task is performed over large texts. It has been shown that data affects the quality of pre-training, and curriculum has been investigated regarding sequence length. We consider a linguistic perspective in the curriculum, where frequent words are learned first and rare words last. This is achieved by substituting syntactic constituents for rare words with their constituent labels. By such syntactic substitutions, a curriculum can be made by gradually introducing words with decreasing frequency levels. Without modifying model architectures or introducing external computational overhead, our data-centric method gives better performances over vanilla BERT on various downstream benchmarks. | We propose a language model pre-training method based on linguistically motivated curriculum learning. |
Human visual perception can easily generalize to out-of-distributed visual data, which is far beyond the capability of modern machine learning models. Domain generalization (DG) aims to close this gap, with existing DG methods mainly focusing on the loss function design. In this paper, we propose to explore an orthogonal direction, i.e., the design of the backbone architecture. It is motivated by an empirical finding that transformer-based models trained with empirical risk minimization (ERM) outperform CNN-based models employing state-of-the-art (SOTA) DG algorithms on multiple DG datasets. We develop a formal framework to characterize a network's robustness to distribution shifts by studying its architecture's alignment with the correlations in the dataset. This analysis guides us to propose a novel DG model built upon vision transformers, namely \emph{Generalizable Mixture-of-Experts (GMoE)}. Extensive experiments on DomainBed demonstrate that GMoE trained with ERM outperforms SOTA DG baselines by a large margin. Moreover, GMoE is complementary to existing DG methods and its performance is substantially improved when trained with DG algorithms. | We theoretically investigate the impact of backbone architecture on DG. We propose a novel SOTA model Generalizable Mixture-of-Experts (GMoE) for DG. |
We propose a globally-accelerated, first-order method for the optimization of smooth and (strongly or not) geodesically-convex functions in Hadamard manifolds. Our algorithm enjoys the same convergence rates as Nesterov's accelerated gradient descent, up to a multiplicative geometric penalty and log factors.
Crucially, we can enforce our method to stay within a compact set we define. Prior fully accelerated works resort to assuming that the iterates of their algorithms stay in some pre-specified compact set, except for two previous methods, whose applicability is limited to local optimization and to spaces of constant curvature, respectively. Achieving global and general Riemannian acceleration without iterates assumptively staying in the feasible set was asked as an open question in (Kim & Yang, 2022), which we solve for Hadamard manifolds.
In our solution, we show that we can use a linearly convergent algorithm for constrained strongly g-convex smooth problems to implement a Riemannian inexact proximal point operator that we use as a subroutine, which is of independent interest. | We propose accelerated first-order methods for Riemannian optimization in Hadamard manifolds by using a proximal method that we design. We can work without undesirable assumptions previous accelerated works made |
Online Knowledge Distillation (KD) has an advantage over traditional KD works in that it removes the necessity for a pre-trained teacher. Indeed, an ensemble of small teachers has become typical guidance for a student's learning trajectory. Previous works emphasized diversity to create helpful ensemble knowledge and further argued that the size of diversity should be significant to prevent homogenization. This paper proposes a well-founded online KD framework with naturally derived specialists. In supervised learning, the parameters of a classifier are optimized by stochastic gradient descent based on a training dataset distribution. If the training dataset is shifted, the optimal point and corresponding parameters change accordingly, which is natural and explicit.
We first introduce a label prior shift to induce evident diversity among the same teachers, which assigns a skewed label distribution to each teacher and simultaneously specializes them through importance sampling. Compared to previous works, our specialization achieves the highest level of diversity and maintains it throughout training. Second, we propose a new aggregation that uses post-compensation in specialist outputs and conventional model averaging. The aggregation empirically exhibits the advantage of ensemble calibration even if applied to previous diversity-eliciting methods. Finally, through extensive experiments, we demonstrate the efficacy of our framework on top-1 error rate, negative log-likelihood, and notably expected calibration error. | Online knowledge distillation with an ensemble of specialized teachers that are explicitly estimated for each imbalanced label prior. |
Message passing graph neural networks (GNNs) are known to have their expressiveness upper-bounded by 1-dimensional Weisfeiler-Lehman (1-WL) algorithm. To achieve more powerful GNNs, existing attempts either require ad hoc features, or involve operations that incur high time and space complexities. In this work, we propose a general and provably powerful GNN framework that preserves the scalability of message passing scheme. In particular, we first propose to empower 1-WL for graph isomorphism test by considering edges among neighbors, giving rise to NC-1-WL. The expressiveness of NC-1-WL is shown to be strictly above 1-WL and below 3-WL theoretically. Further, we propose the NC-GNN framework as a differentiable neural version of NC-1-WL. Our simple implementation of NC-GNN is provably as powerful as NC-1-WL. Experiments demonstrate that our NC-GNN achieves remarkable performance on various benchmarks. | We propose a general GNN framework with provably expressive power, while maintaining the scalability of the message passing scheme. |
Common space learning, associating semantic and visual domains in a common
latent space, is essential to transfer knowledge from seen classes to unseen ones
on Zero-Shot Learning (ZSL) realm. Existing methods for common space learning
rely heavily on structure alignment due to the heterogeneous nature between
semantic and visual domains, but the existing design is sub-optimal. In this paper,
we utilize persistent homology to investigate geometry structure alignment,
and observe two following issues: (i) The sampled mini-batch data points present
a distinct structure gap compared to global data points, thus the learned structure
alignment space inevitably neglects abundant and accurate global structure
information. (ii) The latent visual and semantic space fail to preserve multiple
dimensional geometry structure, especially high dimensional structure information.
To address the first issue, we propose a Topology-guided Sampling Strategy
(TGSS) to mitigate the gap between sampled and global data points. Both theoretical
analyses and empirical results guarantee the effectiveness of the TGSS.
To solve the second issue, we introduce a Topology Alignment Module (TAM)
to preserve multi-dimensional geometry structure in latent visual and semantic
space, respectively. The proposed method is dubbed TopoZero. Empirically, our
TopoZero achieves superior performance on three authoritative ZSL benchmark
datasets. | we utilize persistent homology to investigate geometry structure alignment, based on which, we propose a TopoZero framework to achieve multi-dimensional structure alignment. |
An oft-cited challenge of federated learning is the presence of heterogeneity. \emph{Data heterogeneity} refers to the fact that data from different clients may follow very different distributions. \emph{System heterogeneity} refers to client devices having different system capabilities. A considerable number of federated optimization methods address this challenge. In the literature, empirical evaluations usually start federated training from random initialization. However, in many practical applications of federated learning, the server has access to proxy data for the training task that can be used to pre-train a model before starting federated training. Using four standard federated learning benchmark datasets, we empirically study the impact of starting from a pre-trained model in federated learning. Unsurprisingly, starting from a pre-trained model reduces the training time required to reach a target error rate and enables the training of more accurate models (up to 40\%) than is possible when starting from random initialization. Surprisingly, we also find that starting federated learning from a pre-trained initialization reduces the effect of both data and system heterogeneity. We recommend future work proposing and evaluating federated optimization methods to evaluate the performance when starting from random and pre-trained initializations. This study raises several questions for further work on understanding the role of heterogeneity in federated optimization. | Stop worrying about heterogeneity and start from pre-trained weights. |
Data augmentation has recently emerged as an essential component of modern training recipes for visual recognition tasks.
However, data augmentation for video recognition has been rarely explored despite its effectiveness.
Few existing augmentation recipes for video recognition naively extend the image augmentation methods by applying the same operations to the whole video frames.
Our main idea is that the magnitude of augmentation operations for each frame needs to be changed over time to capture the real-world video's temporal variations.
These variations should be generated as diverse as possible using fewer additional hyper-parameters during training.
Through this motivation, we propose a simple yet effective video data augmentation framework, DynaAugment.
The magnitude of augmentation operations on each frame is changed by an effective mechanism, Fourier Sampling that parameterizes diverse, smooth, and realistic temporal variations.
DynaAugment also includes an extended search space suitable for video for automatic data augmentation methods.
DynaAugment experimentally demonstrates that there are additional performance rooms to be improved from static augmentations on diverse video models.
Specifically, we show the effectiveness of DynaAugment on various video datasets and tasks: large-scale video recognition (Kinetics-400 and Something-Something-v2), small-scale video recognition (UCF-101 and HMDB-51), fine-grained video recognition (Diving-48 and FineGym), video action segmentation on Breakfast, video action localization on THUMOS'14, and video object detection on MOT17Det. | We propose a novel data augmentation framework for video recognition that extends the static nature of image augmentations into temporally dynamic. |
Non-contact heart rate estimation has essential implications for the development of affective computing and telemedicine. However, existing deep learning-based methods often endeavor to achieve real-time measurements, so a simple, fast, pre-processing-free approach is needed. Our work consists of two main parts. Firstly, we proposed SEQ-rPPG, which first transforms the RGB frame sequence into the original BVP signal sequence by learning-based linear mapping and then outputs the final BVP signal using 1DCNN-based spectral transform, and time-domain filtering. Secondly, to address the shortcomings of the existing dataset in training the model, a new large-scale dataset was collected for training and testing. Our approach achieved competitive results on the collected large dataset(the best) and public dataset UBFC-rPPG(0.81 MAE with 30s time window, test only). It requires no complex pre-processing, has the fastest speed, can run in real-time on mobile ARM CPUs, and can achieve real-time beat-to-beat performance on desktop CPUs. Benefiting from the high-quality training set, other deep learning-based models reduced errors by at least 53$\%$. We compared the methods with and without the spectral transformation, and the results show that the processing in the time domain is effective. | A new rPPG method is proposed, which is very simple, fast and accurate. |
Research on text-to-image generation has witnessed significant progress in generating diverse and photo-realistic images, driven by diffusion and auto-regressive models trained on large-scale image-text data. Though state-of-the-art models can generate high-quality images of common entities, they often have difficulty generating images of uncommon entities, such as `Chortai (dog)' or `Picarones (food)'. To tackle this issue, we present the Retrieval-Augmented Text-to-Image Generator (Re-Imagen), a generative model that uses retrieved information to produce high-fidelity and faithful images, even for rare or unseen entities. Given a text prompt, Re-Imagen accesses an external multi-modal knowledge base to retrieve relevant (image, text) pairs, and uses them as references to generate the image. With this retrieval step, Re-Imagen is augmented with the knowledge of high-level semantics and low-level visual details of the mentioned entities, and thus improves its accuracy in generating the entities' visual appearances. We train Re-Imagen on a constructed dataset containing (image,text,retrieval) triples to teach the model to ground on both text prompt and retrieval. Furthermore, we develop a new sampling strategy to interleave the classifier-free guidance for text and retrieval condition to balance the text and retrieval alignment. Re-Imagen achieves new SoTA FID results on two image generation benchmarks, such as COCO (\ie, FID = 5.25) and WikiImage (\ie, FID = 5.82) without fine-tuning. To further evaluate the capabilities of the model, we introduce EntityDrawBench, a new benchmark that evaluates image generation for diverse entities, from frequent to rare, across multiple visual domains. Human evaluation on EntityDrawBench shows that Re-Imagen performs on par with the best prior models in photo-realism, but with significantly better real-world faithfulness, especially on less frequent entities. | A text-to-image generation model that can retrieve from external knowledge base to generate more faithful images. |
Machine Translation systems can produce different types of errors, some of which get characterized as critical or catastrophic due to the specific negative impact they can have on users. Automatic or human evaluation metrics do not necessarily differentiate between such critical errors and more innocuous ones. In this paper we focus on one type of critical error: added toxicity. We evaluate and analyze added toxicity when translating a large evaluation dataset (HOLISTICBIAS, over 472k sentences, covering 13 demographic axes) from English into 164 languages. The toxicity automatic evaluation shows that added toxicity across languages varies from 0% to 5%. The output languages with the most added toxicity tend to be low-resource ones, and the demographic axes with the most added toxicity include sexual orientation, gender and sex, and ability. We also perform human evaluation on a subset of 8 directions, confirming the prevalence of true
added toxicity.
We use a measurement of the amount of source contribution to the translation, where a low source contribution implies hallucination, to interpret what causes toxicity. We observe that the source contribution is somewhat correlated with toxicity but that 45.6% of added toxic words have a high source contribution, suggesting that much of the added toxicity may be due to mistranslations. Combining the signal of source contribution level with a measurement of translation robustness allows us to flag 22.3% of added toxicity, suggesting that added toxicity may be related to both hallucination and the stability of translations in different contexts. Given these findings, our recommendations to reduce added toxicity are to curate training data to avoid mistranslations, mitigate hallucination and check unstable translations. | Quantification, analysis and mitigation recommendations of toxicity for 164 languages |
Time series analysis is of immense importance in extensive applications, such as weather forecasting, anomaly detection, and action recognition. This paper focuses on temporal variation modeling, which is the common key problem of extensive analysis tasks. Previous methods attempt to accomplish this directly from the 1D time series, which is extremely challenging due to the intricate temporal patterns. Based on the observation of multi-periodicity in time series, we ravel out the complex temporal variations into the multiple intraperiod- and interperiod-variations. To tackle the limitations of 1D time series in representation capability, we extend the analysis of temporal variations into the 2D space by transforming the 1D time series into a set of 2D tensors based on multiple periods. This transformation can embed the intraperiod- and interperiod-variations into the columns and rows of the 2D tensors respectively, making the 2D-variations to be easily modeled by 2D kernels. Technically, we propose the TimesNet with TimesBlock as a task-general backbone for time series analysis. TimesBlock can discover the multi-periodicity adaptively and extract the complex temporal variations from transformed 2D tensors by a parameter-efficient inception block. Our proposed TimesNet achieves consistent state-of-the-art in five mainstream time series analysis tasks, including short- and long-term forecasting, imputation, classification, and anomaly detection. Code is available at this repository: https://github.com/thuml/TimesNet. | Based on the multi-periodicity, we analyze the intraperiod- and interperiod-variations in 2D space and propose the TimesNet as a task-general model, which achieves consistent state-of-the-art in five mainstream time series analysis tasks. |
We present the combinatorial node labeling framework, which generalizes many prior approaches to solving hard graph optimization problems by supporting problems where solutions consist of arbitrarily many node labels, such as graph coloring. We then introduce a neural network architecture to implement this framework. Our architecture builds on a graph attention network with several inductive biases to improve solution quality and is trained using policy gradient reinforcement learning. We demonstrate our approach on both graph coloring and minimum vertex cover. Our learned heuristics match or outperform classical hand-crafted greedy heuristics and machine learning approaches while taking only seconds on large graphs. We conduct a detailed analysis of the learned heuristics and architecture choices and show that they successfully adapt to different graph structures. | We present the combinatorial node labeling framework and an accompanying neural network architecture to solve hard graph optimization problems. |
Mixup is a data augmentation technique that relies on training using random convex combinations of data points and their labels. In recent years, Mixup has become a standard primitive used in the training of state-of-the-art image classification models due to its demonstrated benefits over empirical risk minimization with regards to generalization and robustness. In this work, we try to explain some of this success from a feature learning perspective. We focus our attention on classification problems in which each class may have multiple associated features (or views) that can be used to predict the class correctly. Our main theoretical results demonstrate that, for a non-trivial class of data distributions with two features per class, training a 2-layer convolutional network using empirical risk minimization can lead to learning only one feature for almost all classes while training with a specific instantiation of Mixup succeeds in learning both features for every class. We also show empirically that these theoretical insights extend to the practical settings of image benchmarks modified to have additional synthetic features. | Mixup with appropriate hyperparameters can learn multiple predictive features for each class in a dataset, even when empirical risk minimization fails to do so. |
Data augmentation is a crucial component in unsupervised contrastive learning (CL). It determines how positive samples are defined and, ultimately, the quality of the representation. Even if efforts have been made to find efficient augmentations for ImageNet, CL underperforms compared to supervised methods and it is still an open problem in other applications, such as medical imaging, or in datasets with easy-to-learn but irrelevant imaging features. In this work, we propose a new way to define positive samples using kernel theory along with a novel loss called \textit{decoupled uniformity}. We propose to integrate prior information, learnt from generative models viewed as feature extractor, or given as auxiliary attributes, into contrastive learning, to make it less dependent on data augmentation. We draw a connection between contrastive learning and the conditional mean embedding theory to derive tight bounds on the downstream classification loss. In an unsupervised setting, we empirically demonstrate that CL benefits from generative models, such as VAE and GAN, to less rely on data augmentations. We validate our framework on vision and medical datasets including CIFAR10, CIFAR100, STL10, ImageNet100, CheXpert and a brain MRI dataset. In the weakly supervised setting, we demonstrate that our formulation provides state-of-the-art results. | Improving positive sampling in contrastive learning using kernel |
System identification, also known as learning forward models, transfer functions, system dynamics, etc., has a long tradition both in science and engineering in different fields. Particularly, it is a recurring theme in Reinforcement Learning research, where forward models approximate the state transition function of a Markov Decision Process by learning a mapping function from current state and action to the next state. This problem is commonly defined as a Supervised Learning problem in a direct way. This common approach faces several difficulties due to the inherent complexities of the dynamics to learn, for example, delayed effects, high non-linearity, non-stationarity, partial observability and, more important, error accumulation when using bootstrapped predictions (predictions based on past predictions), over large time horizons. Here we explore the use of Reinforcement Learning in this problem. We elaborate on why and how this problem fits naturally and sound as a Reinforcement Learning problem, and present some experimental results that demonstrate RL is a promising technique to solve these kind of problems. | System Identification as a Reinforcement Learning Problem |
Joint-Embedding Self Supervised Learning (JE-SSL) has seen a rapid development, with the emergence of many method variations and few principled guidelines that would help practitioners to successfully deploy those methods. The main reason for that pitfall actually comes from JE-SSL's core principle of not employing any input reconstruction. Without any visual clue, it becomes extremely cryptic to judge the quality of a learned representation without having access to a labelled dataset. We hope to correct those limitations by providing a single --theoretically motivated-- criterion that reflects the quality of learned JE-SSL representations: their effective rank. Albeit simple and computationally friendly, this method ---coined {\em RankMe}--- allows one to assess the performance of JE-SSL representations, even on different downstream datasets, without requiring any labels, training or parameters to tune. Through thorough empirical experiments involving hundreds of repeated training episodes, we demonstrate how RankMe can be used for hyperparameter selection with nearly no loss in final performance compared to the current selection method that involve dataset labels. We hope that RankMe will facilitate the use of JE-SSL in domains with little or no labeled data. | We show how the rank of representations can be used to measure their dowstream performance, even on unseen datasets. We validate this simple metric with thorough experiments and show its power for hyperparameter selection. |
A Markov network characterizes the conditional independence structure, or Markov property, among a set of random variables. Existing work focuses on specific families of distributions (e.g., exponential families) and/or certain structures of graphs, and most of them can only handle variables of a single data type (continuous or discrete). In this work, we characterize the conditional independence structure in general distributions for all data types (i.e., continuous, discrete, and mixed-type) with a Generalized Precision Matrix (GPM). Besides, we also allow general functional relations among variables, thus giving rise to a Markov network structure learning algorithm in one of the most general settings. To deal with the computational challenge of the problem, especially for large graphs, we unify all cases under the same umbrella of a regularized score matching framework. We validate the theoretical results and demonstrate the scalability empirically in various settings. | We investigate scalable estimation of nonparametric Markov networks with general distributions for all data types (i.e., continuous, discrete, and mixed-type). |
Finding accurate Machine Learning pipelines is essential in achieving state-of-the-art AI predictive performance. Unfortunately, most existing Pipeline Optimization techniques rely on flavors of Bayesian Optimization that do not explore the deep interaction between pipeline stages/components (e.g. between hyperparameters of the deployed preprocessing algorithm and the hyperparameters of a classifier). In this paper, we are the first to capture the deep interaction between components of a Machine Learning pipeline. We propose embedding pipelines in a deep latent representation through a novel per-component encoder mechanism. Such pipeline embeddings are used with deep kernel Gaussian Process surrogates inside a Bayesian Optimization setup. Through extensive experiments on large-scale meta-datasets, we demonstrate that learning pipeline embeddings with Deep Neural Networks significantly advances the state-of-the-art in Pipeline Optimization. | How to learn Machine Learning pipelines representations to improve their optimization |
Contrastive learning, especially self-supervised contrastive learning (SSCL), has achieved great success in extracting powerful features from unlabeled data. In this work, we contribute to the theoretical understanding of SSCL and uncover its connection to the classic data visualization method, stochastic neighbor embedding (SNE), whose goal is to preserve pairwise distances. From the perspective of preserving neighboring information, SSCL can be viewed as a special case of SNE with the input space pairwise similarities specified by data augmentation. The established correspondence facilitates deeper theoretical understanding of learned features of SSCL, as well as methodological guidelines for practical improvement. Specifically, through the lens of SNE, we provide novel analysis on domain-agnostic augmentations, implicit bias and robustness of learned features. To illustrate the practical advantage, we demonstrate that the modifications from SNE to $t$-SNE can also be adopted in the SSCL setting, achieving significant improvement in both in-distribution and out-of-distribution generalization. | This work proposes a novel perspective that interprets SSCL methods as a type of SNE methods, which facilitates both deeper theoretical understandings of SSCL, and methodological guidelines for practical improvement. |
Advances in reinforcement learning have led to its successful application in complex tasks with continuous state and action spaces. Despite these advances in practice, most theoretical work pertains to finite state and action spaces. We propose building a theoretical understanding of continuous state and action spaces by employing a geometric lens. Central to our work is the idea that the transition dynamics induce a low dimensional manifold of reachable states embedded in the high-dimensional nominal state space. We prove that, under certain conditions, the dimensionality of this manifold is at most the dimensionality of the action space plus one. This is the first result of its kind, linking the geometry of the state space to the dimensionality of the action space. We empirically corroborate this upper bound for four MuJoCo environments.We further demonstrate the applicability of our result by learning a policy in this low dimensional representation. To do so we introduce an algorithm that learns a mapping to a low dimensional representation, as a narrow hidden layer of a deep neural network, in tandem with the policy using DDPG. Our experiments show that a policy learnt this way perform on par or better for four MuJoCo control suite tasks. | We prove that the state space is a low dimensional manifold and show that DDPG can effectively learn in this low dimensional space |
Monotonic linear interpolation (MLI) --- on the line connecting a random initialization with the minimizer it converges to, the loss and accuracy are monotonic --- is a phenomenon that is commonly observed in the training of neural networks. Such a phenomenon may seem to suggest that optimization of neural networks is easy. In this paper, we show that the MLI property is not necessarily related to the hardness of optimization problems, and empirical observations on MLI for deep neural networks depend heavily on the biases. In particular, we show that interpolating both weights and biases linearly leads to very different influences on the final output, and when different classes have different last-layer biases on a deep network, there will be a long plateau in both the loss and accuracy interpolation (which existing theory of MLI cannot explain). We also show how the last-layer biases for different classes can be different even on a perfectly balanced dataset using a simple model. Empirically we demonstrate that similar intuitions hold on practical networks and realistic datasets. | We explain the long plateau in the loss and error curves along the linear interpolation from the initialization to the minimum. |
Reconstructing natural images from fMRI recordings is a challenging task of great importance in neuroscience. The current architectures are bottlenecked because they fail to effectively capture the hierarchical processing of visual stimuli that takes place in the human brain. Motivated by that fact, we introduce a novel neural network architecture for the problem of neural decoding. Our architecture uses Hierarchical Variational Autoencoders (HVAEs) to learn meaningful representations of natural images and leverages their latent space hierarchy to learn voxel-to-image mappings. By mapping the early stages of the visual pathway to the first set of latent variables and the higher visual cortex areas to the deeper layers in the latent hierarchy, we are able to construct a latent variable neural decoding model that replicates the hierarchical visual information processing. Our model achieves better reconstructions compared to the state of the art and our ablation study indicates that the hierarchical structure of the latent space is responsible for that performance. | We propose a novel architecture for decoding visual imagery from fMRI recordings using Hierarchical VAEs. |
Object rearrangement is a challenge for embodied agents because solving these tasks requires generalizing across a combinatorially large set of configurations of entities and their locations. Worse, the representations of these entities are unknown and must be inferred from sensory percepts. We present a hierarchical abstraction approach to uncover these underlying entities and achieve combinatorial generalization from unstructured visual inputs. By constructing a factorized transition graph over clusters of entity representations inferred from pixels, we show how to learn a correspondence between intervening on states of entities in the agent's model and acting on objects in the environment. We use this correspondence to develop a method for control that generalizes to different numbers and configurations of objects, which outperforms current offline deep RL methods when evaluated on simulated rearrangement tasks. | We demonstrate how to generalize over a combinatorially large space of rearrangement tasks from only pixel observations by constructing from video demonstrations a factorized transition graph over entity state transitions that we use for control. |
The combinatorial pure exploration of causal bandits is the following online learning task: given a causal graph with unknown causal inference distributions, in each round we choose a subset of variables to intervene or do no intervention, and observe the random outcomes of all random variables, with the goal that using as few rounds as possible, we can output an intervention that gives the best (or almost best) expected outcome on the reward variable $Y$ with probability at least $1-\delta$, where $\delta$ is a given confidence level. We provide the first gap-dependent and fully adaptive pure exploration algorithms on two types of causal models --- the binary generalized linear model (BGLM) and general graphs. For BGLM, our algorithm is the first to be designed specifically for this setting and achieves polynomial sample complexity, while all existing algorithms for general graphs have either sample complexity exponential to the graph size or some unreasonable assumptions. For general graphs, our algorithm provides a significant improvement on sample complexity, and it nearly matches the lower bound we prove. Our algorithms achieve such improvement by a novel integration of prior causal bandit algorithms and prior adaptive pure exploration algorithms, the former of which utilize the rich observational feedback in causal bandits but are not adaptive to reward gaps, while the latter of which have the issue in reverse. | Combinatorial pure exploration algorithm of causal bandits on two different models |
Future- or return-conditioned supervised learning is an emerging paradigm for offline reinforcement learning (RL), in which the future outcome (i.e., return) associated with a sequence of actions in an offline dataset is used as input to a policy trained to imitate those same actions. While return-conditioning is at the heart of popular algorithms such as decision transformer (DT), these methods tend to perform poorly in highly stochastic environments, where an occasional high return associated with a sequence of actions may be due more to the randomness of the environment than to the actions themselves. Such situations can lead to a learned policy that is inconsistent with its conditioning inputs; i.e., using the policy – while conditioned on a specific desired return – to act in the environment can lead to a distribution of real returns that is wildly different than desired. In this work, we propose the dichotomy of control (DoC), a future-conditioned supervised learning framework that separates mechanisms within a policy’s control (actions) from those outside of a policy’s control (environment stochasticity). We achieve this by conditioning the policy on a latent variable representation of the future and designing a mutual information constraint that removes any future information from the latent variable that is only due to randomness of the environment. Theoretically, we show that DoC yields policies that are consistent with their conditioning inputs, ensuring that conditioning a learned policy on a desired high-return future outcome will correctly induce high-return behavior. Empirically, we show that DoC is able to achieve significantly better performance than DT on environments with highly stochastic rewards (e.g., Bandit) and transitions (e.g., FrozenLake). | We propose dichotomy of control (DoC) for supervised learning in stochastic environments by separating things within a policy's control (actions) from those outside of a policy’s control (env stochasticity) through a mutual information constraint. |
In a backdoor attack, an attacker injects corrupted examples into the training set. The goal of the attacker is to cause the final trained model to predict the attacker's desired target label when a predefined trigger is added to test inputs. Central to these attacks is the trade-off between the success rate of the attack and the number of corrupted training examples injected. We pose this attack as a novel bilevel optimization problem: construct strong poison examples that maximize the attack success rate of the trained model. We use neural tangent kernels to approximate the training dynamics of the model being attacked and automatically learn strong poison examples. We experiment on subclasses of CIFAR-10 and ImageNet with WideResNet-34 and ConvNeXt architectures on periodic and patch trigger attacks and show that NTBA-designed poisoned examples achieve, for example, an attack success rate of 90% with ten times smaller number of poison examples injected compared to the baseline. We provided an interpretation of the NTBA-designed attacks using the analysis of kernel linear regression. We further demonstrate a vulnerability in overparametrized deep neural networks, which is revealed by the shape of the neural tangent kernel.
| We provide a new algorithm based on intuitions from kernel regression that uses neural tangent kernels to design stronger backdoor attacks. |
Equipping predicted segmentation with calibrated uncertainty is essential for safety-critical applications. In this work, we focus on capturing the data-inherent uncertainty (aka aleatoric uncertainty) in segmentation, typically when ambiguities exist in input images. Due to the high-dimensional output space and potential multiple modes in segmenting ambiguous images, it remains challenging to predict well-calibrated uncertainty for segmentation. To tackle this problem, we propose a novel mixture of stochastic experts (MoSE) model, where each expert network estimates a distinct mode of the aleatoric uncertainty and a gating network predicts the probabilities of an input image being segmented in those modes. This yields an efficient two-level uncertainty representation. To learn the model, we develop a Wasserstein-like loss that directly minimizes the distribution distance between the MoSE and ground truth annotations. The loss can easily integrate traditional segmentation quality measures and be efficiently optimized via constraint relaxation. We validate our method on the LIDC-IDRI dataset and a modified multimodal Cityscapes dataset. Results demonstrate that our method achieves the state-of-the-art or competitive performance on all metrics. | We propose a novel mixture of stochastic experts (MoSE) model training with a Wasserstein-like loss, which produces an efficient two-level representation for the multi-modal aleatoric uncertainty in semantic segmentation. |
We propose ProsodyBERT, a self-supervised approach to learning prosody representations from raw audio. Different from most previous works, which use information bottlenecks to disentangle prosody features from speech content and speaker information, we perform an offline clustering of speaker-normalized prosody-related features (energy, pitch, their dynamics, etc.) and use the cluster labels as targets for HuBERT-like masked unit prediction. A span boundary loss is also introduced to capture long-range prosodic information. We demonstrate the effectiveness of ProsodyBERT on a multi-speaker style-controllable text-to-speech (TTS) system. Experiments show that the TTS system trained with ProsodyBERT features can generate natural and expressive speech samples, surpassing the model supervised by energy and pitch on subjective human evaluation. Also, the style and expressiveness of synthesized audio can be controlled by manipulating the prosody features. In addition, We achieve new state-of-the-art results on the IEMOCAP emotion recognition task by combining our prosody features with HuBERT features, showing that ProsodyBERT is complementary to popular pretrained speech self-supervised models. | a self-supervised approach to learning prosody representations from raw audio |
We introduce Discrete Predictor-Corrector diffusion models (DPC), extending predictor-corrector samplers in Gaussian diffusion models to the discrete case. Predictor-corrector samplers are a class of samplers for diffusion models, which improve on ancestral samplers by correcting the sampling distribution of intermediate diffusion states using MCMC methods. In DPC, the Langevin corrector, which does not have a direct counterpart in discrete space, is replaced with a discrete MCMC transition defined by a learned corrector kernel. The corrector kernel is trained to make the correction steps achieve asymptotic convergence, in distribution, to the correct marginal of the intermediate diffusion states. Equipped with DPC, we revisit recent transformer-based non-autoregressive generative models through the lens of discrete diffusion, and find that DPC can alleviate the compounding decoding error due to the parallel sampling of visual tokens. Our experiments show that DPC improves upon existing discrete latent space models for class-conditional image generation on ImageNet, and outperforms continuous diffusion models and GANs, according to standard metrics and user preference studies. | We propose a learned predictor-corrector sampler for discrete diffusion models and empirically demonstrate its effectiveness for image generation. |
Designing an accurate and explainable reward function for many Reinforcement Learning tasks is a cumbersome and tedious process.
Instead, learning policies directly from the feedback of human teachers naturally integrates human domain knowledge into the policy optimization process.
However, different feedback modalities, such as demonstrations and preferences, provide distinct benefits and disadvantages. For example, demonstrations convey a lot of information about the task but are often hard or costly to obtain from real experts while preferences typically contain less information but are in most cases cheap to generate.
However, existing methods centered around human feedback mostly focus on a single teaching modality, causing them to miss out on important training data while making them less intuitive to use.
In this paper we propose a novel method for policy learning that incorporates two different feedback types, namely \emph{demonstrations} and \emph{preferences}.
To this end, we make use of the connection between discriminator training and density ratio estimation to incorporate preferences into the popular Adversarial Imitation Learning paradigm.
This insight allows us to express loss functions over both demonstrations and preferences in a unified framework.
Besides expert demonstrations, we are also able to learn from imperfect ones and combine them with preferences to achieve improved task performance.
We experimentally validate the effectiveness of combining both preferences and demonstrations on common benchmarks and also show that our method can efficiently learn challenging robot manipulation tasks. | We extend Adversarial Imitation Learning to simultaneously utilize both demonstrations and preferences. |
With the recent advance of geometric deep learning, neural networks have been extensively used for data in non-Euclidean domains. In particular, hyperbolic neural networks have proved successful in processing hierarchical information of data. However, many hyperbolic neural networks are numerically unstable during training, which precludes using complex architectures. This crucial problem makes it difficult to build hyperbolic generative models for real and complex data. In this work, we propose a hyperbolic generative network in which we design novel architecture and layers to guarantee stable training. Our proposed network contains three parts: first, a hyperbolic autoencoder (AE) that produces hyperbolic embedding for input data; second, a hyperbolic generative adversarial network (GAN) for generating the hyperbolic latent embedding of the AE from simple noise; third, a generator that inherits the decoder from the AE and the generator from the GAN. We call this network the hyperbolic AE-GAN, or HAEGAN for short. The architecture of HAEGAN fosters expressive representation in the hyperbolic space, and the specific design of layers ensures numerical stability. Experiments show that HAEGAN is able to generate complex data with state-of-the-art structure-related performance. | A hyperbolic generative network numerically stable for generating complex data. |
Motivated by the interests of social network analysis and network-based recommendation systems, we consider a semi-supervised community detection problem, where the goal is to estimate the community label of a new node by leveraging on the network structure and partially observed community labels of existing nodes.
We model the network with a degree-corrected stochastic block model, which allows for severe degree heterogeneity and potentially non-assortative communities.
We propose a fast algorithm that computes a `structural similarity metric' between the new node and each of the $K$ communities, aggregating information in labeled and unlabeled data. The estimated label of the new node is equal to the value of $k$ that maximizes this similarity metric. Our method is computationally fast and compares favorably with existing semi-supervised algorithms on numerical performance. In theory, we derive explicit bounds for the misclassification error and show the efficiency of our method by comparing it with an ideal classifier. To our best knowledge, our results provide the first semi-supervised community detection algorithm with theoretical guarantees. | We propose a fast semi-supervised community detection algorithm AngleMin+ based on the structural similarity metric of DCBM, which is able to address degree heterogeneity and non-assortative network and possesses nice theoretical guarantees. |
Machine learning systems may encounter unexpected problems when the data distribution changes in the deployment environment. A major reason is that certain combinations of domains and labels are not observed during training but appear in the test environment. Although various invariance-based algorithms can be applied, we find that the performance gain is often marginal. To formally analyze this issue, we provide a unique algebraic formulation of the combination shift problem based on the concepts of homomorphism, equivariance, and a refined definition of disentanglement. The algebraic requirements naturally derive a simple yet effective method, referred to as equivariant disentangled transformation (EDT), which augments the data based on the algebraic structures of labels and makes the transformation satisfy the equivariance and disentanglement requirements. Experimental results demonstrate that invariance may be insufficient, and it is important to exploit the equivariance structure in the combination shift problem. | Learning data augmentations based on the algebraic structures of labels is a promising approach for combination shift. |
Adversarial attacks in reinforcement learning (RL) often assume highly-privileged access to the victim’s parameters, environment, or data. Instead, this paper proposes a novel adversarial setting called a Cheap Talk MDP in which an Adversary can merely append deterministic messages to the Victim’s observation, resulting in a minimal range of influence. The Adversary cannot occlude ground truth, influence underlying environment dynamics or reward signals, introduce non-stationarity, add stochasticity, see the Victim’s actions, or access their parameters. Additionally, we present a simple meta-learning algorithm called Adversarial Cheap Talk (ACT) to train Adversaries in this setting. We demonstrate that an Adversary trained with ACT can still significantly influence the Victim’s training and testing performance, despite the highly constrained setting. Affecting train-time performance reveals a new attack vector and provides insight into the success and failure modes of existing RL algorithms. More specifically, we show that an ACT Adversary is capable of harming performance by interfering with the learner’s function approximation, or instead helping the Victim’s performance by outputting useful features. Finally, we show that an ACT Adversary can manipulate messages during train-time to directly and arbitrarily control the Victim at test-time. | We can cause an RL agent to fail, succeed, or be manipulatable by deterministically perturbing irrelevant features in its observation during training. |
Optimizing functions without access to gradients is the remit of black-box meth- ods such as evolution strategies. While highly general, their learning dynamics are often times heuristic and inflexible — exactly the limitations that meta-learning can address. Hence, we propose to discover effective update rules for evolution strategies via meta-learning. Concretely, our approach employs a search strategy parametrized by a self-attention-based architecture, which guarantees the update rule is invariant to the ordering of the candidate solutions. We show that meta-evolving this system on a small set of representative low-dimensional analytic optimization problems is sufficient to discover new evolution strategies capable of generalizing to unseen optimization problems, population sizes and optimization horizons. Furthermore, the same learned evolution strategy can outperform established neuroevolution baselines on supervised and continuous control tasks. As additional contributions, we ablate the individual neural network components of our method; reverse engineer the learned strategy into an explicit heuristic form, which remains highly competitive; and show that it is possible to self-referentially train an evolution strategy from scratch, with the learned update rule used to drive the outer meta-learning loop. | We meta-learn evolution strategies, which flexibly generalize to unseen optimization problems, population sizes and optimization horizons. |
Modern image classifiers achieve high predictive accuracy, but the predictions typically come without reliable uncertainty estimates. Conformal prediction algorithms provide uncertainty estimates by predicting a set of classes based on the probability estimates of the classifier (for example, the softmax scores). To provide such sets, conformal prediction algorithms often rely on estimating a cutoff threshold for the probability estimates, and this threshold is chosen based on a calibration set. Conformal prediction methods guarantee reliability only when the calibration set is from the same distribution as the test set. Therefore, the methods need to be recalibrated for new distributions. However, in practice, labeled data from new distributions is rarely available, making calibration infeasible. In this work, we consider the problem of predicting the cutoff threshold for a new distribution only based on unlabeled examples. While it is impossible in general to guarantee reliability when calibrating based on unlabeled examples, we show that our method provides excellent uncertainty estimates under natural distribution shifts. | We propose a novel test-time recalibration method for conformal prediction based on unlabeled examples that provides excellent uncertainty estimates under natural distribution shifts. |
Can continuous diffusion models bring the same performance breakthrough on natural language they did for image generation? To circumvent the discrete nature of text data, we can simply project tokens in a continuous space of embeddings, as is standard in language modeling. We propose Self-conditioned Embedding Diffusion (SED), a continuous diffusion mechanism that operates on token embeddings and allows to learn flexible and scalable diffusion models for both conditional and unconditional text generation. Through qualitative and quantitative evaluation, we show that our text diffusion models generate samples comparable with those produced by standard autoregressive language models — while being in theory more efficient on accelerator hardware at inference time. Our work paves the way for scaling up diffusion models for text, similarly to autoregressive models, and for improving performance with recent refinements to continuous diffusion. | Our continuous diffusion framework operates on word embeddings, enabling flexible and scalable diffusion models for text generation. |
Inferring causal structure from data is a challenging task of fundamental importance in science. Observational data are often insufficient to identify a system’s causal structure uniquely. While conducting interventions (i.e., experiments) can improve the identifiability, such samples are usually challenging and expensive to obtain. Hence, experimental design approaches for causal discovery aim to minimize the number of interventions by estimating the most informative intervention target. In this work, we propose a novel gradient-based intervention targeting method, abbreviated GIT, that 'trusts' the gradient estimator of a gradient-based causal discovery framework to provide signals for intervention acquisition function. We provide extensive experiments in simulated and real-world datasets and demonstrate that GIT performs on par with competitive baselines, surpassing them in the low-data regime. | We propose GIT, a novel gradient-based intervention targeting method, which improves the performance of causal discovery, especially in the low data regime. |
Automatically optimizing the hyperparameters of Machine Learning algorithms is one of the primary open questions in AI. Existing work in Hyperparameter Optimization (HPO) trains surrogate models for approximating the response surface of hyperparameters as a regression task. In contrast, we hypothesize that the optimal strategy for training surrogates is to preserve the ranks of the performances of hyperparameter configurations as a Learning to Rank problem. As a result, we present a novel method that meta-learns neural network surrogates optimized for ranking the configurations' performances while modeling their uncertainty via ensembling. In a large-scale experimental protocol comprising 12 baselines, 16 HPO search spaces and 86 datasets/tasks, we demonstrate that our method achieves new state-of-the-art results in HPO. | Meta-learn Deep Ensembles using Ranking Losses to improve the performance on Hyperparameter Optimization |
Construction of a scaffold structure that supports a desired motif, conferring protein function, shows promise for the design of vaccines and enzymes. But a general solution to this motif-scaffolding problem remains open. Current machine-learning techniques for scaffold design are either limited to unrealistically small scaffolds (up to length 20) or struggle to produce multiple diverse scaffolds. We propose to learn a distribution over diverse and longer protein backbone structures via an E(3)-equivariant graph neural network. We develop SMCDiff to efficiently sample scaffolds from this distribution conditioned on a given motif; our algorithm is the first to theoretically guarantee conditional samples from a diffusion model in the large-compute limit. We evaluate our designed backbones by how well they align with AlphaFold2-predicted structures. We show that our method can (1) sample scaffolds up to 80 residues and (2) achieve structurally diverse scaffolds for a fixed motif. | We have created the first generative modeling approach to motif-scaffolding by developing a diffusion probabilistic model of protein backbones and a procedure for generating scaffolds conditional on a motif. |
De novo molecular generation is an essential task for science discovery. Recently, fragment-based deep generative models have attracted much research attention due to their flexibility in generating novel molecules based on existing molecule fragments. However, the motif vocabulary, i.e., the collection of frequent fragments, is usually built upon heuristic rules, which brings difficulties to capturing common substructures from large amounts of molecules. In this work, we propose MiCaM to generate molecules based on mined connection-aware motifs. Specifically, it leverages a data-driven algorithm to automatically discover motifs from a molecule library by iteratively merging subgraphs based on their frequency. The obtained motif vocabulary consists of not only molecular motifs (i.e., the frequent fragments), but also their connection information, indicating how the motifs are connected with each other. Based on the mined connection-aware motifs, MiCaM builds a connection-aware generator, which simultaneously picks up motifs and determines how they are connected. We test our method on distribution-learning benchmarks (i.e., generating novel molecules to resemble the distribution of a given training set) and goal-directed benchmarks (i.e., generating molecules with target properties), and achieve significant improvements over previous fragment-based baselines. Furthermore, we demonstrate that our method can effectively mine domain-specific motifs for different tasks. | We propose a fragment-based model for molecular generation. It first mines connection-aware motifs from the molecule library and then leverage a connection-aware generator to generate novel drug candidates. |
A fundamental challenge in physics-informed machine learning (PIML) is the design of robust PIML methods for out-of-distribution (OOD) forecasting tasks, where the tasks require learning-to-learn from observations of the same (ODE) dynamical system with different unknown parameters, and demand accurate forecasts even under initial conditions outside the training support. In this work we propose a solution for such tasks, which we define as a meta-learning procedure for causal structural discovery (including invariant risk minimization). Using three different OOD tasks, we empirically observe that the proposed approach significantly outperforms existing state-of-the-art PIML and deep learning methods. | This work proposes combining causal structural discovery, invariant risk minimization, and meta-learning in order to make Physics-informed Machine Learning robust to OOD tasks. |
Federated adversarial training can effectively complement adversarial robustness into the privacy-preserving federated learning systems. However, the high demand for memory capacity and computing power makes large-scale federated adversarial training infeasible on resource-constrained edge devices. Few previous studies in federated adversarial training have tried to tackle both memory and computational constraints at the same time. In this paper, we propose a new framework named Federated Adversarial Decoupled Learning (FADE) to enable AT on resource-constrained edge devices. FADE decouples the entire model into small modules to fit into the resource budget of each edge device respectively, and each device only needs to perform AT on a single module in each communication round. We also propose an auxiliary weight decay to alleviate objective inconsistency and achieve better accuracy-robustness balance in FADE. FADE offers a theoretical guarantee for convergence and adversarial robustness, and our experimental results show that FADE can significantly reduce the consumption of memory and computing power while maintaining accuracy and robustness. | We propose a novel framework to enable large-scale federated adversarial training on resource-constrained edge devices. |
The combination of knowledge distillation with contrastive learning has great potential to distill structural knowledge. Most of the contrastive-learning-based distillation methods treat the entire training dataset as the memory bank and maintain two memory banks, one for the student and one for the teacher. Besides, the representations in the two memory banks are updated in a momentum manner, leading to representation inconsistency. In this work, we propose Contrastive Consistent Representation Distillation (CoCoRD) to provide consistent representations for efficient contrastive-learning-based distillation. Instead of momentum-updating the cached representations, CoCoRD updates the encoders in a momentum manner. Specifically, the teacher is equipped with a momentum-updated projection head to generate consistent representations. These teacher representations are cached in a fixed-size queue which serves as the only memory bank in CoCoRD and is significantly smaller than the entire training dataset. Additionally, a slow-moving student, implemented as a momentum-based moving average of the student, is built to facilitate contrastive learning. CoCoRD, which utilizes only one memory bank and much fewer negative keys, provides highly competitive results under typical teacher-student settings. On ImageNet, CoCoRD-distilled ResNet50 outperforms the teacher ResNet101 by 0.2% top-1 accuracy. Furthermore, in PASCAL VOC and COCO detection, the detectors whose backbones are initialized by CoCoRD-distilled models exhibit considerable performance improvements. | We propose Contrastive Consistent Representation Distillation (CoCoRD) to provide consistent representations for efficient contrastive-learning-based distillation. |
Self-Supervised Learning (SSL) methods operate on unlabeled data to learn robust representations useful for downstream tasks. Most SSL methods rely on augmentations obtained by transforming the 2D image pixel map. These augmentations ignore the fact that biological vision takes place in an immersive three-dimensional, temporally contiguous environment, and that low-level biological vision relies heavily on depth cues. Using a signal provided by a pretrained state-of-the-art RGB-to-depth model (the Depth Prediction Transformer, Ranftl et al., 2021), we explore two distinct approaches to incorporating depth signals into the SSL framework. First, we evaluate contrastive learning using an RGB+depth input representation. Second, we use the depth signal to generate novel views from slightly different camera positions, thereby producing a 3D augmentation for contrastive learning. We evaluate these two approaches on three different SSL methods---BYOL, SimSiam, and SwAV---using ImageNette (10 class subset of ImageNet) and ImageNet-100. We find that both approaches to incorporating depth signals improve the robustness and generalization of the baseline SSL methods, though the first approach (with depth-channel concatenation) is superior. | Depth signal improves contrastive learning |
Visual description, detection, and matching of keypoints in images are fundamental components of many computer vision problems, such as camera tracking and (re)localization. Recently, learning-based feature extractors on top of convolutional neural networks (CNNs) have achieved state-of-the-art performance. In this paper, we further explore the usage of CNN and present a new approach that ensembles randomly initialized CNNs without any training. Our observation is that the CNN architecture inherently extracts features with certain extents of robustness to viewpoint/illumination changes and thus, it can be regarded as visual descriptors. Consequently, randomized CNNs serve as descriptor extractors and a subsequent consensus mechanism detects keypoints using them. Such description and detection pipeline can be used to match keypoints in images and achieves higher generalization ability than the state-of-the-art methods in our experiments. | A new approach that uses convolutional neural networks (CNNs) for keypoint description, detection, and matching, without requiring the deep networks to be trained. |
This paper proposes a novel batch normalization strategy for test-time adaptation. Recent test-time adaptation methods heavily rely on the modified batch normalization, i.e., transductive batch normalization (TBN), which calculates the mean and the variance from the current test batch rather than using the running mean and variance obtained from the source data, i.e., conventional batch normalization (CBN). Adopting TBN that employs test batch statistics mitigates the performance degradation caused by the domain shift. However, re-estimating normalization statistics using test data depends on impractical assumptions that a test batch should be large enough and be drawn from i.i.d. stream, and we observed that the previous methods with TBN show critical performance drop without the assumptions. In this paper, we identify that CBN and TBN are in a trade-off relationship and present a new test-time normalization (TTN) method that interpolates the statistics by adjusting the importance between CBN and TBN according to the domain-shift sensitivity of each BN layer. Our proposed TTN improves model robustness to shifted domains across a wide range of batch sizes and in various realistic evaluation scenarios. TTN is widely applicable to other test-time adaptation methods that rely on updating model parameters via backpropagation. We demonstrate that adopting TTN further improves their performance and achieves state-of-the-art performance in various standard benchmarks. | We propose a test-time batch normalization method, which interpolates source and current batch statistics considering each layer's domain-shift sensitivity level that shows robust performance over various realistic evaluation scenarios.. |
Class-incremental few-shot learning, where new sets of classes are provided sequentially with only a few training samples, presents a great challenge due to catastrophic forgetting of old knowledge and overfitting caused by lack of data. During finetuning on new classes, the performance on previous classes deteriorates quickly even when only a small fraction of parameters are updated, since the previous knowledge is broadly associated with most of the model parameters in the original parameter space. In this paper, we introduce WaRP, the \textit{weight space rotation process}, which transforms the original parameter space into a new space so that we can push most of the previous knowledge compactly into only a few important parameters. By properly identifying and freezing these key parameters in the new weight space, we can finetune the remaining parameters without affecting the knowledge of previous classes. As a result, WaRP provides an additional room for the model to effectively learn new classes in future incremental sessions. Experimental results confirm the effectiveness of our solution and show the improved performance over the state-of-the-art methods. | This paper introduced a concept of weight space rotation which makes changes to parameter space itself for solving incremental few-shot learning problem. |
Machine learning models are vulnerable to Out-Of-Distribution (OOD) examples, such a problem has drawn much attention. However, current methods lack a full understanding of different types of OOD data: there are benign OOD data that can be properly adapted to enhance the learning performance, while other malign OOD data would severely degenerate the classification result. To Harness OOD data, this paper proposes HOOD method that can leverage the content and style from each image instance to identify benign and malign OOD data. Particularly, we design a variational inference framework to causally disentangle content and style features by constructing a structural causal model. Subsequently, we augment the content and style through an intervention process to produce malign and benign OOD data, respectively. The benign OOD data contain novel styles but hold our interested contents, and they can be leveraged to help train a style-invariant model. In contrast, the malign OOD data inherit unknown contents but carry familiar styles, by detecting them can improve model robustness against deceiving anomalies. Thanks to the proposed novel disentanglement and data augmentation techniques, HOOD can effectively deal with OOD examples in unknown and open environments, whose effectiveness is empirically validated in three typical OOD applications including OOD detection, open-set semi-supervised learning, and open-set domain adaptation. | Harnessing OOD examples through data augmentation that changes the content and style. |
Large language models (LLMs) have been shown to be capable of impressive few-shot generalisation to new tasks. However, they still tend to perform poorly on multi-step logical reasoning problems. Here we carry out a comprehensive evaluation of LLMs on 46 tasks that probe different aspects of logical reasoning. We show that language models tend to perform fairly well at single step inference or entailment tasks, but struggle to chain together multiple reasoning steps to solve more complex problems. In light of this, we propose a Selection-Inference (SI) framework that exploits pre-trained LLMs as general processing modules, and alternates between selection and inference to generate a series of interpretable, casual reasoning steps leading to the final answer. We show that a 7B parameter LLM used within the SI framework in a 5-shot generalisation setting, with no fine-tuning, yields a performance improvement of over 100% compared to an equivalent vanilla baseline on a suite of 10 logical reasoning tasks. The same model in the same setting even outperforms a significantly larger 280B parameter baseline on the same suite of tasks. Moreover, answers produced by the SI framework are accompanied by a causal natural-language-based reasoning trace, which has important implications for the safety and trustworthiness of the system. | Using language models to produce a human interpretable chain of logical reasoning to answer questions. |
Based on the model's resilience to computational noise, model quantization is important for compressing models and improving computing speed. Existing quantization techniques rely heavily on experience and "fine-tuning" skills. This study shows that in inference process, the mixed-precision quantization is a NP-hard problem and we designed a set of method to map mathematical method into pratical method. And our problem setting can be solved by branch and bound method with less computing resouces. We also show that how to set the quantization parameters in theorical method.
| change mixed precision inference quantization problem into the problem which solved by branch and bound method |
Semi-supervised learning (SSL) tackles the label missing problem by enabling the effective usage of unlabeled data. While existing SSL methods focus on the traditional setting, a practical and challenging scenario called label Missing Not At Random (MNAR) is usually ignored. In MNAR, the labeled and unlabeled data fall into different class distributions resulting in biased label imputation, which deteriorates the performance of SSL models. In this work, class transition tracking based Pseudo-Rectifying Guidance (PRG) is devised for MNAR. We explore the class-level guidance information obtained by the Markov random walk, which is modeled on a dynamically created graph built over the class tracking matrix. PRG unifies the history information of each class transition caused by the pseudo-rectifying procedure to activate the model's enthusiasm for neglected classes, so as the quality of pseudo-labels on both popular classes and rare classes in MNAR could be improved. We show the superior performance of PRG across a variety of the MNAR scenarios and the conventional SSL setting, outperforming the latest SSL solutions by a large margin. Checkpoints and evaluation code are available at the anonymous link https://anonymous.4open.science/r/PRG4SSL-MNAR-8DE2 while the source code will be available upon paper acceptance. | A simple but effective approach yielding tangible improvement in the performance of semi-supervised learning with non-random missing labels. |
Large pretrained foundation models (such as CLIP, DALL-E) are among the most recent significant advances in the AI community. Their implication is profound. This paper examines the value of these foundation models as a model knowledge base -- we aim to distill the knowledge in these foundation models for training lightweight models designed for specific tasks in practical application scenarios with improved performance. Despite abundant progress in knowledge distillation (KD) in traditional models trained under the supervision of class labels in datasets encoded as integers, distilling such text-image contrastive learning model has not been explored extensively. Meanwhile, KD is well-known for being bothered by the capacity gap problem (i.e., distilling knowledge from a teacher significantly larger than a student often degrades the performance of the student). The teacher-student capacity gap in distilling foundation models is even larger. Therefore, how to overcome this potential issue is also elusive now. This paper presents detailed analyses of these questions aiming to successfully tap into a pretrained foundation model (CLIP) to boost the student's performance. Besides the practical performance benefits, several interesting discoveries are unveiled: (1) CLIP is not bothered by the capacity gap, which may let us re-evaluate if the "capacity-gap" issue is really due to the capacity gap (2) We find the reason is largely due to that CLIP is not over-confident on the wrong labels when misclassifies input image samples. | We focus on image classification task, and investigate the capasity gap resistance of CLIP in knowledge distillation. |
Security is becoming increasingly critical in deep learning applications. Recent researches demonstrate that NN models are vulnerable to adversarial attacks, which can mislead them with only small input perturbations. Moreover, adversaries who know the architecture of victim models can conduct more effective attacks. Unfortunately, the architectural knowledge can usually be stolen by the adversaries by exploiting the system-level hints through many side channels, which is referred to as the neural architecture extraction attack. Conventional countermeasures for neural architecture extraction can introduce large overhead, and different hardware platforms have diverse types of side-channel leakages such that many expert efforts are needed in developing hardware-specific countermeasures. In this paper, we propose DeepGuiser, an automatic, hardware-agnostic, and retrain-free neural architecture disguising method, to disguise the neural architectures to reduce the harm of neural architecture extraction attacks. In a nutshell, given a trained model, DeepGuiser outputs a deploy model that is functionally equivalent with the trained model but with a different (i.e., disguising) architecture. DeepGuiser can minimize the harm of the follow-up adversarial transfer attacks to the deploy model, even if the disguising architecture is completely stolen by the architecture extraction attack. Experiments demonstrate that DeepGuiser can effectively disguise diverse architectures and impede the adversarial transferability by 13.87% ∼ 32.59%, while only introducing 10% ∼ 40% extra inference latency. | DeepGuiser is an automatic, hardware-agnostic, and retrain-free neural architecture disguising method to disguise the neural architectures, to resist possible adversarial attacks rendered by the model extraction attacks. |
Updates to Machine Learning as a Service (MLaaS) APIs may affect downstream systems that depend on their predictions. However, performance changes introduced by these updates are poorly documented by providers and seldom studied in the literature. As a result, users are left wondering: do model updates introduce subtle performance changes that could adversely affect my system? Ideally, users would have access to a detailed ChangeList specifying the slices of data where model performance has improved and degraded since the update. But, producing a ChangeList is challenging because it requires (1) discovering slices in the absence of detailed annotations or metadata, (2) accurately attributing coherent concepts to the discovered slices, and (3) communicating them to the user in a digestable manner. We introduce Mocha, an interactive framework for building, verifying and releasing ChangeLists that addresses these challenges. Using it, we perform a large-scale analysis of three real-world MLaaS API updates. We produce a ChangeList for each, identifying over 100 coherent data slices on which the model’s performance changed significantly. Notably, we find 63 instances where an update improves performance globally, but hurts performance on a coherent slice – a phenomenon not previously documented at scale in the literature. These findings underscore the importance of producing a detailed ChangeList when the model behind an API is updated. | In this work, we study MLaaS API updates. We introduce, Mocha, a new framework for describing model updates. Then we use Mocha to demonstrate how subtle, but significant, shifts are commonly introduced by updates. |
The optimal transportation map finds the most economical way to transport one probability measure to another, and it has been applied in a broad range of applications in machine learning and computer vision. By the Brenier theory, computing the optimal transport map is equivalent to solving a Monge-Amp\`ere equation, which is highly non-linear. Therefore, the computation of optimal transportation maps is intrinsically challenging.
In this work, we propose a novel and powerful method, the FFT-OT (fast Fourier transform-optimal transport), to compute the 3-dimensional OT problems. The method is based on several key ideas: first, the Monge-Amp\`ere equation is linearized to a sequence of linear elliptic PDEs with spacial and temporal variant coefficients; second, the obliqueness property of optimal transportation maps is reformulated as a Neumann boundary condition; and third, the variant coefficient elliptic PDEs are approximated by constant coefficient elliptic PDEs and solved by FFT on GPUs. We also prove that the algorithm converges linearly, namely the approximation error decreases exponentially fast. Experimental results show that the FFT-OT algorithm is more than a hundred times faster than the conventional methods based on the convex geometry. Furthermore, the method can be directly applied for sampling from complex 3D density functions in machine learning and magnifying the volumetric data in medical imaging. | Optimal transport, Monge-Amp\`ere equation, Elliptic PDE, Fast Fourier transform |
Polyak-Łojasiewicz (PL) [Polyak, 1963] condition is a weaker condition than the strong convexity but suffices to ensure a global convergence for the Gradient Descent algorithm. In this paper, we study the lower bound of algorithms using first-order oracles to find an approximate optimal solution. We show that any first-order algorithm requires at least $\Omega\left((L/\mu)^{1-\alpha} \right)$ gradient costs to find an $\epsilon$-approximate optimal solution for a general $L$-smooth function that has an $\mu$-PL constant for any $\alpha>0$. This result demonstrates the near optimality of the Gradient Descent algorithm to minimize smooth PL functions in the sense that there exists a ``hard'' PL function such that no first-order algorithm can be faster by a polynomial order. In contrast, it is well-known that the momentum technique, e.g. [Nesterov, 2003, chap. 2] can provably accelerate Gradient Descent to $O\left(\sqrt{L/\hat{\mu}}\log\frac{1}{\epsilon}\right)$ gradient costs for functions that are $L$-smooth and $\hat{\mu}$-strongly convex. Therefore, our result distinguishes the hardness of minimizing a smooth PL function and a smooth strongly convex function as the complexity of the former cannot be improved by any polynomial order in general. | We show that any first-order algorithm requires at least $\tilde{\Omega}\left((L/\mu)^{1-\alpha} \right)$ gradient costs to find an $\epsilon$-approximate optimal solution for a general $L$-smooth, $\mu$-PL function for any $\alpha>0$ . |
Modern earphones come equipped with microphones and inertial measurement units (IMU). When a user wears the earphone, the IMU can serve as a second modality for detecting speech signals. Specifically, as humans speak to their earphones (e.g., during phone calls), the throat’s vibrations propagate through the skull to ultimately induce a vibration in the IMU. The IMU data is heavily distorted
(compared to the microphone’s recordings), but IMUs offer a critical advantage — they are not interfered by ambient sounds. This presents an opportunity in multi-modal speech enhancement, i.e., can the distorted but uninterfered IMU signal enhance the user’s speech when the microphone’s signal suffers from strong ambient interference?
We combine the best of both modalities (microphone and IMU) by designing a cooperative and self-supervised network architecture that does not rely on clean speech data from the user. Instead, using only noisy speech recordings, the IMU learns to give hints on where the target speech is likely located. The microphone uses this hint to enrich the speech signal, which then trains the IMU to improve subsequent hints. This iterative approach yields promising results, comparable to a supervised denoiser trained on clean speech signals. When clean signals are also available to our architecture, we observe promising SI-SNR improvement. We believe this result can aid speech-related applications in earphones and hearing aids, and potentially generalize to others, like audio-visual denoising. | Using clean low resolution IMU data to supervise the multimodal denoiser |
The Transformer-based models have shown superior performance in the long sequence time-series forecasting problem. The sparsity assumption on self-attention dot-product reveals that not all inputs are equally significant for Transformers. Instead of implicitly utilizing weighted time-series, we build a new learning framework by cascaded teaching Transformers to reweight samples. We formulate the framework as a multi-level optimization and design three different dataset-weight generators. We perform extensive experiments on five datasets, which shows that our proposed method could significantly outperform the SOTA Transformers. | Cascaded Teaching Trasnformers |
Fine-tuning large-scale pretrained models has led to remarkable progress in well-studied modalities such as vision and NLP. However, similar gains have not been observed in many other tasks due to an assumed lack of relevant pretrained models for these diverse modalities. In this work, we revisit this assumption by studying the cross-modal transfer ability of large-scale pretrained models. We introduce ORCA, a general cross-modal fine-tuning workflow that enables fast and automatic exploitation of existing pretrained models for diverse tasks. ORCA achieves task-specific adaptation by learning feature embeddings that minimize an optimal transport distance metric to map the data distribution in the end-task modality to the pretraining modality. We test ORCA on 13 tasks with varying modalities and input-output types. ORCA performs the best on 10 of them and is in the top three on the others. We further quantify the importance of embedding distance for downstream performance, highlight ORCA’s utility for data-limited tasks, and demonstrate its compatibility with same-modality transfer. | We study how to effectively transfer pretrained models to problems outside the pretraining modalities. |
We tackle the domain generalisation (DG) problem by posing it as a domain adaptation (DA) task where we adversarially synthesise the worst-case `target' domain and adapt a model to that worst-case domain, thereby improving the model’s robustness. To synthesise data that is challenging yet semantics-preserving, we generate Fourier amplitude images and combine them with source domain phase images, exploiting the widely believed conjecture from signal processing that amplitude spectra mainly determines image style, while phase data mainly captures image semantics. To synthesise a worst-case domain for adaptation, we train the classifier and the amplitude generator adversarially. Specifically, we exploit the maximum classifier discrepancy (MCD) principle from DA that relates the target domain performance to the discrepancy of classifiers in the model hypothesis space. By Bayesian hypothesis modeling, we express the model hypothesis space effectively as a posterior distribution over classifiers given the source domains, making adversarial MCD minimisation feasible. On the DomainBed benchmark including the large-scale DomainNet dataset, the proposed approach yields significantly improved domain generalisation performance over the state-of-the-art. | We tackle the domain generalisation problem by posing it as a domain adaptation task where we adversarially synthesise the worst-case target domain via Fourier amplitude generation. |
Prompt tuning approaches, which learn task-specific soft prompts for a downstream task conditioning on frozen pre-trained models, have attracted growing interest due to its parameter efficiency. With large language models and sufficient training data, prompt tuning performs comparably to full-model tuning. However, with limited training samples in few-shot settings, prompt tuning fails to match the performance of full-model fine-tuning. In this work, we focus on improving the few-shot performance of prompt tuning by transferring knowledge from soft prompts of source tasks with abundant training samples. Recognizing the good generalization capabilities of ensemble methods in low-data regime, we first experiment and show that a simple ensemble of model predictions based on different source prompts, outperforms existing multi-prompt knowledge transfer approaches such as source prompt fusion in the few-shot setting. Motivated by this observation, we further investigate model ensembles and propose Sample-specific Ensemble of Source Models (SESoM). SESoM learns to adjust the contribution of each source model for each target sample separately when ensembling source model outputs. Through this way, SESoM inherits the superior generalization of ensemble methods and simultaneously captures the sample-specific competence of each source prompt. We conduct experiments across a diverse set of eight NLP tasks using models of different scales (T5-\{base, large, XL\}) and find that SESoM consistently outperforms the existing models of the same as well as larger parametric scale by a large margin. | For improving few-shot prompt tuning, we propose a Sample-specific Ensemble of Source Models to transfer knowledge from soft prompts trained on source tasks to target tasks by adjusting the contribution of source models for each target sample. |
The ability to rapidly generalize is crucial for reinforcement learning to be practical in real-world tasks. However, generalization is complicated by the fact that, in many settings, some state features reliably support generalization while others do not. We consider the problem of learning generalizable policies and skills (in the form of options) by identifying feature sets that generalize across instances. We propose an attention-ensemble approach, where a collection of minimally overlapping feature masks is learned, each of which individually maximizes performance on the source instance. Subsequent tasks are instantiated using the ensemble, and transfer performance is used to update the estimated probability that each feature set will generalize in the future. We show that our approach leads to fast policy generalization for eight tasks in the Procgen benchmark. We then show its use in learning portable options in Montezuma's Revenge, where it is able to generalize skills learned in the first screen to the remainder of the game. | We learn a generalizing state feature for skill transfer using an attention-based ensemble. |
A zero-shot RL agent is an agent that can solve any RL task in a given environment, instantly with no additional planning or learning, after an initial reward-free learning phase. This marks a shift from the reward-centric RL paradigm towards controllable agents that can follow arbitrary instructions in an environment. Current RL agents can solve families of related tasks at best, or require planning anew for each task. Strategies for approximate zero-shot RL have been suggested using successor features (SFs) (Borsa et al., 2018) or forward-backward (FB) representations (Touati & Ollivier, 2021), but testing has been limited.
After clarifying the relationships between these schemes, we introduce improved losses and new SF models, and test the viability of zero-shot RL schemes systematically on tasks from the Unsupervised RL benchmark (Laskin et al., 2021). To disentangle universal representation learning from exploration, we work in an offline setting and repeat the tests on several existing replay buffers.
SFs appear to suffer from the choice of the elementary state features. SFs with Laplacian eigenfunctions do well, while SFs based on auto-encoders, inverse curiosity, transition models, low-rank transition matrix, contrastive learning, or diversity (APS), perform unconsistently. In contrast, FB representations jointly learn the elementary and successor features from a single, principled criterion. They perform best and consistently across the board, reaching $85\%$ of supervised RL performance with a good replay buffer, in a zero-shot manner. | We revisit zero-shot RL based on successor representations, we introduce improved losses and new models and evaluate them systematically on the unsupervised RL benchmark. |
Few-shot incremental segmentation is the task of updating a segmentation model, as novel classes are introduced online overtime with a small number of training images. Although incremental segmentation methods exist in the literature, they tend to fall short in the few-shot regime and when given partially-annotated training images, where only the novel class is segmented. This paper proposes a data synthesizer, Guided copy-And-Paste Synthesis (GAPS), that improves the performance of few-shot incremental segmentation in a model-agnostic fashion. Despite the great success of copy-paste synthesis in the conventional offline visual recognition, we demonstrate substantially degraded performance of its naive extension in our online scenario, due to newly encountered challenges. To this end, GAPS (i) addresses the partial-annotation problem by leveraging copy-paste to generate fully-labeled data for training, (ii) helps augment the few images of novel objects by introducing a guided sampling process, and (iii) mitigates catastrophic forgetting by employing a diverse memory-replay buffer. Compared to existing state-of-the-art methods, GAPS dramatically boosts the novel IoU of baseline methods on established few-shot incremental segmentation benchmarks by up to 80%. More notably, GAPS maintains good performance in even more impoverished annotation settings, where only single instances of novel objects are annotated. | This paper proposes a guided copy-paste synthesis process for few-shot incremental semantic segmentation, which can be combined with existing methods to achieve dramatic performance increase. |
The long runtime of high-fidelity partial differential equation (PDE) solvers makes them unsuitable for time-critical applications. We propose to accelerate PDE solvers using reduced-order modeling (ROM). Whereas prior ROM approaches reduce the dimensionality of discretized vector fields, our continuous reduced-order modeling (CROM) approach builds a low-dimensional embedding of the continuous vector fields themselves, not their discretization. We represent this reduced manifold using continuously differentiable neural fields, which may train on any and all available numerical solutions of the continuous system, even when they are obtained using diverse methods or discretizations. We validate our approach on an extensive range of PDEs with training data from voxel grids, meshes, and point clouds. Compared to prior discretization-dependent ROM methods, such as linear subspace proper orthogonal decomposition (POD) and nonlinear manifold neural-network-based autoencoders, CROM features higher accuracy, lower memory consumption, dynamically adaptive resolutions, and applicability to any discretization. For equal latent space dimension, CROM exhibits 79$\times$ and 49$\times$ better accuracy, and 39$\times$ and 132$\times$ smaller memory footprint, than POD and autoencoder methods, respectively. Experiments demonstrate 109$\times$ and 89$\times$ wall-clock speedups over unreduced models on CPUs and GPUs, respectively. Videos and codes are available on the project page: https://crom-pde.github.io | We accelerate PDE solvers via rapid latent space traversal of continuous vector fields leveraging implicit neural representations. |
Variational autoencoders (VAEs) are powerful generative modelling methods, however they suffer from blurry generated samples and reconstructions compared to the images they have been trained on. Significant research effort has been spent to increase the generative capabilities by creating more flexible models but often flexibility comes at the cost of higher complexity and computational cost. Several works have focused on altering the reconstruction term of the evidence lower bound (ELBO), however, often at the expense of losing the mathematical link to maximizing the likelihood of the samples under the modeled distribution. Here we propose a new formulation of the reconstruction term for the VAE that specifically penalizes the generation of blurry images while at the same time still maximizing the ELBO under the modeled distribution.
We show the potential of the proposed loss on three different data sets, where it outperforms several recently proposed reconstruction losses for VAEs. | We propose a new reconstruction term for VAEs that explicitly focuses on minimizing the blur of generated/reconstructed images while still optimizing the ELBO. |
This work studies an algorithm, which we call magnetic mirror descent, that is inspired by mirror descent and the non-Euclidean proximal gradient algorithm. Our contribution is demonstrating the virtues of magnetic mirror descent as both an equilibrium solver and as an approach to reinforcement learning in two-player zero-sum games. These virtues include: 1) Being the first quantal response equilibria solver to achieve linear convergence for extensive-form games with first order feedback; 2) Being the first standard reinforcement learning algorithm to achieve empirically competitive results with CFR in tabular settings; 3) Achieving favorable performance in 3x3 Dark Hex and Phantom Tic-Tac-Toe as a self-play deep reinforcement learning algorithm. | A single algorithm for both single-agent reinforcement learning and approximating quantal response and Nash equilibria in two-player zero-sum games. |
Graph-structured physical mechanisms are ubiquitous in real-world scenarios, thus revealing underneath formulas is of great importance for scientific discovery. However, classical symbolic regression methods fail on this task since they can only handle input-output pairs that are not graph-structured. In this paper, we propose a new approach that generalizes symbolic regression to graph-structured physical mechanisms. The essence of our method is to model the formula skeleton with a message-passing flow, which helps transform the discovery of the skeleton into the search for the message-passing flow. Such a transformation guarantees that we are able to search a message-passing flow, which is efficient and Pareto-optimal in terms of both accuracy and simplicity. Subsequently, the underneath formulas can be identified by interpreting component functions of the searched message-passing flow, reusing classical symbolic regression methods. We conduct extensive experiments on datasets from different physical domains, including mechanics, electricity, and thermology, and on real-world datasets of pedestrian dynamics without ground-truth formulas. The experimental results not only verify the rationale of our design but also demonstrate that the proposed method can automatically learn precise and interpretable formulas for graph-structured physical mechanisms. | We generalize symbolic regression to graph data and propose a novel method with the key insight of learning formula skeleton by searching message-passing flows of graph neural networks. |