type
stringclasses
1 value
name
stringlengths
14
183
virtualsite_url
stringlengths
46
46
speakers/authors
stringlengths
8
1.31k
abstract
stringlengths
246
3.59k
Poster
When Does Curriculum Learning Help? A Theoretical Perspective
https://neurips.cc//virtual/2025/poster/117698
Kaibo Zhang, Yunjuan Wang, Raman Arora
Curriculum learning has emerged as an effective strategy to enhance the training efficiency and generalization of machine learning models. However, its theoretical underpinnings remain relatively underexplored. In this work, we develop a theoretical framework for curriculum learning based on biased regularized empirica...
Poster
When Do Transformers Outperform Feedforward and Recurrent Networks? A Statistical Perspective
https://neurips.cc//virtual/2025/poster/118434
Alireza Mousavi-Hosseini, Clayton Sanford, Denny Wu, Murat Erdogdu
Theoretical efforts to prove advantages of Transformers in comparison with classical architectures such as feedforward and recurrent neural networks have mostly focused on representational power. In this work, we take an alternative perspective and prove that even with infinite compute, feedforward and recurrent networ...
Poster
When Less Language is More: Language-Reasoning Disentanglement Makes LLMs Better Multilingual Reasoners
https://neurips.cc//virtual/2025/poster/116811
Weixiang Zhao, Jiahe Guo, Yang Deng, Tongtong Wu, Wenxuan Zhang, Yulin Hu, Xingyu Sui, Yanyan Zhao, Wanxiang Che, Bing Qin, Tat-Seng Chua, Ting Liu
Multilingual reasoning remains a significant challenge for large language models (LLMs), with performance disproportionately favoring high-resource languages. Drawing inspiration from cognitive neuroscience, which suggests that human reasoning functions largely independently of language processing, we hypothesize that ...
Poster
When Lower-Order Terms Dominate: Adaptive Expert Algorithms for Heavy-Tailed Losses
https://neurips.cc//virtual/2025/poster/116542
Antoine Moulin, Emmanuel Esposito, Dirk van der Hoeven
We consider the problem setting of prediction with expert advice with possibly heavy-tailed losses, i.e.\ the only assumption on the losses is an upper bound on their second moments, denoted by $\theta$. We develop adaptive algorithms that do not require any prior knowledge about the range or the second moment of the l...
Poster
When majority rules, minority loses: bias amplification of gradient descent
https://neurips.cc//virtual/2025/poster/118044
François Bachoc, Jerome Bolte, Ryan Boustany, Loubes Jean-Michel
Despite growing empirical evidence of bias amplification in machine learning, its theoretical foundations remain poorly understood. We develop a formal framework for majority-minority learning tasks, showing how standard training can favor majority groups and produce stereotypical predictors that neglect minority-speci...
Poster
When Models Don’t Collapse: On the Consistency of Iterative MLE
https://neurips.cc//virtual/2025/poster/119185
Daniel Barzilai, Ohad Shamir
The widespread use of generative models has created a feedback loop in which each generation of models is trained on data partially produced by its predecessors. This process has raised concerns about model collapse: A critical degradation in performance caused by repeated training on synthetic data. However, different...
Poster
When Models Know More Than They Can Explain: Quantifying Knowledge Transfer in Human-AI Collaboration
https://neurips.cc//virtual/2025/poster/119536
Quan Shi, Carlos Jimenez, Shunyu Yao, Nick Haber, Diyi Yang, Karthik Narasimhan
As large language models (LLMs) increasingly serve as close collaborators for humans, it is crucial that they express their reasoning in ways that humans can understand and learn from. However, this capability remains relatively less understood and under-evaluated. To address this, we introduce a conceptual framework f...
Poster
When No Paths Lead to Rome: Benchmarking Systematic Neural Relational Reasoning
https://neurips.cc//virtual/2025/poster/121739
Anirban Das, Muhammad Irtaza Khalid, Rafael Peñaloza, Steven Schockaert
Designing models that can learn to reason in a systematic way is an important and long-standing challenge. In recent years, a wide range of solutions have been proposed for the specific case of systematic relational reasoning, including Neuro-Symbolic approaches, variants of the Transformer architecture, and specialize...
Poster
When One Moment Isn't Enough: Multi-Moment Retrieval with Cross-Moment Interactions
https://neurips.cc//virtual/2025/poster/118798
Zhuo Cao, Heming Du, Bingqing Zhang, Xin Yu, Xue Li, Sen Wang
Existing Moment retrieval (MR) methods focus on Single-Moment Retrieval (SMR). However, one query can correspond to multiple relevant moments in real-world applications. This makes the existing methods insufficient for video temporal grounding. By revisiting the gap between current MR tasks and real-world applications,...
Poster
When Semantics Mislead Vision: Mitigating Large Multimodal Models Hallucinations in Scene Text Spotting and Understanding
https://neurips.cc//virtual/2025/poster/119366
Yan Shu, Hangui Lin, Yexin Liu, Yan Zhang, Gangyan Zeng, Yan Li, Yu Zhou, Ser Nam Lim, Harry Yang, Nicu Sebe
Large Multimodal Models (LMMs) have achieved impressive progress in visual perception and reasoning. However, when confronted with visually ambiguous or non-semantic scene text, they often struggle to accurately spot and understand the content, frequently generating semantically plausible yet visually incorrect answers...
Poster
When Thinking Drifts: Evidential Grounding for Robust Video Reasoning
https://neurips.cc//virtual/2025/poster/115892
Romy Luo, Zihui (Sherry) Xue, Alex Dimakis, Kristen Grauman
Video reasoning, the task of enabling machines to infer from dynamic visual content through multi-step logic, is crucial for advanced AI. While the Chain-of-Thought (CoT) mechanism has enhanced reasoning in text-based tasks, its application to video understanding remains underexplored. This paper presents a systematic ...
Poster
When Thinking Fails: The Pitfalls of Reasoning for Instruction-Following in LLMs
https://neurips.cc//virtual/2025/poster/115354
Xiaomin Li, Zhou Yu, Zhiwei Zhang, Xupeng Chen, Ziji Zhang, Yingying Zhuang, Narayanan Sadagopan, Anurag Beniwal
Reasoning-enhanced large language models (RLLMs), whether explicitly trained for reasoning or prompted via chain-of-thought (CoT), have achieved state-of-the-art performance on many complex reasoning tasks. However, we uncover a surprising and previously overlooked phenomenon: explicit CoT reasoning can significantly d...
Poster
Where and How to Perturb: On the Design of Perturbation Guidance in Diffusion and Flow Models
https://neurips.cc//virtual/2025/poster/116229
Donghoon Ahn, Jiwon Kang, Sanghyun Lee, Minjae Kim, Wooseok Jang, Jaewon Min, Sangwu Lee, Sayak Paul, Seungryong Kim
Recent guidance methods steer diffusion sampling by perturbing the model to construct an implicit bad model and guide generation away from it. Among them, attention perturbation has demonstrated strong empirical performance, yet existing approaches remain fragmented, necessitating a unified framework. We propose a gene...
Poster
Where Does It Exist from the Low-Altitude: Spatial Aerial Video Grounding
https://neurips.cc//virtual/2025/poster/116334
Yang Zhan, Yuan Yuan
The task of localizing an object's spatial tube based on language instructions and video, known as spatial video grounding (SVG), has attracted widespread interest. Existing SVG tasks have focused on ego-centric fixed front perspective and simple scenes, which only involved a very limited view and environment. However,...
Poster
Where Graph Meets Multi-View Learning : Collarborative Graph Mixture of Experts
https://neurips.cc//virtual/2025/poster/116976
Zhihao Wu, Jinyu Cai, Yunhe Zhang, Jielong Lu, Zhaoliang Chen, Shuman Zhuang, Haishuai Wang
The convergence of multi-view learning and graph learning has propelled the emergence of multi-view graph neural networks (MvGNNs), offering unprecedented capabilities to address complex real-world data characterized by heterogeneous yet interconnected information. While existing MvGNNs exploit the potential of multi-v...
Poster
Which Algorithms Have Tight Generalization Bounds?
https://neurips.cc//virtual/2025/poster/115861
Michael Gastpar, Ido Nachum, Jonathan Shafer, Thomas Weinberger
We study which machine learning algorithms have tight generalization bounds with respect to a given collection of population distributions. Our results build on and extend the recent work of Gastpar et al. (2023). First, we present conditions that preclude the existence of tight generalization bounds. Specifically, we ...
Poster
Which Data Attributes Stimulate Math and Code Reasoning? An Investigation via Influence Functions
https://neurips.cc//virtual/2025/poster/117204
Siqi Kou, Qingyuan Tian, Hanwen Xu, Zihao Zeng, Zhijie Deng
Large language models (LLMs) have demonstrated remarkable reasoning capabilities in math and coding, often bolstered by post-training on the chain-of-thoughts (CoTs) generated by stronger models. However, existing strategies for curating such training data predominantly rely on heuristics, limiting generalizability and...
Poster
Whitened Score Diffusion: A Structured Prior for Imaging Inverse Problems
https://neurips.cc//virtual/2025/poster/119040
Jeffrey Alido, Tongyu Li, Yu Sun, Lei Tian
Conventional score-based diffusion models (DMs) may struggle with anisotropic Gaussian diffusion processes due to the required inversion of covariance matrices in the denoising score matching training objective \cite{vincent_connection_2011}. We propose Whitened Score (WS) diffusion models, a novel SDE-based framework ...
Poster
Whole-Body-Conditioned Ego-Centric Video Prediction
https://neurips.cc//virtual/2025/poster/117535
Yutong Bai, Danny Tran, Amir Bar, Trevor Darrell, Yann LeCun, Jitendra Malik
We train models to predict ego-centric video from human actions (PEVA), given the past video and an action represented by the relative 3D body pose. By conditioning on kinematic pose trajectories, structured by the joint hierarchy of the body, our model learns to simulate how physical human actions shape the environmen...
Poster
Who Reasons in the Large Language Models?
https://neurips.cc//virtual/2025/poster/117526
Jie Shao, Jianxin Wu
Despite the impressive performance of large language models (LLMs), the process of endowing them with new capabilities---such as mathematical reasoning---remains largely empirical and opaque. A critical open question is whether reasoning abilities stem from the entire model, specific modules, or are merely artifacts of...
Poster
Whose Instructions Count? Resolving Preference Bias in Instruction Fine-Tuning
https://neurips.cc//virtual/2025/poster/117781
Jiayu Zhang, Changbang Li, Yinan Peng, Weihao Luo, Peilai Yu, Xuan Zhang
Instruction fine-tuning (IFT) has emerged as a ubiquitous strategy for specializing large language models (LLMs), yet it implicitly assumes a single, coherent "ground-truth" preference behind all human-written instructions. In practice, annotators differ in the styles, emphases, and granularities they prefer, introduci...
Poster
Whose View of Safety? A Deep DIVE Dataset for Pluralistic Alignment of Text-to-Image Models
https://neurips.cc//virtual/2025/poster/121857
Charvi Rastogi, Tian Huey Teh, Pushkar Mishra, Roma Patel, Ding Wang, Mark Díaz, Alicia Parrish, Aida Mostafazadeh Davani, Zoe Ashwood, Michela Paganini, Vinodkumar Prabhakaran, Verena Rieser, Lora Aroyo
Current text-to-image (T2I) models often fail to account for diverse human experiences, leading to misaligned systems. We advocate for pluralistic alignment, where an AI understands and is steerable towards diverse, and often conflicting, human values. Our work provides three core contributions to achieve this in T2I m...
Poster
Who Speaks for the Trigger? Dynamic Expert Routing in Backdoored Mixture-of-Experts Transformers
https://neurips.cc//virtual/2025/poster/115307
Xin Zhao, Xiaojun Chen, Bingshan Liu, Haoyu Gao, Zhendong Zhao, Yilong Chen
Large language models (LLMs) with Mixture-of-Experts (MoE) architectures achieve impressive performance and efficiency by dynamically routing inputs to specialized subnetworks, known as experts. However, this sparse routing mechanism inherently exhibits task preferences due to expert specialization, introducing a new ...
Poster
Who You Are Matters: Bridging Interests and Social Roles via LLM-Enhanced Logic Recommendation
https://neurips.cc//virtual/2025/poster/120358
Qing Yu, Xiaobei Wang, Shuchang Liu, yandong.bai, Xiaoyu Yang, Xueliang Wang, Chang Meng, Shanshan Wu, HailanYang, Bin Wen, Huihui Xiao, Xiang Li, Fan Yang, Xiaoqiang Feng, Lantao Hu, Han Li, Kun Gai, Lixin Zou
Recommender systems filter contents/items valuable to users by inferring preferences from user features and historical behaviors. Mainstream approaches follow the learning-to-rank paradigm, which focus on discovering and modeling item topics (e.g.,categories), and capturing user preferences on these topics based on his...
Poster
Why 1 + 1 < 1 in Visual Token Pruning: Beyond Naive Integration via Multi-Objective Balanced Covering
https://neurips.cc//virtual/2025/poster/119476
Yangfu Li, Hongjian Zhan, Tianyi Chen, Qi Liu, Yu-Jie Xiong, Yue Lu
Existing visual token pruning methods target prompt alignment and visual preservation with static strategies, overlooking the varying relative importance of these objectives across tasks, which leads to inconsistent performance. To address this, we derive the first closed-form error bound for visual token pruning based...
Poster
Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associations
https://neurips.cc//virtual/2025/poster/118673
Yiyou Sun, Yu Gai, Lijie Chen, Abhilasha Ravichander, Yejin Choi, Nouha Dziri, Dawn Song
Large language models (LLMs) frequently generate hallucinations—content that deviates from factually inaccurate or deviates from provided context—posing challenges for diagnosis. However, diagnosing the causes of hallucination is challenging due to the complex interplay of underlying causes. This paper introduces a fra...
Poster
Why Diffusion Models Don’t Memorize: The Role of Implicit Dynamical Regularization in Training
https://neurips.cc//virtual/2025/poster/119372
Tony Bonnaire, Raphaël Urfin, Giulio Biroli, Marc Mezard
Diffusion models have achieved remarkable success across a wide range of generative tasks. A key challenge is understanding the mechanisms that prevent their memorization of training data and allow generalization. In this work, we investigate the role of the training dynamics in the transition from generalization to me...
Poster
Why Do Multi-Agent LLM Systems Fail?
https://neurips.cc//virtual/2025/poster/121528
Mert Cemri, Melissa Z Pan, Shuyi Yang, Lakshya A Agrawal, Bhavya Chopra, Rishabh Tiwari, Kurt Keutzer, Aditya Parameswaran, Dan Klein, Kannan Ramchandran, Matei A Zaharia, Joseph Gonzalez, Ion Stoica
Despite enthusiasm for Multi-Agent LLM Systems (MAS), their performance gains on popular benchmarks are often minimal. This gap highlights a critical need for a principled understanding of why MAS fail. Addressing this question requires systematic identification and analysis of failure patterns. We introduce MAD, a com...
Poster
Why Do Some Language Models Fake Alignment While Others Don't?
https://neurips.cc//virtual/2025/poster/120242
Abhay Sheshadri, John Hughes, Julian Michael, Alex Mallen, Arun Jose, Fabien Roger
*Alignment faking in large language models* presented a demonstration of Claude 3 Opus and Claude 3.5 Sonnet selectively complying with a helpful-only training objective to prevent modification of their behavior outside of training. We expand this analysis to 23 models and find that only 5 (Claude 3 Opus, Claude 3.5 So...
Poster
Why Knowledge Distillation Works in Generative Models: A Minimal Working Explanation
https://neurips.cc//virtual/2025/poster/116625
Sungmin Cha, Kyunghyun Cho
Knowledge distillation (KD) is a core component in the training and deployment of modern generative models, particularly large language models (LLMs). While its empirical benefits are well documented—enabling smaller student models to emulate the performance of much larger teachers—the underlying mechanisms by which KD...
Poster
Why Masking Diffusion Works: Condition on the Jump Schedule for Improved Discrete Diffusion
https://neurips.cc//virtual/2025/poster/115376
Alan Amin, Nate Gruver, Andrew Wilson
Discrete diffusion models, like continuous diffusion models, generate high-quality samples by gradually undoing noise applied to datapoints with a Markov process. Gradual generation in theory comes with many conceptual benefits; for example, inductive biases can be incorporated into the noising Markov process. In pract...
Poster
Why Playing Against Diverse and Challenging Opponents Speeds Up Coevolution: A Theoretical Analysis on Combinatorial Games
https://neurips.cc//virtual/2025/poster/115319
Alistair Benford, Per Kristian Lehre
Competitive coevolutionary algorithms (CoEAs) have a natural application to problems that are adversarial or feature strategic interaction. However, there is currently limited theoretical insight into how to avoid pathological behaviour associated to CoEAs. In this paper we use impartial combinatorial games as a challe...
Poster
Why Popular MOEAs are Popular: Proven Advantages in Approximating the Pareto Front
https://neurips.cc//virtual/2025/poster/119624
Mingfeng Li, Qiang Zhang, Weijie Zheng, Benjamin Doerr
Recent breakthroughs in the analysis of multi-objective evolutionary algorithms (MOEAs) are mathematical runtime analyses of those algorithms which are intensively used in practice. So far, most of these results show the same performance as previously known for simple algorithms like the GSEMO. The few results indicati...
Poster
Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree Search
https://neurips.cc//virtual/2025/poster/116491
Yuichi Inoue, Kou Misaki, Yuki Imajuku, So Kuroki, Taishi Nakamura, Takuya Akiba
Recent advances demonstrate that increasing inference-time computation can significantly boost the reasoning capabilities of large language models (LLMs). Although repeated sampling (i.e., generating multiple candidate outputs) is a highly effective strategy, it does not leverage external feedback signals for refinemen...
Poster
WildCAT: Appearance-Aware Multi-View Diffusion in the Wild
https://neurips.cc//virtual/2025/poster/118317
Morris Alper, David Novotny, Filippos Kokkinos, Hadar Averbuch-Elor, Tom Monnier
Despite recent advances in sparse novel view synthesis (NVS) applied to object-centric scenes, scene-level NVS remains a challenge. A central issue is the lack of available clean multi-view training data, beyond manually curated datasets with limited diversity, camera variation, or licensing issues. On the other hand, ...
Poster
Win Fast or Lose Slow: Balancing Speed and Accuracy in Latency-Sensitive Decisions of LLMs
https://neurips.cc//virtual/2025/poster/119027
Hao Kang, Qingru Zhang, Han Cai, Weiyuan Xu, Tushar Krishna, Yilun Du, Tsachy Weissman
Large language models (LLMs) have shown remarkable performance across diverse reasoning and generation tasks, and are increasingly deployed as agents in dynamic environments such as code generation and recommendation systems. However, many real-world applications, such as high-frequency trading and real-time competitiv...
Poster
WISA: World simulator assistant for physics-aware text-to-video generation
https://neurips.cc//virtual/2025/poster/119925
Jing Wang, Ao Ma, Ke Cao, Jun Zheng, Jiasong Feng, Zhanjie Zhang, Wanyuan Pang, Xiaodan Liang
Recent advances in text-to-video (T2V) generation, exemplified by models such as Sora and Kling, have demonstrated strong potential for constructing world simulators. However, existing T2V models still struggle to understand abstract physical principles and to generate videos that faithfully obey physical laws. This li...
Poster
Wisdom is Knowing What not to Say: Hallucination-Free LLMs Unlearning via Attention Shifting
https://neurips.cc//virtual/2025/poster/118563
Chenchen Tan, Youyang Qu, Xinghao Li, Hui Zhang, Shujie Cui, Cunjian Chen, Longxiang Gao
The increase in computing power and the necessity of AI-assisted decision-making boost the growing application of large language models (LLMs). Along with this, the potential retention of sensitive data of LLMs has spurred increasing research into machine unlearning. However, existing unlearning approaches face a criti...
Poster
With Limited Data for Multimodal Alignment, Let the STRUCTURE Guide You
https://neurips.cc//virtual/2025/poster/118769
Fabian Gröger, Shuo Wen, Huyen Le, Maria Brbic
Multimodal models have demonstrated powerful capabilities in complex tasks requiring multimodal alignment including zero-shot classification and cross-modal retrieval. However, existing models typically rely on millions of paired multimodal samples, which are prohibitively expensive or infeasible to obtain in many doma...
Poster
WKV-sharing embraced random shuffle RWKV high-order modeling for pan-sharpening
https://neurips.cc//virtual/2025/poster/116699
man zhou, Xuanhua He, Danfeng Hong, Bo Huang
Pan-sharpening aims to generate a spatially and spectrally enriched multi-spectral image by integrating complementarycross-modality information from low-resolution multi-spectral image and texture-rich panchromatic counterpart. In this work, we propose aWKV-sharing embraced random shuffle RWKV high-order modeling parad...
Poster
WMCopier: Forging Invisible Watermarks on Arbitrary Images
https://neurips.cc//virtual/2025/poster/117840
Ziping Dong, Chao Shuai, Zhongjie Ba, Peng Cheng, Zhan Qin, Qinglong Wang, Kui Ren
Invisible Image Watermarking is crucial for ensuring content provenance and accountability in generative AI. While Gen-AI providers are increasingly integrating invisible watermarking systems, the robustness of these schemes against forgery attacks remains poorly characterized. This is critical, as forging traceable wa...
Poster
WolBanking77: Wolof Banking Speech Intent Classification Dataset
https://neurips.cc//virtual/2025/poster/121817
Abdou Karim KANDJI, Frederic Precioso, Cheikh BA, Augustin NDIONE, Samba NDIAYE
Intent classification models have made a lot of progress in recent years. However, previous studies primarily focus on high-resource languages datasets, which results in a gap for low-resource languages and for regions with a high rate of illiterate people where languages are more spoken than read or written. This is t...
Poster
Wonder Wins Ways: Curiosity-Driven Exploration through Multi-Agent Contextual Calibration
https://neurips.cc//virtual/2025/poster/120209
Yiyuan Pan, Zhe Liu, Hesheng Wang
Unlocking autonomous exploration in complex Multi-Agent Reinforcement Learning (MARL) systems, especially under sparse rewards, hinges on endowing agents with effective motivation. While artificial curiosity offers a powerful self-supervised drive, its efficacy is undermined when agents cannot robustly discern meaningf...
Poster
Word-Level Emotional Expression Control in Zero-Shot Text-to-Speech Synthesis
https://neurips.cc//virtual/2025/poster/117924
Tianrui Wang, Haoyu Wang, Meng Ge, Cheng Gong, Chunyu Qiang, Ziyang Ma, Zikang Huang, Guanrou Yang, Xiaobao Wang, Eng-Siong Chng, Xie Chen, Longbiao Wang, Jianwu Dang
While emotional text-to-speech (TTS) has made significant progress, most existing research remains limited to utterance-level emotional expression and fails to support word-level control. Achieving word-level expressive control poses fundamental challenges, primarily due to the complexity of modeling multi-emotion tran...
Poster
Words That Unite The World: A Unified Framework for Deciphering Central Bank Communications Globally
https://neurips.cc//virtual/2025/poster/121548
Agam Shah, Siddhant Sukhani, Huzaifa Pardawala, Saketh Budideti, Riya Bhadani, Rudra Gopal, Siddhartha Somani, Michael Galarnyk, Soungmin Lee, Arnav Hiray, Akshar Ravichandran, Eric Kim, Pranav Aluru, Joshua Zhang, Sebastian Jaskowski, Veer Guda, Meghaj Tarte, Liqin Ye, Spencer Gosden, Rutwik Routu, Rachel Yuh, Sloka C...
Central banks around the world play a crucial role in maintaining economic stability. Deciphering policy implications in their communications is essential, especially as misinterpretations can disproportionately impact vulnerable populations. To address this, we introduce the World Central Banks (WCB) dataset, the most...
Poster
World-aware Planning Narratives Enhance Large Vision-Language Model Planner
https://neurips.cc//virtual/2025/poster/116820
Junhao Shi, Zhaoye Fei, Siyin Wang, Qipeng Guo, Jingjing Gong, Xipeng Qiu
Large Vision-Language Models (LVLMs) show promise for embodied planning tasks but struggle with complex scenarios involving unfamiliar environments and multi-step goals. Current approaches rely on environment-agnostic imitation learning that disconnects instructions from environmental contexts, causing models to strugg...
Poster
WorldMem: Long-term Consistent World Simulation with Memory
https://neurips.cc//virtual/2025/poster/117127
Zeqi Xiao, Yushi LAN, Yifan Zhou, Wenqi Ouyang, Shuai Yang, Yanhong Zeng, Xingang Pan
World simulation has gained increasing popularity due to its ability to model virtual environments and predict the consequences of actions. However, the limited temporal context window often leads to failures in maintaining long-term consistency, particularly in preserving 3D spatial consistency. In this work, we prese...
Poster
WorldModelBench: Judging Video Generation Models As World Models
https://neurips.cc//virtual/2025/poster/121570
Dacheng Li, Yunhao Fang, Yukang Chen, Shuo Yang, Shiyi Cao, Justin Wong, Michael Luo, Xiaolong Wang, Hongxu Yin, Joseph Gonzalez, Ion Stoica, Song Han, Yao Lu
Video generation models have rapidly progressed, positioning themselves as video world models capable of supporting decision-making applications like robotics and autonomous driving. However, current benchmarks fail to rigorously evaluate these claims, focusing only on general video quality, ignoring important factors ...
Poster
World Models as Reference Trajectories for Rapid Motor Adaptation
https://neurips.cc//virtual/2025/poster/115217
Carlos Stein Brito, Daniel McNamee
Deploying learned control policies in real-world environments poses a fundamental challenge. When system dynamics change unexpectedly, performance degrades until models are retrained on new data. We introduce Reflexive World Models (RWM), a dual control framework that uses world model predictions as implicit reference ...
Poster
WorldWeaver: Generating Long-Horizon Video Worlds via Rich Perception
https://neurips.cc//virtual/2025/poster/115883
Zhiheng Liu, Xueqing Deng, Shoufa Chen, Angtian Wang, Qiushan Guo, Mingfei Han, Zeyue Xue, Mengzhao Chen, Ping Luo, Linjie Yang
Generative video modeling has made significant strides, yet ensuring structural and temporal consistency over long sequences remains a challenge. Current methods predominantly rely on RGB signals, leading to accumulated errors in object structure and motion over extended durations. To address these issues, we introduce...
Poster
Worse than Zero-shot? A Fact-Checking Dataset for Evaluating the Robustness of RAG Against Misleading Retrievals
https://neurips.cc//virtual/2025/poster/121659
Linda Zeng, Rithwik Gupta, Divij Motwani, Diji Yang, Yi Zhang
Retrieval-augmented generation (RAG) has shown impressive capabilities in mitigating hallucinations in large language models (LLMs). However, LLMs struggle to maintain consistent reasoning when exposed to misleading or conflicting evidence, especially in real-world domains such as politics, where information is polariz...
Poster
WritingBench: A Comprehensive Benchmark for Generative Writing
https://neurips.cc//virtual/2025/poster/121666
Yuning Wu, Jiahao Mei, Ming Yan, Chenliang Li, Shaopeng Lai, Yuran Ren, Zijia Wang, Ji Zhang, Mengyue Wu, Qin Jin, Fei Huang
Recent advancements in large language models (LLMs) have significantly enhanced text generation capabilities, yet evaluating their performance in generative writing remains a challenge. Existing benchmarks primarily focus on generic text generation or limited in writing tasks, failing to capture the diverse requirement...
Poster
Wukong's 72 Transformations: High-fidelity 3D Morphing via Flow Models
https://neurips.cc//virtual/2025/poster/115638
Minghao Yin, Yukang Cao, Kai Han
We present WUKONG, a novel training-free framework for high-fidelity textured 3D morphing that takes a pair of source and target prompts (text or images) as input. Unlike conventional methods -- which rely on manual correspondence matching and deformation trajectory estimation (limiting generalization and requiring cos...
Poster
X-Field: A Physically Grounded Representation for 3D X-ray Reconstruction
https://neurips.cc//virtual/2025/poster/117962
Feiran Wang, Jiachen Tao, Junyi Wu, Haoxuan Wang, Bin Duan, Kai Wang, Zongxin Yang, Yan Yan
X-ray imaging is indispensable in medical diagnostics, yet its use is tightly regulated due to radiation exposure. Recent research borrows representations from the 3D reconstruction area to complete two tasks with reduced radiation dose: X-ray Novel View Synthesis (NVS) and Computed Tomography (CT) reconstruction. Howe...
Poster
XIFBench: Evaluating Large Language Models on Multilingual Instruction Following
https://neurips.cc//virtual/2025/poster/121440
Zhenyu Li, Kehai Chen, Yunfei Long, Xuefeng Bai, Yaoyin Zhang, Xuchen Wei, Juntao Li, Min Zhang
Large Language Models (LLMs) have demonstrated remarkable instruction-following capabilities across various applications. However, their performance in multilingual settings lacks systematic investigation, with existing evaluations lacking fine-grained constraint analysis across diverse linguistic contexts. We introduc...
Poster
xLSTM-Mixer: Multivariate Time Series Forecasting by Mixing via Scalar Memories
https://neurips.cc//virtual/2025/poster/118697
Maurice Kraus, Felix Divo, Devendra Singh Dhami, Kristian Kersting
Time series data is prevalent across numerous fields, necessitating the development of robust and accurate forecasting models. Capturing patterns both within and between temporal and multivariate components is crucial for reliable predictions. We introduce xLSTM-Mixer, a model designed to effectively integrate tempora...
Poster
X-Mahalanobis: Transformer Feature Mixing for Reliable OOD Detection
https://neurips.cc//virtual/2025/poster/116875
Bolin Wang, Tong Wei, Jiang-Xin Shi, Yu-Feng Li, Min-Ling Zhang
Recognizing out-of-distribution (OOD) samples is essential for deploying robust machine learning systems in the open-world environments. Conventional OOD detection approaches rely on feature representations from the final layer of neuron networks, often neglecting the rich information encapsulated in shallow layers. Le...
Poster
X-Scene: Large-Scale Driving Scene Generation with High Fidelity and Flexible Controllability
https://neurips.cc//virtual/2025/poster/118117
Yu Yang, Alan Liang, Jianbiao Mei, Yukai Ma, Yong Liu, Gim Hee Lee
Diffusion models are advancing autonomous driving by enabling realistic data synthesis, predictive end-to-end planning, and closed-loop simulation, with a primary focus on temporally consistent generation. However, the generation of large-scale 3D scenes that require spatial coherence remains underexplored. In this pap...
Poster
X-Smart-Kitchen-30: Densely annotated cooking dataset with 3D kinematics to challenge video and language models
https://neurips.cc//virtual/2025/poster/121758
Andy Bonnetto, Haozhe Qi, Franklin Leong, Matea Tashkovska, Mahdi Rad, Solaiman Shokur, Friedhelm C. Hummel, Silvestro Micera, Marc Pollefeys, Alexander Mathis
Understanding behavior requires datasets that capture humans while carrying out complex tasks. The kitchen is an excellent environment for assessing human motor and cognitive function, as many complex actions are naturally exhibited in kitchens from chopping to cleaning. Here, we introduce the X-Smart-Kitchen-30 datase...
Poster
XVerse: Consistent Multi-Subject Control of Identity and Semantic Attributes via DiT Modulation
https://neurips.cc//virtual/2025/poster/119968
Bowen Chen, Brynn zhao, Haomiao Sun, Li Chen, Xu Wang, Daniel Du, Xinglong Wu
Achieving fine-grained control over subject identity and semantic attributes (pose, style, lighting) in text-to-image generation, particularly for multiple subjects, often undermines the editability and coherence of Diffusion Transformers (DiTs). Many approaches introduce artifacts or suffer from attribute entanglement...
Poster
YEAST: Yet Another Sequential Test
https://neurips.cc//virtual/2025/poster/117238
Alexey Kurennoy, Majed Dodin, Tural Gurbanov, Ana Peleteiro Ramallo
Online evaluation of machine learning models is typically conducted through randomised experiments. Sequential statistical tests are valuable tools for analysing these experiments, as they enable researchers to stop data collection early without increasing the risk of false discoveries. However, existing sequential tes...
Poster
Yggdrasil: Bridging Dynamic Speculation and Static Runtime for Latency-Optimal Tree-Based LLM Decoding
https://neurips.cc//virtual/2025/poster/119964
Yue Guan, Changming Yu, Shihan Fang, Weiming Hu, Zaifeng Pan, Zheng Wang, Zihan Liu, Yangjie Zhou, Yufei Ding, Minyi Guo, Jingwen Leng
Speculative decoding improves LLM inference by generating and verifying multiple tokens in parallel, but existing systems suffer from suboptimal performance due to a mismatch between dynamic speculation and static runtime assumptions. We present Yggdrasil, a co-designed system that enables latency-optimal speculative d...
Poster
YOLOv12: Attention-Centric Real-Time Object Detectors
https://neurips.cc//virtual/2025/poster/116765
Yunjie Tian, Qixiang Ye, DAVID DOERMANN
Enhancing the network architecture of the YOLO framework has been crucial for a long time. Still, it has focused on CNN-based improvements despite the proven superiority of attention mechanisms in modeling capabilities. This is because attention-based models cannot match the speed of CNN-based models. This paper propos...
Poster
You Can Trust Your Clustering Model: A Parameter-free Self-Boosting Plug-in for Deep Clustering
https://neurips.cc//virtual/2025/poster/118692
Hanyang Li, Yuheng Jia, Hui LIU, Junhui Hou
Recent deep clustering models have produced impressive clustering performance. However, a common issue with existing methods is the disparity between global and local feature structures. While local structures typically show strong consistency and compactness within class samples, global features often present intertw...
Poster
You Have to Be Realistic: On Investigating Feature Emergence in Deep Learning-based Side-channel Analysis
https://neurips.cc//virtual/2025/poster/118946
Sengim Karayalcin, Marina Krček, Stjepan Picek
Side-channel analysis (SCA) poses a real-world threat by exploiting unintentional physical signals to extract secret information from secure devices. Evaluation labs also use the same techniques to certify device security. In recent years, deep learning has emerged as a prominent method for SCA, achieving state-of-the-...
Poster
You Only Communicate Once: One-shot Federated Low-Rank Adaptation of MLLM
https://neurips.cc//virtual/2025/poster/119011
Binqian Xu, Haiyang Mei, Zechen Bai, Jinjin Gong, Rui Yan, Guosen Xie, Yazhou Yao, Basura Fernando, Xiangbo Shu
Multimodal Large Language Models (MLLMs) with Federated Learning (FL) can quickly adapt to privacy-sensitive tasks, but face significant challenges such as high communication costs and increased attack risks, due to their reliance on multi-round communication. To address this, One-shot FL (OFL) has emerged, aiming to c...
Poster
You Only Spectralize Once: Taking a Spectral Detour to Accelerate Graph Neural Network
https://neurips.cc//virtual/2025/poster/117732
Yi Li, Zhichun Guo, Guanpeng Li, Bingzhe Li
Training Graph Neural Networks (GNNs) often relies on repeated, irregular, and expensive message-passing operations over all nodes (e.g., $N$), leading to high computational overhead. To alleviate this inefficiency, we revisit the GNNs training from a spectral perspective. In many real-world graphs, node features and e...
Poster
Your Pre-trained LLM is Secretly an Unsupervised Confidence Calibrator
https://neurips.cc//virtual/2025/poster/118852
Beier Luo, Shuoyuan Wang, Sharon Li, Hongxin Wei
Post-training of large language models is essential for adapting pre-trained language models (PLMs) to align with human preferences and downstream tasks. While PLMs typically exhibit well-calibrated confidence, post-trained language models (PoLMs) often suffer from over-confidence, assigning high confidence to both cor...
Poster
Zebra-Llama: Towards Extremely Efficient Hybrid Models
https://neurips.cc//virtual/2025/poster/116316
Mingyu Yang, Mehdi Rezagholizadeh, Guihong Li, Vikram Appia, Emad Barsoum
With the growing demand for deploying large language models (LLMs) across diverse applications, improving their inference efficiency is crucial for sustainable and democratized access. However, retraining LLMs to meet new user-specific requirements is prohibitively expensive and environmentally unsustainable. In this w...
Poster
ZEBRA: Towards Zero-Shot Cross-Subject Generalization for Universal Brain Visual Decoding
https://neurips.cc//virtual/2025/poster/119305
Haonan Wang, Jingyu Lu, Hongrui Li, Xiaomeng Li
Recent advances in neural decoding have enabled the reconstruction of visual experiences from brain activity, positioning fMRI-to-image reconstruction as a promising bridge between neuroscience and computer vision. However, current methods predominantly rely on subject-specific models or require subject-specific fine-t...
Poster
ZeCO: Zero-Communication Overhead Sequence Parallelism for Linear Attention
https://neurips.cc//virtual/2025/poster/116936
YUHONG CHOU, Zehao Liu, Rui-Jie Zhu, Xinyi Wan, Tianjian Li, Congying Chu, Qian Liu, Jibin Wu, Zejun MA
Linear attention mechanisms deliver significant advantages for Large Language Models (LLMs) by providing linear computational complexity, enabling efficient processing of ultra-long sequences (e.g., 1M context). However, existing Sequence Parallelism (SP) methods, essential for distributing these workloads across devic...
Poster
ZeroPatcher: Training-free Sampler for Video Inpainting and Editing
https://neurips.cc//virtual/2025/poster/116549
Shaoshu Yang, Yingya Zhang, Ran He
Video inpainting and editing have long been challenging tasks in the video generation community, requiring extensive computational resources and large datasets to train models with satisfactory performance. Recent breakthroughs in large-scale video foundation models have greatly enhanced text-to-video generation capabi...
Poster
ZeroSep: Separate Anything in Audio with Zero Training
https://neurips.cc//virtual/2025/poster/118822
Chao Huang, Yuesheng Ma, Junxuan Huang, Susan Liang, Yunlong Tang, Jing Bi, Wenqiang Liu, Nima Mesgarani, Chenliang Xu
Audio source separation is fundamental for machines to understand complex acoustic environments and underpins numerous audio applications. Current supervised deep learning approaches, while powerful, are limited by the need for extensive, task-specific labeled data and struggle to generalize to the immense variability ...
Poster
Zero-Shot Blind-Spot Denoising via Pixel Refilling
https://neurips.cc//virtual/2025/poster/118602
Qilong Guo, Tianjing Zhang, Zhiyuan Ma, Hui Ji
Blind-spot networks offer a powerful paradigm for zero-shot image denoising by training models to predict masked target pixels from their neighbors. However, they struggle with real-world noise exhibiting strong local correlations, where efforts to suppress noise correlation often weaken essential pixel dependencies, ...
Poster
Zero-Shot Context Generalization in Reinforcement Learning from Few Training Contexts
https://neurips.cc//virtual/2025/poster/115730
James Chapman, Kedar Karhadkar, Guido Montufar
Deep reinforcement learning (DRL) has achieved remarkable success across multiple domains, including competitive games, natural language processing, and robotics. Despite these advancements, policies trained via DRL often struggle to generalize to evaluation environments with different parameters. This challenge is typ...
Poster
Zero-shot Denoising via Neural Compression: Theoretical and algorithmic framework
https://neurips.cc//virtual/2025/poster/119178
Ali Zafari, Xi Chen, Shirin Jalali
Zero-shot denoising aims to denoise observations without access to training samples or clean reference images. This setting is particularly relevant in practical imaging scenarios involving specialized domains such as medical imaging or biology. In this work, we propose the Zero-Shot Neural Compression Denoiser (Z...
Poster
Zero-Shot Detection of LLM-Generated Text via Implicit Reward Model
https://neurips.cc//virtual/2025/poster/120141
Runheng Liu, Heyan Huang, Xingchen Xiao, Zhijing Wu
Large language models (LLMs) have demonstrated remarkable capabilities across various tasks. However, their ability to generate human-like text has raised concerns about potential misuse. This underscores the need for reliable and effective methods to detect LLM-generated text. In this paper, we propose IRM, a novel ze...
Poster
Zero-Shot Performance Prediction for Probabilistic Scaling Laws
https://neurips.cc//virtual/2025/poster/115947
Viktoria Schram, Markus Hiller, Daniel Beck, Trevor Cohn
The prediction of training and learning curves of Natural Language Processing (NLP) models enables targeted decision-making to achieve performance objectives, reduces computational overhead and minimizes costs associated with dataset acquisition and curation. In this work, we formulate both prediction tasks as multitas...
Poster
Zero-Shot Trajectory Planning for Signal Temporal Logic Tasks
https://neurips.cc//virtual/2025/poster/118254
Ruijia Liu, Ancheng Hou, Xiao Yu, Xiang Yin
Signal Temporal Logic (STL) is a powerful specification language for describing complex temporal behaviors of continuous signals, making it well-suited for high-level robotic task descriptions. However, generating executable plans for STL tasks is challenging, as it requires consideration of the coupling between the ta...
Poster
Zero-shot World Models via Search in Memory
https://neurips.cc//virtual/2025/poster/115426
Federico Malato, Ville Hautamäki
World Models have vastly permeated the field of Reinforcement Learning. Their ability to model the transition dynamics of an environment have led to tremendous improvements in sample efficiency for online RL. Among them, the most notorious example is Dreamer, a model that learns to act in a diverse set of image-based e...
Poster
ZeroS: Zero‑Sum Linear Attention for Efficient Transformers
https://neurips.cc//virtual/2025/poster/118425
Jiecheng Lu, Xu Han, Yan Sun, Viresh Pati, Yubin Kim, Siddhartha Somani, Shihao Yang
Linear attention methods offer Transformers $O(N)$ complexity but typically underperform standard softmax attention. We identify two fundamental limitations affecting these approaches: the restriction to convex combinations that only permits additive information blending, and uniform accumulated weight bias that dilute...
Poster
Zeroth-Order Optimization Finds Flat Minima
https://neurips.cc//virtual/2025/poster/116554
Liang Zhang, Bingcong Li, Kiran Thekumparampil, Sewoong Oh, Michael Muehlebach, Niao He
Zeroth-order methods are extensively used in machine learning applications where gradients are infeasible or expensive to compute, such as black-box attacks, reinforcement learning, and language model fine-tuning. Existing optimization theory focuses on convergence to an arbitrary stationary point, but less is known on...
Poster
ZEUS: Zero-shot Embeddings for Unsupervised Separation of Tabular Data
https://neurips.cc//virtual/2025/poster/116816
Patryk Marszałek, Tomasz Kuśmierczyk, Witold Wydmański, Jacek Tabor, Marek Śmieja
Clustering tabular data remains a significant open challenge in data analysis and machine learning. Unlike for image data, similarity between tabular records often varies across datasets, making the definition of clusters highly dataset-dependent. Furthermore, the absence of supervised signals complicates hyperparamete...
Poster
ZigzagPointMamba: Spatial-Semantic Mamba for Point Cloud Understanding
https://neurips.cc//virtual/2025/poster/116930
LinshuangDiao, Sensen Song, Dayong Ren, Yurong Qian
State Space Models (SSMs) like PointMamba provide efficient feature extraction for point cloud self-supervised learning with linear complexity, surpassing Transformers in computational efficiency. However, existing PointMamba-based methods rely on complex token ordering and random masking, disrupting spatial continuity...
Poster
zip2zip: Inference-Time Adaptive Vocabularies for Language Models via Token Compression
https://neurips.cc//virtual/2025/poster/118871
Saibo Geng, Nathan Ranchin, Yunzhen Yao, Maxime Peyrard, Chris Wendler, Michael Gastpar, Robert West
Tokenization efficiency plays a critical role in the performance and cost of large language models (LLMs), yet most models rely on static tokenizers optimized for general-purpose corpora. These tokenizers' fixed vocabularies often fail to adapt to domain- or language-specific inputs, leading to longer token sequences a...
Poster
Zooming from Context to Cue: Hierarchical Preference Optimization for Multi-Image MLLMs
https://neurips.cc//virtual/2025/poster/116709
Xudong Li, Mengdan Zhang, Peixian Chen, Xiawu Zheng, Yan Zhang, Jingyuan Zheng, Yunhang Shen, Ke Li, Chaoyou Fu, Xing Sun, Rongrong Ji
Multi-modal Large Language Models (MLLMs) excel at single-image tasks but struggle with multi-image understanding due to cross-modal misalignment, leading to hallucinations (context omission, conflation, and misinterpretation). Existing methods using Direct Preference Optimization (DPO) constrain optimization to a soli...
Poster
ZPressor: Bottleneck-Aware Compression for Scalable Feed-Forward 3DGS
https://neurips.cc//virtual/2025/poster/115036
Weijie Wang, Donny Y. Chen, Zeyu Zhang, Duochao Shi, Akide Liu, Bohan Zhuang
Feed-forward 3D Gaussian Splatting (3DGS) models have recently emerged as a promising solution for novel view synthesis, enabling one-pass inference without the need for per-scene 3DGS optimization. However, their scalability is fundamentally constrained by the limited capacity of their encoders, leading to degraded pe...