Datasets:

debajyotidasgupta

/

dblp_2022_2023

like 0

Title: When Is Multiple Access Beneficial? An Analysis of Multi-User Performance in IEEE 802.11ax Abstract: ABSTRACTIEEE 802.11ax introduces OFDMA (Orthogonal Frequency Division Multiple Access), that allows multiple users to transmit or receive frames concurrently. The standard suggests that OFDMA will provide reduced latency and increased throughput in dense scenarios compared to single-user OFDM. This work uses the latest 802.11ax models supplied by the widely used open-source network simulator NS-3 (version 3.34) to investigate the OFDMA performance under various downlink traffic loads and various application settings for both UDP and TCP traffic. Our simulation results show that the actual benefits of OFDMA over OFDM can only be extracted under intermediate traffic loads for UDP and TCP. We observe that in comparison to OFDM, OFDMA provides more consistent performance in the case of video streaming and video conferencing. Further, it provides lesser latency for web-based applications.

Title: Spectrum Sensing and SINR Estimation in an IEEE 802.11s Dynamic Spectrum Access Wireless Mesh Network Abstract: ABSTRACTWireless Mesh Networks (WMNs) that make opportunistic use of licensed spectrum (also called Cognitive Radio Ad Hoc Networks or CRAHNs) need to perform spectrum sensing (SS) to find an advantageous channel assignment. However, current SS methods for cognitive radios typically rely on binary decisions on the presence or absence of interference, or the presence or absence of a licensed user. Interference estimation in WMNs also relies on binary measures such as the presence or absence of conflicts. Yet, performance can be significantly improved if measures have smaller resolutions, i.e., are more fine-grained than binary decisions. On the matter of when SS is performed, existing work tends to require a trade-off between SS time and data transmission. This work presents Signal-to-Interference-and-Noise-Ratio estimation for Dynamic Spectrum Access WMNs that is granular and accurate. We also propose to use the idle time inherent in the backoff mechanism of Enhanced Distributed Channel Access (EDCA) for performing SS, which has the benefit of causing no disruption to data transmissions. The estimation method we suggest that uses the SS results employs the maximum a posteriori estimate, which is the same as the maximum likelihood estimator. We show that this is the best possible estimator as it is the minimum variance unbiased estimator and it is efficient since the Cramer-Rao bound is satisfied by equality. The performance is evaluated in terms of confidence intervals. We show that the time available during EDCA backoff for performing sensing is sufficient. We also show the margin of error that can be achieved based on the number of sensing windows employed.

Title: X-Haul Solutions for Different Functional Split Options Using THz and Sub-THz Bands Abstract: ABSTRACTIn a variety of scenarios, from temporal events to emergencies, mobile cell also known as cell on wheels (COW) or cell on light truck (COLT) is considered as a widespread solution for temporarily increasing the capacity of the cellular networks. Flying cells are small-sized, low-cost, fast deployment options of mobile cells. In the fifth generation (5G) of cellular systems, the functionalities of radio access network (RAN) components i.e., centralized unit (CU), distributed unit (DU), and radio unit (RU) vary in different functional splits and thus have different data rate requirements for the interfaces between the units. The initial target of the paper is to provide a basic overview of different functional splits introduced in 5G, and highlight the throughput requirements of the transport X-Haul links of those functional splits. Moreover, the target is to investigate the option of utilizing terahertz (THz) and sub-THz wireless radio link to meet a high data rate requirement. In this work, we considered two frequency bands i.e., 105 GHz and 220 GHz, and estimated the handling capacity of the X-Haul link for different supporting bandwidths and distances. The X-Haul link capacity requirements also depend upon the air interface configuration, therefore, in this work different bandwidths, the number of antennas, and MIMO layer configurations are considered. The analysis is applicable for the terrestrial as well as for the flying platforms. Interestingly, it is found that with a lower functional split i.e., split 8, the X-Haul data requirement is above 150 Gbps for a simple 5G system with only 100 MHz bandwidth and 32 antennas.

Title: Crowding Game and Deep Q-Networks for Dynamic RAN Slicing in 5G Networks Abstract: ABSTRACTFifth generation (5G) mobile networks do not solely support services with very high throughput, but also answer the requirements of other heterogeneous services, characterized by different Quality of Service (QoS) criteria such as maximum tolerated delay, minimum guaranteed throughput, or capacity constraints. This heterogeneity necessitates the partitioning of the available radio resources into multiple slices, where each slice is characterized by specific QoS and isolation constraints. Nonetheless, a static resource allocation scheme among slices might not be an effective solution to face the dynamic and sporadic nature of traffic and users' service types. Therefore, a dynamic slicing scheme that leverages on the ease of slice selection per user in the fully virtualized open Radio Access Networks (RAN), is paramount to accommodate the various demands and load conditions. The main contribution of this paper is to apply traffic engineering in the scope of RAN slicing. In fact, instead of solely re-dimensioning the slices by injecting more resources when congestion hits, we also allot users to slices that may differ from their service type if their performance target is met. This is a novel definition of dynamic slicing, the dynamicity not being in re-dimensioning on the fly but in accommodating active users to existing resources for higher resource utilization efficiency without hindering users' performances. Additionally, the available bandwidth is dynamically adjusted among slices based on their load and affected users' performances. To reach that goal, two Dynamic RAN Slicing schemes are proposed: a centralized scheme based on Deep Reinforcement Learning (DRL) and a distributed scheme based on non-cooperative game theory. Exhaustive numerical simulations demonstrate the high efficiency of our proposed approach in comparison with the state-of-the-art where users are allocated to slices according to their service type and required Key Performance Indicators (KPIs).

Title: Assessing the Electromagnetic Field Exposure of 5G Transmitters: Challenges & Solutions Abstract: ABSTRACTThe ever-increasing densification of cellular base stations, combined with the use of active antenna systems, leads to concerns about human exposure to radio-frequency electromagnetic fields (EMF). From a regulatory point of view, adapting assessment and validation methodologies for new 5G installations is also a complex exercise, as it will be for future generations. Most methodologies used by regulators rely on calculations or measurement approaches that are not always appropriate - leading to an overestimation of EMF levels. This paper gives an overview of existing calculation methods to correctly estimate the EMF exposure of a 5G mobile network, and provides an analysis to select the most appropriate framework. To do so, we first describes the key definitions and concepts framing EMF exposure assessment. Then, we review the state-of-the-art assessment methods that consider 5G's key features: Massive MIMO, millimeter Wave, and precise beamforming. Then, we propose an analysis based on Non-standalone (NSA) and Standalone standards-based (SA) paths to 5G. Next, we analyse the key standard used in Europe, i.e. IEC62232, and compare it with other standards and approaches to give a comprehensive analysis of what is missing and needed for the next versions. Finally, we select the most appropriate framework for calculation and computation considering the new wireless technologies.

Title: Optimization of Path Loss Prediction in Urban and Suburban Environments in 2.3-2.4 GHz using OLS Nonlinear Regression and Artificial Neural Networks Abstract: ABSTRACTThe path loss models for urban and suburban scenarios play a key role since they provide RF power estimations that allow an optimized design and performance of the wireless network. Although different models of propagation can be found in the literature, it is always important to consider the particular characteristics of urbanization in each region where a 5G technology will be implemented. In this work, the empirical results of propagation loss in the 2.3-2.4 GHz band in an urban and suburban environment are compared with some widely used empirical models. Subsequently, the nonlinear regression of the propagation loss values obtained in the field is performed based on an Ordinary Least Squares (OLS) approach and an Artificial Neural Networks (ANN) approach. The empirical results, both in urban and suburban scenarios, were better represented by the SUI model. The nonlinear regression results generated by the ANN modeling were better than the OLS regression results.

Title: Environment-Aware Link Quality Prediction for Millimeter-Wave Wireless LANs Abstract: ABSTRACTMillimeter-wave (mmWave) communications have been regarded as one of the most promising solutions to deliver ultra-high data rates in wireless local-area networks. A significant barrier to delivering consistently high rate performance is the rapid variation in quality of mmWave links due to blockages and small changes in user locations. If link quality can be predicted in advance, proactive resource allocation techniques such as link-quality-aware scheduling can be used to mitigate this problem. In this paper, we propose a link quality prediction scheme based on knowledge of the environment. We use geometric analysis to identify the shadowed regions that separate LoS and NLoS scenarios, and build LoS and NLoS link-quality predictors based on an analytical model and a regression-based approach, respectively. For the more challenging NLoS case, we use a synthetic dataset generator with accurate ray tracing analysis to train a deep neural network (DNN) to learn the mapping between environment features and link quality. We then use the DNN to efficiently construct a map of link quality predictions within given environments. Extensive evaluations with additional synthetically generated scenarios show a very high prediction accuracy for our solution. We also experimentally verify the scheme by applying it to predict link quality in an actual 802.11ad environment, and the results show a close agreement between predicted values and measurements of link quality.

Title: Fuzzy representation of vague spatial descriptions in real estate advertisements Abstract: ABSTRACTGeocoding a spatial description is challenging since vernacular place names and vague spatial expressions give uncertainty and ambiguity to the description. Usually, digital gazetteers are used to match geospatial objects to their boundaries. However, gazetteers do not contain all places. Therefore, a number of studies have proposed to enrich gazetteers by estimating and representing the vernacular places. Nevertheless, only a few approaches have taken into account vague spatial expressions such as "nearby", and have represented geospatial objects as sharp boundaries. In this work, we present an automatic workflow to retrieve a location approximation of vague spatial description. We propose a model to estimate a fuzzy representation of each mentioned geospatial information and spatial expressions. Then, we perform information fusion to find a location approximation of a property. Lastly, we demonstrate our proposed method by applying it to the case of French Real Estate advertisements with two real-world datasets in Nice and Paris. Real Estate advertisements allow us to deal with uncertain geospatial objects since avague and exaggerated property location's description is usually provided. Our results show that our proposed method is promising and able to correctly approximate a location from uncertain spatial descriptions.

Title: Preference aware route recommendation using one billion geotagged tweets Abstract: ABSTRACTTwitter is a popular social networking service where people send short messages called tweets. Tweets contain metadata such as language, hashtags, geotags, and time of creation. We focus on the geotags of tweets. A Geo-tag is georeferenced information that indicates the geographical origin of a tweet. Geotagged tweets provide an excellent opportunity to understand the underlying user behavior. We propose a preference-aware route recommendation method relying on over one billion geotagged tweets. The method can recommend routes based on user preference by extracting a subset of one billion geotagged tweets according to user preference and using that subset to generate a cost function for route discovery. The proposed method assumes that areas with a high density of geotagged tweets are areas of high interest. In other words, if the density of geotagged tweets with user preference is superimposed on the cost of the route search, the users' preference can be considered when recommending a route. We highlight a nighttime route recommendation mechanism for a case study of our method. We hypothesize that geotagged tweets sent out at night indicate human activity at night. In other words, areas with a high density of geo-tagged tweets are considered to be areas that are vibrant at night. In addition, it is empirically clear that nighttime vibrant is also based on brightness. Therefore, we utilize nighttime tweets and nighttime light data to recommend routes. We extract a subset by calculating nighttime from tweet metadata. Tweets data are divided into grids and used to calculate a vibrant grid from a weighted tweets grid and a nighttime lights grid. Edge is weighted from vibrant cell values and road network edge lengths to recommend a vibrant route based on weighted road network edges. We experimented in Shinjuku, Tokyo, Japan, between two stations. As a result, based on the objective evaluation, we recommended a vibrant route.

Title: Addressing the location A/B problem on Twitter: the next generation location inference research Abstract: ABSTRACTOften, global and regional topics on Twitter across multiple thematic areas, such as disasters, politics, protests, entertainment, epidemics, literature, travel, culture, weather, etc., witness an unprecedented level of exchange of conversations. An issue with those conversations is that a user can be at location A and participate in a public discourse specific to location B, which we refer to as the Location A/B problem. Location profiling of users solely based on locations mentioned in their tweets leads to ineffective location-based recommendations. The problem is deemed solved if location candidates could be categorized as either origin locations (Location As) or non-origin locations (Location Bs); however, real-world tweets are much more complex, and currently, no public datasets are available for training such classifiers. To the best of our knowledge, this study yields the first steps in addressing the Location A/B problem on Twitter. We propose a theoretical framework that utilizes the existing literature on location inference to categorize location candidates as either origin locations or non-origin locations. We envision that: (i) the framework provides the grounds for designing models that aim to solve the Location A/B problem, and (ii) the location profiling of users based on origin locations leads to improved geotargeted recommendations.

Title: Doing groceries again: towards a recommender system for grocery stores selection Abstract: ABSTRACTChoosing a store (i.e. grocery, restaurant etc.) depends on different decision criteria. If the data for these criteria is distributed among different sources a user might need to invest a substantial amount of time to aggregate the necessary information from different resources or base their decisions only on a subset of criteria. Additionally, visualising all criteria can augment the user's decision making. In this work, we demonstrate a prototype that is able to combine the data of different decision criteria from different (online) resources and provides recommendations of the combined decision criteria. Additionally, a skyline facilitates the choice of stores that dominate specific features. As a concrete example, we state a query of the type "Get me all stores of a supermarket (of a particular company) in the vicinity". The data for the chosen criteria of traffic time, distance, or occupancy of stores were obtained from Google traffic, popular times and timeline. The timeline data is used for our introduced decision criterion 'utility' which is an indicator of how much added value is gained by visiting a particular store. The visualization of this allows users to see at one glance different criteria of their decision-making of which supermarket to choose, which can make the difference between an efficient and hassle-free groceries experience and one that comes at a high cost of money, time and nerves.

Title: DRIFT: E2EE spatial feature sharing & instant messaging Abstract: ABSTRACTMost online communication today is inherently temporal and aspatial. Instant messaging (IM) services are structured around a timeline interface which prioritizes a linear succession of events and guides our attention towards the novel. In this way, the different textures of social life are lost in linear reduction. In this paper, we present DRIFT, a novel and open-source IM application framework, based on a different paradigm of communication that preserves temporality but organizes it around space. Instead of the timeline, our application grounds messaging in the map and its pins, offering users a tool that encourages spatio-temporal communication and the sharing of spatial features. Given increasing concerns about the safety and privacy of online user interaction, we integrate state-of-the art encryption as a core feature of our application. Firstly, to protect user messages and map pins, we implement end-to-end encryption with the Double Ratchet key management algorithm and the open standard Matrix protocol. Secondly, to maintain location privacy, we allow users to batch download map tilesets and machine learning models to perform operations such as search entirely on device, avoiding compromising API calls to cloud services. With these combined features, DRIFT aims to introduce a new model for online interaction that upends the short attention span imposed by the narrow timeline and replace it with a spatio-temporally rich and secure IM tool for both laymen and more vulnerable users such as journalists, human rights activists, and whistleblowers.

Title: Utilizing location-based social media for trip mining and recommendation Abstract: ABSTRACTWolfgang Wörndl has given an invited talk at the 6th ACM SIGSPATIAL Workshop on Location-based Recommendations, Geosocial Networks and Geoadvertising (LocalRec) on November 1, 2022. This paper provides a summary of the topics addressed.

Title: Leveraging Multi-modal Interactions among the Intermediate Representations of Deep Transformers for Emotion Recognition Abstract: ABSTRACTMulti-modal emotion recognition aims to recognize emotion states from multi-modal inputs. Existing end-to-end models typically fuse the uni-modal representations in the last layers without leveraging the multi-modal interactions among the intermediate representations. In this paper, we propose the multi-modal Recurrent Intermediate-Layer Aggregation (RILA) model to explore the effectiveness of leveraging the multi-modal interactions among the intermediate representations of deep pre-trained transformers for end-to-end emotion recognition. At the heart of our model is the Intermediate-Representation Fusion Module (IRFM), which consists of the multi-modal aggregation gating module and multi-modal token attention module. Specifically, at each layer, we first use the multi-modal aggregation gating module to capture the utterance-level interactions across the modalities and layers. Then we utilize the multi-modal token attention module to leverage the token-level multi-modal interactions. The experimental results on IEMOCAP and CMU-MOSEI show that our model achieves the state-of-the-art performance, benefiting from fully exploiting the multi-modal interactions among the intermediate representations.

Title: Bridging the Gap: End-to-End Domain Adaptation for Emotional Vocalization Classification using Adversarial Learning Abstract: ABSTRACTGood classification performance on a hold-out partition can only be expected if the data distribution of the test data matches the training data. However, in many real-life use cases, this constraint is not met. In this work, we explore if it is feasible to use existing methods of an adversarial domain transfer to bridge this inter-domain gap. To do so, we use a CycleGAN that was trained on converting between the domains. We demonstrate that the quality of the generated data has a substantial impact on the effectiveness of the domain adaptation, and propose an additional step to overcome this problem. To evaluate the approach, we classify emotions in female and male vocalizations. Furthermore, we show that our model successfully approximates the distribution of acoustic features and that our approach can be employed to improve emotion classification performance. Since the presented approach is domain and feature independent it can therefore be applied to any classification task.

Title: Transformer-based Non-Verbal Emotion Recognition: Exploring Model Portability across Speakers' Genders Abstract: ABSTRACTRecognizing emotions in non-verbal audio tracks requires a deep understanding of their underlying features. Traditional classifiers relying on excitation, prosodic, and vocal traction features are not always capable of effectively generalizing across speakers' genders. In the ComParE 2022 vocalisation sub-challenge we explore the use of a Transformer architecture trained on contrastive audio examples. We leverage augmented data to learn robust non-verbal emotion classifiers. We also investigate the impact of different audio transformations, including neural voice conversion, on the classifier capability to generalize across speakers' genders. The empirical findings indicate that neural voice conversion is beneficial in the pretraining phase, yielding an improved model generality, whereas is harmful at the finetuning stage as hinders model specialization for the task of non-verbal emotion recognition.

Title: Emotional Reaction Analysis based on Multi-Label Graph Convolutional Networks and Dynamic Facial Expression Recognition Transformer Abstract: ABSTRACTAutomatically predicting and understanding human emotional reactions have wide applications in human-computer interaction. In this paper, we present our solutions to the MuSe-Reaction sub-challenge in MuSe 2022. The task of this sub-challenge is to predict the intensity of 7 emotional expressions from human reactions to a wide range of emotionally evocative stimuli. Specifically, we design an end-to-end model, which is composed of a Spatio-Temporal Transformer for dynamic facial representation learning and a multi-label graph convolutional network for emotion dependency modeling.We also explore the effects of a temporal model with a variety of features from acoustic and visual modalities. Our proposed method achieves mean Pearson's correlation coefficient of 0.3375 on the test set of MuSe-Reaction, which outperforms the baseline system(i.e., 0.2801) by a large margin.

Title: ViPER: Video-based Perceiver for Emotion Recognition Abstract: ABSTRACTRecognizing human emotions from videos requires a deep understanding of the underlying multimodal sources, including images, audio, and text. Since the input data sources are highly variable across different modality combinations, leveraging multiple modalities often requires ad hoc fusion networks. To predict the emotional arousal of a person reacting to a given video clip we present ViPER, a multimodal architecture leveraging a modality-agnostic transformer based model to combine video frames, audio recordings, and textual annotations. Specifically, it relies on a modality-agnostic late fusion network which makes ViPER easily adaptable to different modalities. The experiments carried out on the Hume-Reaction datasets of the MuSe-Reaction challenge confirm the effectiveness of the proposed approach.

Title: Multimodal Temporal Attention in Sentiment Analysis Abstract: ABSTRACTIn this paper, we present the solution to the MuSe-Stress sub-challenge in the MuSe 2022 Multimodal Sentiment Analysis Challenge. The task of MuSe-Stress is to predict a time-continuous value (i.e., physiological arousal and valence) based on multimodal data of audio, visual, text, and physiological signals. In this competition, we find that multimodal fusion has good performance for physiological arousal on the validation set, but poor prediction performance on the test set. We believe that problem may be due to the over-fitting caused by the model's over-reliance on some specific modal features. To deal with the above problem, we propose Multimodal Temporal Attention (MMTA), which considers the temporal effects of all modalities on each unimodal branch, realizing the interaction between unimodal branches and adaptive inter-modal balance. The concordance correlation coefficient (CCC) of physiological arousal and valence are 0.6818 with MMTA and 0.6841 with early fusion, respectively, both ranking Top 1, outperforming the baseline system by a large margin (i.e., 0.4761 and 0.4931) on the test set.

Title: Improving Dimensional Emotion Recognition via Feature-wise Fusion Abstract: ABSTRACTThis paper introduces the solution of the RiHNU team for the MuSe-Stress sub-challenge as part of Multimodal Sentiment Analysis (MuSe) 2022. The MuSe-Stress is a task to discern human emotional states via internal or external responses (e.g., audio, physiological signal, and facial expression) in a job-interview setting. Multimodal learning is extensively considered an available approach for multimodal sentiment analysis tasks. However, most multimodal models fail to capture the association among each modality, resulting in limited generalizability. We argue that those methods are incapable of establishing discriminative features, mainly because they typically neglect fine-grained information. To address this problem, we first encode spatio-temporal features via a feature-wise fuse mechanism to learn more informative representations. Then we exploit the late fusion strategy to capture fine-grained relations between multiple modalities. The ensemble strategy is also used to enhance the final performance. Our method achieves CCC of 0.6803 and 0.6689 for valence and physiological arousal, respectively, on the test set.

Title: Towards Multimodal Prediction of Time-continuous Emotion using Pose Feature Engineering and a Transformer Encoder Abstract: ABSTRACTMuSe-Stress 2022 aims at building sequence regression models for predicting valence and physiological arousal levels of persons who are facing stressful conditions. To that end, audio-visual recordings, transcripts, and physiological signals can be leveraged. In this paper, we describe the approach we developed for Muse-Stress 2022. Specifically, we engineered a new pose feature that captures the movement of human body keypoints. We also trained a Long Short-Term Memory (LSTM) network and a Transformer encoder on different types of feature sequences and different combinations thereof. In addition, we adopted a two-pronged strategy to tune the hyperparameters that govern the different ways the available features can be used. Finally, we made use of late fusion to combine the predictions obtained for the different unimodal features. Our experimental results show that the newly engineered pose feature obtains the second highest development CCC among the seven unimodal features available. Furthermore, our Transformer encoder obtains the highest development CCC for five out of fourteen possible combinations of features and emotion dimensions, with this number increasing from five to nine when performing late fusion. In addition, when searching for optimal hyperparameter settings, our two-pronged hyperparameter tuning strategy leads to noticeable improvements in maximum development CCC, especially when the underlying models are based on an LSTM. In summary, we can conclude that our approach is able to achieve a test CCC of 0.6196 and 0.6351 for arousal and valence, respectively, securing a Top-3 rank in Muse-Stress 2022.

Title: Behavioral Consequences of Reminder Emails on Students’ Academic Performance: a Real-world Deployment Abstract: ABSTRACTPost-secondary institutions have experienced a continuous trend of student procrastination on course work, thus leading to lower academic performance and potentially worse knowledge retention and other longer-term impacts. Sending reminders about deliverables is a simple approach, but it has the potential to be a valuable tool to mitigate such issues and assist students with time management. This paper will introduce specific processes of conducting such experiments, especially email designs and randomization, to help instructors and researchers conduct similar experiments or field-deploying reminders. To evaluate homework reminder messages, we designed and deployed a real-world randomized A/B experiment at a North American university in a CS1 course where students were randomly assigned to either receive these reminder messages or not. Our findings suggest that students who received the reminder messages have a higher homework completion rate (p < .05) and performed significantly better (p < .01) on the following midterm test than students who did not receive the reminder message. Finally, we discuss how a homework reminder can improve student behaviours, as well as how this type of reminder can be enhanced for future interventions.

Title: Comparing Biosignal and Acoustic feature Representation for Continuous Emotion Recognition Abstract: ABSTRACTAutomatic recognition of human emotion has a wide range of applications. Human emotions can be identified across different modalities, such as biosignal, speech, text, and mimics. This paper is focusing on time-continuous prediction of level of valence and psycho-physiological arousal. In that regard, we investigate, (a) the use of different feature embeddings obtained from neural networks pre-trained on different speech tasks (e.g., phone classification, speech emotion recognition) and self-supervised neural networks, (b) estimation of arousal and valence from physiological signals in an end-to-end manner and (c) combining different neural embeddings. Our investigations on the MuSe-Stress sub-challenge shows that (a) the embeddings extracted from physiological signals using CNNs trained in an end-to-end manner improves over the baseline approach of modeling physiological signals, (b) neural embeddings obtained from phone classification neural network and speech emotion recognition neural network trained on auxiliary language data sets yield improvement over baseline systems purely trained on the target data, and (c) task-specific neural embeddings yield improved performance over self-supervised neural embeddings for both arousal and valence. Our best performing system on test-set surpass the DeepSpectrum baseline (combined score) by a relative 7.7% margin

Title: Modular experiential learning for secure, safe, and reliable AI: Curricular Initiative to Promote Education in Trustworthy AI Abstract: ABSTRACTArtificial intelligence (AI) is increasingly applied to IT systems. However, AI can be manipulated to perform undesirably, exhibit biases or abusive behaviors. When AI algorithms are parallelized on high-performance computing-based cyberinfrastructure (CI), such misbehaviors and uncertainty can multiply to obscure the root causes. Secure, safe, and reliable computing techniques can mitigate these problems. The project described in this paper aims to inform curriculum and develop materials to educate students who use AI from the outset, so that they will first become aware of the issues and secondly practical considerations will be integrated with theory in classes. Intensive, multi-faceted, modular, experiential learning units are designed to rapidly upgrade the skills of current and future CI users, so they can apply new skills to their tasks. The loosely coupled modules can be taken as standalone self-directed units or integrated into existing classes, starting with CS 1 and CS 2, which are taken by many non-CS STEM students. In a sandpit environment, learners take measured risks when guided on a journey of discovery. The primary purpose of this paper is to present key findings of the research following a 2-year pilot. A secondary purpose of the paper is to disseminate this exciting endeavor broadly, so that likeminded educators and researchers can consider participating in the project.

Title: iThem: Programming Internet of Things Beyond Trigger-Action Pattern Abstract: ABSTRACT With emerging technologies bringing Internet of Things (IoT) devices into domestic environments, trigger-action programming such as IFTTT with its simple if-this-then-that pattern provides an effective way for end-users to connect fragmented intelligent services and program their own smart home/work space automation. While the simplicity of trigger-action programming can be effective for non-programmers with its straightforward concepts and graphical user interface, it does not allow the algorithmic expressivity that a programming language has. For instance, the simple if-this-then-that structure cannot cover complex algorithms that arise from real world scenarios involving multiple conditions or keeping track of a sequence of conditions (e.g., incrementing counters, triggering one action if two conditions are both true). In this exploratory work, we take an alternative approach by creating a programmable channel between application programming interfaces (APIs), which allows programmers to preserve states and to use them to write complex algorithms. We propose iThem, which stands for intelligence of them—internet of things, that allow programmers to author any complex algorithms that can connect different IoT services and fully unleash the freedom of a general programming language. In this poster, we share the design, development, and ongoing validation progress of iThem, which piggybacks on existing programmable IoT system IFTTT, and which allows for a programmable channel that connects triggers and actions in IFTTT with versatility.

Title: FullPull : A Stretchable UI to Input Pulling Strength on Touch Surfaces Abstract: ABSTRACTTouch surfaces are used as input interfaces for many devices such as smartphones, tablets, and smartwatches. However, the flexibility of the input surface is low, and the possible input operations are limited to planar ones such as touch and swipe. In contrast, in the field of HCI, there has been much research on increasing the number of input interactions by attaching augmented devices with various physical characteristics to the touch surface. However, most of these interactions are limited to operations where pressure is applied to the input surface. In this study, we propose FullPull, which consists of a rubber tube filled with conductive ink and a suction cup to attach the rubber tube to the surface. FullPull allows users to input pulling depth and strength on the touch surface. We implemented a prototype FullPull device which can be attached to an existing capacitive touch surface and can be pulled by a user. We then evaluated the accuracy of tensile strength estimation of the implemented device. The results showed that the outflow current value when stretched could be classified into four tensile strength levels.

Title: Managerial and Professional Skills and Dispositions from Professionals’ Interviews Abstract: ABSTRACTEmployability should be a primary objective for computing programs, as the majority of IT and other computing graduates go to work in industry upon graduation. Furthermore, students want to be prepared for a career, not just an entry-level job. However, literature has shown a gap between employers’ needs and undergraduates’ preparation in non-technical areas. Competencies (skills, knowledge, and dispositions) can be a common language used by both employers and educators. The more we learn about competencies employers expect, the more we can ensure programs match their expectations. This study focuses on competencies required by managers, by interviewing ten directors/managers, project managers, and product managers who had prior experience in computing-related roles. Each was asked to discuss competencies most important to their current position. Emerging themes identified the most important managerial skills (project management, evaluation of candidates, mentorship, managers’ own technical skills and knowledge, adjusting management style as needed, and appropriately assigning team members), professional skills (communication, problem solving, and relationship building), and dispositions (lifelong learning; adaptability/flexibility; being self-driven; self-awareness; being helpful, positive and pleasant; valuing communication and collaboration; having passion for technical work; and perseverance). Implications for education are discussed. This study is part of a larger NSF-funded project related to investigating the competencies required by computing professionals, and the design of educational resources to promote the development of these competencies.

Title: A Personalised Approach to Audiovisual Humour Recognition and its Individual-level Fairness Abstract: ABSTRACTHumour is one of the most subtle and contextualised behavioural patterns to study in social psychology and has a major impact on human emotions, social cognition, behaviour, and relations. Consequently, an automatic understanding of humour is crucial and challenging for a naturalistic human-robot interaction. Recent artificial intelligence (AI)-based methods have shown progress in multimodal humour recognition. However, such methods lack a mechanism in adapting to each individual's characteristics, resulting in a decreased performance, e.g., due to different facial expressions. Further, these models are faced with generalisation problems when being applied for recognition of different styles of humour. We aim to address these challenges by introducing a novel multimodal humour recognition approach in which the models are personalised for each individual in the Passau Spontaneous Football Coach Humour (Passau-SFCH) dataset. We begin by training a model on all individuals in the dataset. Subsequently, we fine-tune all layers of this model with the data from each individual. Finally, we use these models for the prediction task. Using the proposed personalised models, it is possible to significantly (two-tailed t-test, p < 0.05) outperform the non-personalised models. In particular, the mean Area Under the Curve (AUC) is increased from .7573 to .7731 for the audio modality, and from .9203 to .9256 for the video modality. In addition, we apply a weighted late fusion approach which increases the overall performance to an AUC of .9308, demonstrating the complementarity of the features. Finally, we evaluate the individual-level fairness of our approach and show which group of subjects benefits most of using personalisation.

Title: Search with Space: Find and Visualize Furniture in Your Space Abstract: ABSTRACT Online shopping in the home category enables quick and convenient access to large catalog of products. In particular, users can simultaneously browse for functional requirements, such as size and material, while evaluating aesthetic fit, such as color and style, across hundreds of product offerings. However, the typical user flow requires navigating to an e-commerce retailer’s website first, setting the search/filter parameters that may be generic, and then landing on product pages, one at a time, to make a decision. Upon purchase, ”does not fit” is among the top reasons for returning a product. Amalgamating the above information, we present Search with Space, a novel interactive approach that a) inputs the user’s space as a search parameter to b) filter for product matches that will physically fit, and c) visualize these matches in the user’s space at true scale and in a format that facilitates simultaneous comparison. Briefly, the user leverages augmented reality (AR) to set a proxy 3d product in the desired location, updates the proxy’s dimensions, and takes photos from preset angles. Using spatial information captured with AR, a web-based gallery page is curated with all the product matches that will physically fit and products are shown at true scale in their original photos. The user may now browse products visualized in the context of their space and evaluate based on their shopping criteria, share the gallery page with designers or partners for asynchronous feedback, re-use the photos for a different product class, or re-capture their space with different criteria altogether. Search with Space inverts the typical user journey by starting with the user’s space and maintaining that context across all touch points with the catalog.

Title: VLOGS: Virtual Laboratory Observation Tool for Monitoring a Group of Students Abstract: ABSTRACT Virtual laboratories (VLs) enable students to conduct lab experiments in the virtual world using Virtual Reality (VR) technology, providing benefits in areas such as availability, safety as well as scalability. While there are existing platforms that provide VLs with rich content as well as research works on promoting effective learning in VLs, less attention has been paid on VLs from a teaching perspective. Students usually learn and practice in VL sessions with limited help from the instructors. Instructors, on the other hand, could only receive limited information on the performance of the students and could not provide timely feedback to facilitate students’ learning. In this work, we present a prototype virtual laboratory monitoring tool, created using a design thinking approach, for addressing teaching needs when conducting a virtual laboratory session simultaneously for multiple students, similar to that in a physical lab environment.

Title: Integrating Cross-modal Interactions via Latent Representation Shift for Multi-modal Humor Detection Abstract: ABSTRACTMulti-modal sentiment analysis has been an active research area and has attracted increasing attention from multi-disciplinary communities. However, it is still challenging to fuse the information from different modalities in an efficient way. In prior studies, the late fusion strategy has been commonly adopted due to its simplicity and efficacy. Unfortunately, it failed to model the interactions across different modalities. In this paper, we propose a transformer-based hierarchical framework to effectively model both the intrinsic semantics and cross-modal interactions of the relevant modalities. Specifically, the features from each modality are first encoded via standard transformers. Later, the cross-modal interactions from one modality to other modalities are calculated using cross-modal transformers. The derived intrinsic semantics and cross-modal interactions are used to determine the latent representation shift of a particular modality. We evaluate the proposed approach on the MuSe-Humor sub-challenge of Multi-modal Sentiment Analysis Challenge (MuSe) 2022. Experimental results show that an Area Under the Curve (AUC) of 0.9065 can be achieved on the test set of MuSe-Humor. With the promising results, our best submission ranked first place in the sub-challenge.

Title: LipLearner: Customizing Silent Speech Commands from Voice Input using One-shot Lipreading Abstract: ABSTRACTWe present LipLearner, a lipreading-based silent speech interface that enables in-situ command customization on mobile devices. By leveraging contrastive learning to learn efficient representations from existing datasets, it performs instant fine-tuning for unseen users and words using one-shot learning. To further minimize the labor of command registration, we incorporate speech recognition to automatically learn new commands from voice input. Conventional lipreading systems provide limited pre-defined commands due to the time cost and user burden of data collection. In contrast, our technique provides expressive silent speech interaction with minimal data requirements. We conducted a pilot experiment to investigate the real-time performance of LipLearner, and the result demonstrates that an average accuracy of is achievable with only one training sample for each command.

Title: A Dangerous Infodemic: An Examination of the Impact Social Media Misinformation has on COVID-19 Vaccination Status Abstract: ABSTRACT The concept of misinformation is not new, but the digital age has created a new environment for the rapid spreading of misinformation. The overabundance of information that is available online has made it challenging for individuals to identify trustworthy and reliable sources. Social media in particular provides a global network connecting users, and the information there is created by the users themselves; therefore, it can be inaccurate and subjective. Throughout the COVID-19 pandemic, social media sites have acted as facilitators and multipliers of COVID-19-related misinformation. This misinformation can have a significant impact on global health by impacting individuals’ behaviors and has the potential to cause significant harm. In this paper, we explore how COVID-19 misinformation found via social media impacts individuals’ decisions to get vaccinated against COVID-19. The results from our study suggest that as one’s beliefs in misinformation and conspiracies related to COVID-19 increase, so does their decision to not obtain a vaccine.

Title: Amplified Carousel: Amplifying the Perception of Vertical Movement using Optical Illusion Abstract: ABSTRACT With the spread of virtual reality (VR) attractions, vector generation techniques that enhance the sense of realism are gaining attention. Additionally, mixed reality (MR) attractions, which overlay VR onto a real-world display, are expected to become more prevalent in the future. However, with MR, it is impossible to move all the coordinates of the visual stimuli to generate proper vection effects. Therefore, we have created an optical illusion method that provides a three-dimensional impression of a two-dimensional visual stimulation. The technique amplifies the sensation of vertical movement by placing the illusion on the floor.

Title: The Dos and Don'ts of Affect Analysis Abstract: ABSTRACTAs an inseparable and crucial component of communication affects play a substantial role in human-device and human-human interaction. They convey information about a person's specific traits and states [1, 4, 5], how one feels about the aims of a conversation, the trustworthiness of one's verbal communication [3], and the degree of adaptation in interpersonal speech [2]. This multifaceted nature of human affects poses a great challenge when it comes to applying machine learning systems for their automatic recognition and understanding. Contemporary self-supervised learning architectures such as Transformers, which define state-of-the-art (SOTA) in this area, have shown noticeable deficits in terms of explainability, while more conventional, non-deep machine learning methods, which provide more transparency, often fall (far) behind SOTA systems. So, is it possible to get the best of these two 'worlds'? And more importantly, at what price? In this talk, I provide a set of Dos and Don'ts guidelines for addressing affective computing tasks w. r. t. (i) preserving privacy for affective data and individuals/groups, (ii) being efficient in computing such data in a transparent way, (iii) ensuring reproducibility of the results, (iv) knowing the differences between causation and correlation, and (v) properly applying social and ethical protocols.

Title: Covert Eye Op App: An Offense Based Learning Approach Towards Developing Mobile Security Awareness and Interest in Cybersecurity Abstract: ABSTRACTThis paper introduces a unique approach of teaching mobile security awareness at the high school level through a nifty offense-based learning strategy. Our approach involves creating an eye-opening experience for learners through our own mobile app, which has been designed and developed strategically, so that it requests unnecessary permissions from users and secretly exploits them in the form of a covert offensive operation, that includes recording their audio plus tracking their location. When the users notice this exploit activity orchestrated by our app and realize how their provided permissions have backfired on them, they get to learn first-hand about the ways in which a mobile app can misuse user permissions and covertly compromise user information. We have used this app to implement a hands-on experiential learning activity that is intended to teach users the importance of privacy and security in mobile devices by breaching them and making them self-discover issues with how users grant permissions to mobile apps. To our knowledge, there has been limited prior work that focuses on studying how offense-based user hacking techniques impact leaning of mobile security topics. In this paper, we attempt to address this research gap. This paper describes our mobile app, as well as our offense-based lesson plan, which has been used in several workshop sessions as a hands-on learning activity for the high school community since 2019. It also includes our learner assessment study that involves analysis of the quantitative and qualitive data that we have collected in the form of survey responses from different users at the high school level. The results from our study indicate that our offense-based learning approach using our unique app was able to successfully engage users and create a positive learning experience for the high school community by developing user awareness of mobile security related issues, plus overall interest in cybersecurity topics.

Title: ASTREL: Prototyping Shape-changing Interface with Variable Stiffness Soft Robotics Module Abstract: ABSTRACTPrototyping a shape-changing interface is challenging because it requires knowledge of both electronics and mechanical engineering. In this study, we introduced a prototyping platform using a soft pneumatic artificial muscle(PAMs) and modular 3D printed reinforcement. To facilitate a wide variety of applications we propose six types of reinforcement modules capable of either shape deformation and/or variable stiffness. Users can create an approximate prototype using lego-built modules with magnetic connectors. A modeling toolkit can then be used to recreate and customize the prototype structure. After 3D printing, the shape-changing interface can be assembled by threading the PAMs through holes in the reinforcement. We envision that this prototyping platform can be useful in shape-changing interface exploration, where researchers can create working prototypes easily, rapidly, and at a low cost.

Title: Uncovering the Nuanced Structure of Expressive Behavior Across Modalities Abstract: ABSTRACTGuided by semantic space theory, large-scale computational studies have advanced our understanding of the structure and function of expressive behavior. I will integrate findings from experimental studies of facial expression (N=19,656), vocal bursts (N=12,616), speech prosody (N=20,109), multimodal reactions (N=8,056), and an ongoing study of dyadic interactions (N=1,000+). These studies combine methods from psychology and computer science to yield new insights into what expressive behaviors signal, how they are perceived, and how they shape social interaction. Using machine learning to extract cross-cultural dimensions of behavior while minimizing biases due to demographics and context, we arrive at objective measures of the structural dimensions that make up human expression. Expressions are consistently found to be high-dimensional and blended, with their meaning across cultures being efficiently conceptualized in terms of a wide range of specific emotion concepts. Altogether, these findings generate a comprehensive new atlas of expressive behavior, which I will explore through a variety of visualizations. This new taxonomy departs from models such as the basic six and affective circumplex, suggesting a new way forward for expression understanding and sentiment analysis.

Title: Industrial Use Cases: 3D Connectivity for Digital Twins: Decoupling 3D data utilization from delivery and file formats on an infrastructure level. Abstract: ABSTRACT With the rapidly growing size of 3D data sets as well as configuration complexity e.g., of assembly structures, the way to an enterprise-wide utilization of 3D product data is already a difficult one. While the 3D Master for engineering & manufacturing is rather established, availability of 3D data along the full value chain e.g., as part of the Digital Thread, and the ability to easily relate it to other business information is still a huge challenge. On top of this, Digital Twins introduce the demand to dynamically link 3D data with IIoT, AI and other business information to visually bring such Digital Twins to live or even into Mixed Reality applications, adding another level of complexity. Building on established Web technology patterns, we experienced that an API-based harmonization over (brownfield) data backends and file formats can help to establish a "single source of truth" as a prerequisite for the required agile utilization of 3D data and highly flexible interconnectivity. Many solutions rely on explicitly exporting and preparing use case or application specific subsets of 3D data before making such available, often heavily limiting its utilization on the way. Instead, a high level of interconnectivity is achievable by combining a virtualization of 3D data and algorithms with a unified addressing scheme to abstract over file boundaries and formats while enabling dynamic resolution and mapping of contained 3D data elements with business information on many levels. In our talk, we will present the underlying approach and explain its utilization throughout several industrial use cases in the area of Digital Twins with a focus on dynamic linkage and enrichment of 3D product data across software, solution and format boundaries.

Title: Industrial Use-Case : Digital Twin for Autonomous Earthwork in Virtual-Reality Abstract: ABSTRACT While advances in autonomous robotics and earthwork automation are paving the way for the construction site of tomorrow by means of fleets of autonomous machinery, they also force the issue of the supervision of such machines when they get stuck or stop for safety. When traditional supervision approaches impose an operator to stay on site and physically reach into the paused machine to re-gain control, our method consists in controlling the machines remotely in VR by leveraging a digital twin of the machine in its environment, fulled by LiDAR mapping. Our approach also allows the supervisor to orchestrate the autonomous fleet by means of a high-level "bird-view" interface of the whole site. As an attempt to help designing the tools of future earthwork supervisors, our preliminary results demonstrate the feasibility of a VR digital-twin approach for single-machine control, coupled with a 2D fleet-level control interface.

Title: One-Dimensional Eye-Gaze Typing Interface for People with Locked-in Syndrome Abstract: ABSTRACT People with Locked-in syndrome (LIS) suffer from complete loss of voluntary motor functions for speech or hand-writing. They are mentally intact, retaining only the control of vertical eye movements and blinking. In this work, we present a one-dimensional typing interface controlled exclusively by vertical eye movements and dwell-time for them to communicate at will. Hidden Markov Model and Bigram Models are used as auto-completion on both word and sentence level. We conducted two preliminary user studies on non-disabled users. The typing interface achieved 3.75 WPM without prediction and 11.36 WPM with prediction.

Title: Little Garden: An augmented reality game for older adults to promote body movement Abstract: ABSTRACT Physical activity is one of the most effective ways to help older adults stay healthy, but traditional training methods for older adults use single tasks and are boring, often making it difficult for the elderly to achieve good exercise results. In contrast to existing digital games, games based on augmented reality technology have the potential to promote physical activity in the elderly. This paper presents Little Garden, an interactive augmented-reality game designed for older adults. It uses projective augmented reality technology, physical card manipulation, virtual social scenarios to increase user engagement and motor initiation. The pilot data show that the game system promotes physical engagement and provides a good user experience. We believe that augmented reality technology provides a new approach to interface design for age-appropriate user-interface experiences.

Title: FormSense: A Fabrication Method to Support Shape Exploration of Interactive Prototypes Abstract: ABSTRACTWhen exploring the shape of interactive objects, existing prototyping methods can conflict with the iterative process. In this paper, we present FormSense: a simple, fast, and modifiable fabrication method to support the exploration of shape when prototyping interactive objects. FormSense enables touch and pressure sensing through a multi-layer coating approach and a custom touch sensor built from commodity electronic components. We use FormSense to create four interactive prototypes of diverse geometries and materials.

Title: Digital Twin and 3D Web-based Use Cases in Industry Abstract: ABSTRACT Multi-physical modeling combined with data-driven decision making is giving rise to a new paradigm, the "digital twin." The digital twin is a living digital model of a system or physical asset that continuously adapts to operational changes based on real-time data. When properly designed, a digital twin can help predict the future behavior of its corresponding physical counterpart. This paper presents a series of use cases that illustrate the role of a digital twin in different stages of the industrial product lifecycle. The use cases are implemented using 3D web technology for user interfaces and web standards (X3D and glTF) for data exchange between modules. The contribution of this work consists of a set of lessons learnt and some hints on future synergies between digital twin and 3D web technologies.

Title: Availability of Voice Deepfake Technology and its Impact for Good and Evil Abstract: ABSTRACTArtificial Intelligence and especially Machine Learning and Deep Learning techniques are increasingly populating today's technological and social landscape. These advancements have overwhelmingly contributed to the development of Speech Synthesis, also known as Text-To-Speech, where speech is artificially produced from text by means of computer technology [1]. But currently, there is a fundamental common drawback: unnatural, robotic and impersonal synthesized voices [2]. So, what happens when the robotic computer voice no longer sounds like a computer, but sounds like you? That's where Voice Cloning technology comes into play, which allows one to generate an artificial speech that resembles a targeted human voice. This new practice offers many benefits, but with its development, the generation of fake voices and videos, known as “deepfakes”, has risen, causing a loss of trust and greater fear towards technology [3]. In this way, the objective of this paper is to analyze the availability of voice deepfake technologies, its ease of construction and its impact for good and evil. We chose to focus on the educational field by implementing a “deepfake professor” via a survey of readily available voice deepfake technologies. The goal is then to demonstrate the potential capabilities for good and for evil that need to be considered with this technology, so we also conduct an analysis about the misuse, the current regulation, and the future of it. The results of the case study show that it is possible to clone someone's voice with a standard laptop, with no need of high-performance computing resources and based on just a few seconds of reference audio, which creates a superior user experience, but at the same time, reveals how easily can anyone have access to voice cloning. This expresses very well the importance of the new challenges opened by this potential technology and the need of safeguarding and regulation that future generations will have to deal with. There is no doubt that to understand the dynamics and impact of voice cloning and to reach more solid conclusions, future research is needed.

Title: LUNAChair: Remote Wheelchair System that Links Up a Remote Caregiver and Wheelchair Surroundings Abstract: ABSTRACTWe introduce LUNAChair, a remote control and communication system that uses omnidirectional video to connect a remote caregiver to a wheelchair user and a third person around the wheelchair. With the recent growing need for wheelchairs, much of the wheelchair research has focused on wheelchair control, such as fully automatic driving and remote operation. For wheelchair users, conversations with caregivers and third persons around them are also important. Therefore, we propose a system that connects a wheelchair user and a remote caregiver using omnidirectional cameras, which allows the remote caregiver to control the wheelchair while observing both the wheelchair user and his/her surroundings. Moreover, the system facilitates communication by using gaze and hand pointing estimation from an omnidirectional video.

Title: Operationalizing Computing Verb Enhancements to Bloom's Revised Taxonomy: Leveraging Bloom's for Computing to Create Learning Outcomes for Various Computing Disciplines Abstract: ABSTRACTLearning outcomes serve as reference points for assessing course success, for building prerequisite structures, for determining credit transferability to other institutions, and for ensuring program completion. Technical disciplines have found it more difficult to design learning outcomes because discipline-specific terminology has not been compatible with the existing verbs listed in Bloom's Revised Taxonomy. Computing faculty representing different disciplines, including Information Technology, Cybersecurity, and Computer Science, will serve as panelists to explain their approach to creating effective learning outcomes using the Bloom's for Computing list of computing-related verbs to enhance Bloom's Revised Taxonomy.

Title: Methods of Gently Notifying Pedestrians of Approaching Objects when Listening to Music Abstract: ABSTRACT Many people now listen to music with earphones while walking, and are less likely to notice approaching people, cars, etc. Many methods of detecting approaching objects and notifying pedestrians have been proposed, but few have focused on low urgency situations or music listeners, and many notification methods are unpleasant. Therefore, in this work, we propose methods of gently notifying pedestrians listening to music of approaching objects using environmental sound. We conducted experiments in a virtual environment to assess directional perception accuracy and comfort. Our results show the proposed method allows participants to detect the direction of approaching objects as accurately as explicit notification methods, with less discomfort.

Title: Visual Rehabilitation for Learning Disorders in Virtual Reality: Visual Rehabilitation for Learning Disorder in VR Abstract: ABSTRACTCurrent dyslexia rehabilitations methods, although efficient, suffer from the lack of adherence from young patients due to their repetitive and arduous tasks. Digital Therapeutics (DT) have grown exponentially in the last decade, and could be a stepping stone for dyslexia therapy. Making full use of new technologies, they offer new treatments for various disorders. The advancement and diffusion of Virtual Reality (VR) technologies are a new step in the therapeutic domain, notably for the treatment of neurological troubles. In this paper we propose a hybrid VR interface using eye-tracking (ET) and Brain-Computer Interface (BCI) with a gamified application for the rehabilitation of dyslexia. This prototype was designed in collaboration with medical professionals to create a gamified set of exercises adapted in 3D for dyslexia rehabilitation. The interface VR-ET-BCI serves as a monitoring device for the patient and a therapy evaluator for the practitioner. As of today, it lacks yet the clinical trials to show validated results, but an increase in motivation and adherence to therapy is expected.

Title: Summer Research with Undergraduate Students: A Multi-Thread Design Abstract: ABSTRACT In recent years, many universities have made it a priority to provide students with more opportunities for undergraduate research. This growth has been particularly notable in STEM, with undergraduate research appearing in a variety of forms (e.g., as part of a course, as a multi-day event, or as a summer-focused project). Some endeavors, like the National Science Foundation’s Research Experiences for Undergraduates (REU) program, are large-scale ventures, while others are localized. Regardless of scope, the benefits of adopting the recognized high-impact practice of undergraduate research are multifold. First, students who are exposed to state-of-the-art research can acquire research and professional skills; increase their knowledge, interest, and confidence in the subject; and improve their academic performance. Second, undergraduate research fosters the professional development of faculty advisors by providing a stage for sharpening skills in research guidance and advising, as well as working collaboratively with students to advance their work. Third, it benefits the field of research at large by increasing student retention and persistence, promoting recruitment for graduate programs, and diversifying the demographics of the field. These benefits are well evidenced in the literature. A successful undergraduate research experience, however, relies on a design that accounts for the particular constraints of the host institution. Like many other non-R1 institutions, we are limited in terms of budget, facilities, space, and faculty and student time to dedicate to undergraduate research projects, and we also face the uncertainty of short-term research (due to its experimental nature). In this work, we present our design for a summer-focused undergraduate research (SUR) program in CS/IT that has been implemented on a regular basis over the past decade. The program aims to provide the key features of a well-rounded graduate research experience, scaled down to fit the constraints under which we are operating. Additionally, it serves as an opportunity to increase participation in CS/IT by students from under-represented minority groups and to facilitate collaboration with faculty and students in other STEM fields. In addition to presenting the design, we discuss the outcomes, summarize the design strategies, and share lessons learned from the program, with the hope of motivating the development of similar endeavors at other institutions.

Title: Early IT Week: The Expansion of Stakeholder Engagement for a Growing Early IT High School Partnership Program Abstract: ABSTRACT Expansion of the Early Information Technology ecosystem brings the challenge of supporting and engaging stakeholders while maintaining a high-quality partnership program. Utilizing the strategies outlined in the original design initiatives, the Early IT Support Team can effectively disseminate through infectious engagement, promote practice of skill, and seek continuous improvement as it sets goals to expand opportunities for students while developing the next generation of IT talent.

Title: AIx speed: Playback Speed Optimization using Listening Comprehension of Speech Recognition Models Abstract: ABSTRACT In recent years, more and more time has been spent watching videos for online seminars, lectures, and entertainment. In order to improve time efficiency, people often adjust the playback speed to a speed that suits them best. However, it is troublesome to adjust the optimal speed for each video and even more challenging to change and adjust the speed for each speaker within a single video. Therefore, we propose ”AIx speed,” a system that maximizes the playback speed within the range where the speech recognition model can recognize and flexibly adjusts the playback speed for the entire video. This system makes it possible to set a flexible playback speed that balances playback time and content comprehension, compared to fixing the playback speed for the entire video.

Title: WireSketch: Bimanual Interactions for 3D Curve Networks in VR Abstract: ABSTRACT 3D content authoring in immersive environments has the advantage of allowing users to see a design result on its actual scale in real time. We present a system to intuitively create and modify 3D curve networks using bimanual gestures in virtual reality (VR). Our system provides a rich vocabulary of interactions in which both hands are used harmoniously following simple and intuitive grammar, and supports comprehensive manipulation of 3D curve networks.

Title: Over-The-Shoulder Training Between Redundant Wearable Sensors for Unified Gesture Interactions Abstract: ABSTRACT Wearable computers are now prevalent, and it is not uncommon to see people wearing multiple wearable devices. These wearable devices are often equipped with sensors to detect the user’s interactions and context. As more devices are worn on the user’s body, there is an increasing redundancy between the sensors. For example, swiping gestures on a headphone are detected by its touch sensor, but the movement it caused can also be measured by the sensors in a smartwatch or smart rings. We present a new mechanism to train a gesture recognition model using redundant sensor data so that measurements from other sensors can be used to detect gestures performed on another device even if the device is missing. Our preliminary study with 13 participants revealed that a unified gesture recognition model for touch gestures achieved accuracy for 25 gestures (5 gestures × 5 scenarios) where gestures were trained by leveraging the available sensors.

Title: Exploring the Internet of Things: Hands-On IoT Learning Mapped to the IT Pillars Abstract: ABSTRACTThe Internet of Things (IoT) is transforming our homes, businesses, and other enterprises into digital domains with a presence on the Internet. Increasingly enterprises are adopting IoT products, and they will employ individuals to design and support their IoT infrastructures, analyze their data and respond to the benefits and challenges of IoT. During the months of May and June 2022, the Department of Integrated Information Technology (IIT) within the College of Engineering and Computing at the University of South Carolina designed and taught a special topics course focused on IoT from the perspective of the IT Pillars – networking, programming, databases, HCI, and web systems. The course's focus on hands-on learning labs mapped to the IT Pillars distinguished this course from other IoT courses offered by the University of South Carolina. Student performance and feedback demonstrated that the course structure and content, including hands-on labs and projects, was an effective way to introduce students to IoT from the broad perspective of IT applied to real-world applications. Another benefit was a greater appreciation of the depth, scope, and impact of IoT on each of the IT Pillars and its importance in our curricula as we prepare students for their professional careers.

Title: Leveling Up: EmTech Education Providing New Pathway Opportunities in IT for Women at Any Stage of Their Careers Abstract: ABSTRACTTech industry, especially, some areas within tech fields, such as Emerging Technology (EmTech), like cybersecurity, machine learning, AI, and cloud computing, are expected to experience immense increases in job opportunities in coming years. While a variety of solutions are necessary to address the growing workforce needs in the EmTech industry, one of the largest untapped talent pools is women and underrepresented students. The panel postulates the labor needs for more women in Tech could be met by actively nurturing and enticing recent graduates, early career professionals, reskilled and returning career professionals. This panel offers a discussion platform that can provide insight into the stories of Women in EmTech and further research how to improve diversity in the technology education and industry.

Title: Towards Semantically Aware Word Cloud Shape Generation Abstract: ABSTRACT Word clouds are a data visualization technique that showcases a subset of words from a body of text in a cluster form, where a word’s font size encodes some measure of its relative importance—typically frequency—in the text. This technique is primarily used to help viewers glean the most pertinent information from long text documents and to compare and contrast different pieces of text. Despite their popularity, previous research has shown that word cloud designs are often not optimally suited for analytical tasks such as summarization or topic understanding. We propose a solution for generating more effective visualization technique that shapes the word cloud to reflect the key topic(s) of the text. Our method automates the processes of manual image selection and masking required from current word cloud tools to generate shaped word clouds, better allowing for quick summarization. We showcase two approaches using classical and state-of-the-art methods. Upon successfully generating semantically shaped word clouds using both methods, we performed preliminary evaluations with 5 participants. We found that although most participants preferred shaped word clouds over regular ones, the shape can be distracting and detrimental to information extraction if it is not directly relevant to the text or contains graphical imperfections. Our work has implications on future semantically-aware word cloud generation tools as well as efforts to balance visual appeal of word clouds with their effectiveness in textual comprehension.

Title: Hermod: principled and practical scheduling for serverless functions Abstract: ABSTRACTServerless computing has seen rapid growth due to the ease-of-use and cost-efficiency it provides. However, function scheduling, a critical component of serverless systems, has been overlooked. In this paper, we take a fist-principles approach toward designing a scheduler that caters to the unique characteristics of serverless functions as seen in real-world deployments. We first create a taxonomy of scheduling policies along three dimensions. Next, we use simulation to explore the scheduling policy space and show that frequently used features such as late binding and random load balancing are sub-optimal for common execution time distributions and load ranges. We use these insights to design Hermod, a scheduler for serverless functions with two key characteristics. First, to avoid head-of-line blocking due to high function execution time variability, Hermod uses a combination of early binding and processor sharing for scheduling at individual worker machines. Second, Hermod is cost, load, and locality-aware. It improves consolidation at low load, it employs least-loaded balancing at high load to retain high performance, and it reduces the number of cold starts compared to pure load-based policies. We implement Hermod for Apache OpenWhisk and demonstrate that, for the case of the function patterns observed in real-world traces, it achieves up to 85% lower function slowdown and 60% higher throughput compared to existing production and state-of-the-art research schedulers.

Title: A Comparison of Students’ Learning Behaviors and Performance Among Pre, During and Post COVID-19 Pandemic Abstract: ABSTRACT The advent of the COVID-19 pandemic fundamentally changed humans’ lifestyle, especially in the education sector. Educators and learners have to shift from traditional face-to-face learning to online or e-learning. Several novel online learning technologies were promoted during pandemic, and the corresponding advantages and disadvantages were analyzed and pointed out by existing research. However, there is limited research which compare students’ learning behaviors or performance among pre, during and post pandemic periods. In this paper, we performed a quantitative analysis and comparison to reveal the patterns on students’ learning behaviors or performance during different pandemic periods in information technology educations. More specifically, students’ behaviors on assignment submissions and their final grades were extracted and analyzed for the purpose of quantitative comparison. Our experimental results discover significant impacts on students’ learning by the COVID-19 pandemic in information technology educations, which may further benefit different stakeholders in the educational community in their future development.

Title: Bouncing Forward from COVID in Higher Education Abstract: ABSTRACT This paper is a call to arms to bounce forward in the classroom as we emerge from the COVID crisis. The predominant return to in-person classes in higher education should not be a return to the same normal classroom conditions that existed prior to the pandemic. In the last 2+ years, we have come an extraordinarily long way in our abilities and in our inclinations to employ technologies and techniques in a blended classroom environment that truly improves the learning experience. In this paper, we call for and contribute to such an effort. Tying into the abundance of literature dealing with the COVID educational environment, we present our findings and ideas from carefully studying our own faculty. We summarize our overall findings as well as describe in detail three general categories that we believe hold great promise for improving the higher education classroom in the post-crisis era, namely digital chalkboards / screen sharing; remote participation and collaboration; and a paperless classroom. We argue that educators have an obligation and opportunity to not simply return to pre-crisis methods.

Title: Gustav: Cross-device Cross-computer Synchronization of Sensory Signals Abstract: ABSTRACT Temporal synchronization of behavioral and physiological signals collected through different devices (and sometimes through different computers) is a longstanding challenge in HCI, neuroscience, psychology, and related areas. Previous research has proposed to synchronize sensory signals using (1) dedicated hardware; (2) dedicated software; or (3) alignment algorithms. All these approaches are either vendor-locked, non-generalizable, or difficult to adopt in practice. We propose a simple but highly efficient alternative: instrument the stimulus presentation software by injecting supervisory event-related timestamps, followed by a post-processing step over the recorded log files. Armed with this information, we introduce Gustav, our approach to orchestrate the recording of sensory signals across devices and computers. Gustav ensures that all signals coincide exactly with the duration of each experiment condition, with millisecond precision. Gustav is publicly available as open source software.

Title: Tie Memories to E-souvenirs: Hybrid Tangible AR Souvenirs in the Museum Abstract: ABSTRACT Traditional physical souvenirs in museums have three major limits: monotonous interaction, lack of personalization, and disconnection to the exhibition. To conquer these problems and to make personalized souvenirs a part of the visiting experience, we create a hybrid tangible Augmented Reality(AR) souvenir that combines a physical firework launcher and AR models. An application called AR Firework is designed for customizing the hybrid souvenir as well as interactive learning in an exhibition in the wild. Multiple interaction methods including mobile user interface, hand gestures, and voice are adopted to create a multi-sensory product. As the first research to propose tangible AR souvenirs, we find that they establish a long-lasting connection between visitors and their personal visiting experiences. This paper promotes the understanding of personalization, socialization and tangible AR.

Title: It’s Great to Be Back: An Experience Report Comparing Course Satisfaction Surveys Before, During and After Pandemic Abstract: ABSTRACT University and college level course satisfaction is usually appraised through feedback surveys. Course satisfaction items and comments commonly orbit around course content and quality of teaching or teaching staff abilities. With the COVID-19 outbreak, teaching strategies and course delivery tools were forced to change, bringing up considerations that previously were not necessarily regarded as key elements for course evaluation. After almost two years of fully remote courses, the declining of the pandemic allowed several institutions to return to in-person activities, bringing back the traditional in-person course format. With the closure of the first courses carried out again in person, it is possible to review course satisfaction surveys in a post-pandemic environment, and gain insight about the trends, changes, and evolution brought by these difficult times. In this experience report, we discuss the results of the course satisfaction survey of an undergraduate Software Engineering course through a time span of three years that includes pre-pandemic, pandemic, and post-pandemic conditions. Results show that even though by the beginning of the pandemic the course satisfaction level were kept high, as the pandemic aged there was a mild declining trend on course satisfaction and general student engagement. These indicators came back to their traditional levels as the institution brought back in-person courses. The goal of this experience report is to call attention to the production of research works that assist on understanding what are the elements that were relevant for course satisfaction before and during the pandemic, how these prevailed as after the outbreak, and what indicators returned to their original, pre-pandemic standards.

Title: Towards Education 5.0: Instruction with Learners in the Loop Abstract: ABSTRACT Based on the current view of Education 4.0 and Industry 4.0 and developments in Industry 5.0, we extrapolate onto the resulting advances that might be following in Education 5.0: Both are presumed to focus on a reintegration of human factors and place human actors into controlling loops, which we introduce here.

Title: iMarker: Instant and True-to-scale AR with Invisible Markers Abstract: ABSTRACTAugmented Reality (AR) has been widely used in modern mobile devices for various applications. To achieve a stable and precise AR experience, mobile devices are equipped with various sensors (e.g., dual camera, LiDAR) to increase the robustness of camera tracking. Those sensors largely increased the cost of mobile devices and are usually not available on low-cost devices. We propose a novel AR system that leverage the advance of marker-based camera tracking to produce fast and true-to-scale AR rendering on any device with a single camera. Our method enables the computer monitor to be the host of AR markers, without taking up valuable screen space nor impacting the user experience. Unlike traditional marker-based methods, we utilize the difference between human vision and camera system, making AR markers to be invisible to human vision. We propose an efficient algorithm that allows the mobile device to detect those markers accurately and later recover the camera pose for AR rendering. Since the markers are invisible to human vision, we can embed them on any website and the user will not notice the existence of these markers. We also conduct extensive experiments that evaluate the efficacy of our method. The experimental results show that our method is faster and has a more accurate scale of the virtual objects compared to the state-of-the-art AR solution.

Title: National Cybersecurity Curriculum Task Force: How you can contribute to and benefit from high-quality, high-impact cybersecurity curriculum Abstract: ABSTRACTFunded by the NSA through the NCAE-C program, the mission of the National Cybersecurity Curriculum Task Force is to catalog and create high-quality and relevant curricula on emerging cybersecurity topics, mapping to curricular and workforce guidelines, and make them freely available. This lightning talk will share the progress of the project so far, point the audience to where they can find vetted cybersecurity curriculum materials, and offer opportunities to contribute to the project.

Title: Summer Computing Camp to Compare and Contrast CS/IT/CE Programs Abstract: ABSTRACTHigh school students need to better understand the differences between majors in computing fields. Misunderstanding the differences may lead students to drop out or switch majors, harming retention rates and degree completion times. This interdisciplinary team is developing a week-long camp curriculum for high school students to promote success in selecting a computing major. Evaluation methods for the program aim to identify which elements are most impactful in helping participants choose disciplines of study; and the camp will assist students in effectively preparing for college education based on their personal preferences, experiences, and high school coursework.

Title: Rapid Aspect-Oriented Assessment of Relational Database Design Assignments Abstract: ABSTRACT It can be argued that assignments in a Databases course, where students are required to build a database instance based on specifications provided in the form of Entity-Relationship Diagrams and descriptions, are one of the best ways to evaluate students’ practical knowledge because they simultaneously cover a variety of topics and cognitive levels. Organizing an exam with such assignments is a time-consuming process, even more so in a large course. This work proposes a rapid aspect-oriented assessment process and presents a software system that automates most of the tasks needed to administer and assess an exam that includes such assignments.

Title: IT Baccalaureate Programs and Program Leaders in the U.S.: A Profile, Characteristics, and Perceptions Abstract: ABSTRACTThe number of information technology baccalaureate programs continues to grow to meet the demand for rapid growth in IT occupations. With the growth of a relatively new academic discipline, it is important to identify its players, namely IT faculty and administrators, and the programmatic landscape. To gather data about IT programs, faculty, and administrators, a survey was conducted in spring 2021. This paper is presents survey data related to the profiles and perceptions of IT program leaders and characteristics of IT baccalaureate programs in the U.S.

Title: Achieving low latency in public edges by hiding workloads mutual interference Abstract: ABSTRACTOn multi-tenant platforms, such as public clouds and edges, workloads interfere with each other through shared resources. The performance degradation caused by such interference is a notoriously challenging problem. Though many solutions have been proposed for clouds, they can hardly help the application in edges, where workloads are mostly latency-critical, highly dynamic, and more sensitive to interference. Aggressive resource over-provisioning looks to be the only practical solution, albeit it causes significant resource waste. The paper proposes dynamic asymmetric scheduling for edge computing (DASEC) as a unique approach to achieve low latency in public edges and improve resource utilization. DASEC makes application performance less sensitive to the interference between workloads by making the interference affect mostly the tasks on non-critical paths and rarely the tasks on critical paths. With DASEC, the interference is largely hidden from being reflected on the end-to-end performance observed by users. The paper has investigated the techniques to implement DASEC in the task schedulers for edge workloads and tested its effectiveness in managing the interference caused by sharing CPU cores. For different types of edges that schedule tasks at different system levels, the paper implemented DASEC prototypes based on Linux/KVM vCPU scheduler, the completely fair scheduler (CFS) in Linux OS, and Google user-level scheduling framework. Extensive experiments with diverse real-world applications show that DASEC can reduce the latencies of the workloads consolidated on the same edge server by 32% ~ 52%.

Title: TaskScape: Fostering Holistic View on To-do List With Tracking Plan and Emotion Abstract: ABSTRACT Despite advancements with intelligence and connectivity in the workspace, productivity tools, such as to-do list applications, still, measure workers’ performance by a binary state—completed, yet completed, and thus the number of tasks completed. Such quantitative measurements can often overlook human values and individual well-being. While concepts such as positive computing and digital well-being are on the rise in the HCI community, few systems have been proposed to effectively integrate holistic considerations for mental and emotional well-being into productivity tools. In this work, we depart from the classic task list management tool and explore the construction of well-being-centered to-do list software. We propose a task management system–TaskScape—, which allow users to have awareness on the following two aspects: (1) how they plan and complete tasks and (2) how they feel towards their work. With the proposed system, we will investigate if having holistic view on their tasks can facilitate reflection on what they work on, how they stick to their plans, and how their tasks portfolio support their emotional well-being, nudging users to reflect upon their work, planning performance, and their emotional values towards their work. In this poster, we share the design, development, and ongoing validation progress of TaskScape, which is aimed to nudge workers to holistically view work productivity, reminding users that work is more than just work but life.

Title: TrackItPipe: A Fabrication Pipeline To Incorporate Location and Rotation Tracking Into 3D Printed Objects Abstract: ABSTRACTThe increasing convergence of the digital and physical world creates a growing urgency to integrate 3D printed physical tangibles with virtual environments. A precise position and rotation tracking are essential to integrate such physical objects with a virtual environment. However, available 3D models commonly do not provide tracking support on their composition, which requires modifications by CAD experts. This poses a challenge for users with no prior CAD experience. This work presents TrackItPipe, a fabrication pipeline supporting users by semi-automatically adding tracking capabilities for 3D printable tangibles tailored to environmental requirements. TrackItPipe integrates modifications to the 3D model, produces the respective tangibles for 3D printing, and provides integration scripts for Mixed Reality. Using TrackItPipe, users can rapidly equip objects with tracking capabilities.

Title: Using a Professional Skills Framework to Support the Assessment of Dispositions in IT Education Abstract: ABSTRACT The IT2017 ACM/IEEE Curriculum Guidelines for Baccalaureate Degree Programs introduced a significantly new way of framing computing education curricula. Instead of focusing on how to structure the Information Technology body of knowledge, the IT2017 report centered its curricular recommendations on the competencies that IT programs are expected to develop in their graduates. A major contribution of the report is a model of IT competency that identifies three interrelated components: content knowledge, skills, and dispositions, where dispositions represent personal qualities desirable in the workplace. Building on the IT2017 contributions, the ACM/IEEE Computing Curricula 2020 (CC2020) report enriches the disposition concept by identifying eleven dispositions, such as being adaptable or self-directed, that all computing programs should include in the career preparation of their graduates. While the importance of dispositions is widely appreciated by both academic programs and employers, how to develop and assess dispositions in a degree program remains a challenge. A recent mapping of the eleven CC2020 dispositions to the responsibility characteristics of the Skills Framework for the Information Age (SFIA) opened a promising route for addressing this challenge. Inspired by this mapping, this paper’s aim is to help educators assess students’ achievement of CC2020 dispositions through the SFIA responsibility characteristics. Our proposed assessment method and tool are based on evidence of performance of tasks in real-world work settings, demonstrating SFIA responsibility characteristics, which are mapped to CC2020 dispositions. Furthermore, this paper validates the selection of SFIA to operationalize the CC2020 dispositions, by demonstrating that other IT skills frameworks would pose significant challenges for our assessment approach.

Title: An Academic and Professional Profile of IT Faculty in the U.S. Abstract: ABSTRACTInformation technology (IT) as a university discipline continues grow and evolve. For that reason, IT educators and administrators are interested in understanding the academic and professional profile of IT faculty. A comprehensive survey of IT faculty was conducted in spring 2021, and this paper presents data that builds an academic and professional profile of both part time and full-time faculty in the U.S.

Title: Demystifying the Abstractness: Teaching Programming Concepts with Visualization Abstract: ABSTRACT The abstract nature of programming concepts in CS/IT courses causes challenges for undergraduate students to grasp them. Visualization tools can mitigate this challenge by providing a ”what you see is what you get” experience. Given the diversity of the tools and the fast pace of software updates, it remains a question to select the best-fit tools that can be of current use for the respective programming concepts. This study aims at providing an assessment and classification of visualization tools based on four defined metrics which suggests a hybrid approach for selecting combination of visualization tools in teaching.

Title: Virtual Cybersecurity Camps: Lessons Learned Abstract: ABSTRACT Cybersecurity camps provide participants with an opportunity to learn about cybersecurity in a fun and safe environment. Traditionally, such camps, like many others, are held in-person. However, the COVID-19 pandemic created unique challenges and also an opportunity to counter those challenges—holding a cybersecurity camp virtually. While countless other camps, both cybersecurity-related and others, moved to a virtual environment so that such camps could continue to be held, this paper presents some lessons learned and suggestions that may be helpful to others deciding to hold a virtual camp in the future. Some of the lessons learned may be specific to a cybersecurity camp, but most would be applicable to a broad audience.

Title: The Reflective Maker: Using Reflection to Support Skill-learning in Makerspaces Abstract: ABSTRACTIn recent years, while HCI researchers have developed several systems that leverage the use of reflection for skill learning, the use of reflection-based learning of maker skills remains unexplored. We present ReflectiveMaker - a toolkit for experts and educators to design reflection exercises for novice learners in makerspaces. We describe the three components of our toolkit: (a) a designer interface to author the reflection prompts during fabrication activities, (b) a set of fabrication tools to sense the user’s activities and (c) a reflection diary interface to record the user’s reflections and analyze data on their learning progress. We then outline future work and envision a range of application scenarios.

Title: Crowdsourcing Quality Concerns: An Examination of Amazon’s Mechanical Turk Abstract: ABSTRACT The use of crowdsourcing platforms, such as Amazon’s Mechanical Turk (MTurk), have been an effective and frequent tool for researchers to gather data from participants for a study. It provides a fast, efficient, and cost-effective method for acquiring large amounts of data for a variety of research projects, such as surveys that may be conducted to assess the use of information technology or to better understand cybersecurity perceptions and behaviors. While the use of such crowdsourcing platforms has gained both popularity and acceptance over the past several years, quality concerns remain a significant issue for the researcher. This paper examines these issues.

Title: SilentWhisper: faint whisper speech using wearable microphone Abstract: ABSTRACTVoice interaction is a fundamental human capacity, and we can use voice user interfaces just speaking. However, in public spaces, we are hesitant to use them because of consideration for their surroundings and low privacy. Silent speech, a method that recognizes the movement of speech in silence, has been proposed as a solution to this problem, and it allows us to maintain our privacy when speaking. However, existing silent speech interfaces are burdensome because the sensor must be kept in contact with the face and mouth, and commands must be prepared for each user. In this study, we propose a method to input whispered speech at a quiet volume that cannot be heard by others using a pin microphone. Experimental results show that a recognition rate was 13.9% WER and 6.4% CER for 210 phrases. We showed that privacy-preserving vocal input is possible by whispering voices which are not comprehensible to others.

Title: Enabling Open Source, Open Data for Closed Source, Closed Data Learning Management Systems Abstract: ABSTRACT Learning analytics has attracted significant interest for research as well as implementation in education systems. Similar to other analytical approaches to improve underlying processes, the availability of raw data (open-data) of the learning process is paramount. Despite several open-sourced Learning Management System (LMS) solutions that enable an open access to data for learning analytics by/for stakeholders, commercially provided solutions prevalent in the education system in the U.S. are typically closed-source and closed-data. In this contribution, we provide an approach to enable open-sourced, open-data research and implementations within the confined environments of commercial closed-source, closed-data LMS implementations.

Title: LV-Linker: Supporting Fine-grained User Interaction Analyses by Linking Smartphone Log and Recorded Video Data Abstract: ABSTRACT Data-driven mobile design as an important UI/UX research technique often requires analyzing recorded screen video data and time-series usage log data, because it helps to obtain a deeper understanding of fine-grained usage behaviors. However, there is a lack of interactive tools that support simultaneously navigation of both mobile usage log and video data. In this paper, we propose LV-Linker (Log and Video Linker), a web-based data viewer system for synchronizing both smartphone usage log and video data to help researchers quickly to analyze and easily understand user behaviors. We conducted a preliminary user study and evaluated the benefits of linking both data by measuring task completion time, helpfulness, and subjective task workload. Our results showed that offering linked navigation significantly lowers the task completion time and task workload, and promotes data understanding and analysis fidelity.

Title: Serving unseen deep learning models with near-optimal configurations: a fast adaptive search approach Abstract: ABSTRACTPublic clouds provide a bewildering choice of configurations for Deep Learning (DL) models, and the choice of configuration will significantly impact the performance and budget. However, it is an obvious challenge to recommend a near-optimal configuration for a particular DL model from a wide range of candidates. The huge search overhead of finding such a configuration is the notorious cold start problem in state-of-the-art efforts, and this problem becomes more severe when they are faced with unseen DL models. In this paper, we present Falcon, a novel configuration recommender system that can quickly adapt to unseen DL models. Through a large-scale evaluation, we find that there are some Key Operators (KOPs) that can be used to estimate the performance of DL models, and their resource sensitivity can be represented by four typical Key Operator Resource Curves (KOP-RCs). This work can effectively alleviate the cold start problem, because an unseen DL model can be characterized by its KOPs and corresponding KOP-RCs, and this characterization can be constructed as a tree structure in which near-optimal configurations can be searched quickly through a combination of Monte Carlo Tree Search and Bayesian optimization (MCTS-BO). Experiments show that Falcon can effectively reduce the search overhead for unseen DL models by up to 80% compared to state-of-the-art efforts.

Title: The Future of the Body in Tomorrow’s Workplace Abstract: ABSTRACT Even in the most hectic or time-conscious workplace, employees gather in person to chat. And even in the most networked workplace, employees still make time for face-to-face collaboration. This isn’t surprising if we consider that the ability and desire to engage in face-to-face communication (using eye gaze to manage turn-taking, head nods to indicate listening, and smiles to indicate attention) starts soon after birth – well before infants even learn to talk. We might say that we are built to communicate face-to-face! But what role will embodied interaction play in the future workplace, when we will be interacting with autonomous robots, engaging with other people through presence robots, and working in a virtual world where we and our colleagues are represented by avatars? In this talk I will describe some of the ways that embodied interaction is likely to change in the future, and some of the ways that we need to take those future scenarios into account as we design and implement multimodal interfaces.

Title: Focus on People: Five Questions from Human-Centered Computing Abstract: ABSTRACT A substantial body of research in multimodal interaction has studied how people naturally interact –face-to-face and through machines– and developed technology to analyze, support, and extend such forms of interaction. The talk will share personal experiences and views on how audio-visual and ubiquitous research on social interaction has evolved over the past two decades. Five recurrent questions, then and now, include how to study interaction in everyday life; how to learn from and collaborate with the humanities and social sciences; how to think about data; how to address the challenges brought by automation; and how to engage and empower individuals and communities to take part in research projects. Today, the limitations of technology-centric solutions are more evident than ever. Future research with a people-first focus will continue to call for reflection, commitment, and action for a long-term alignment with societal needs and nature’s limits.

Title: Introducing Zero Trust in a Cybersecurity Course Abstract: ABSTRACTZero trust (ZT) is a conceptual and architectural model for cybersecurity teams to design networks into secure micro-perimeters and strengthen data security by systematically integrating state-of-the-art technology, risk management, and threat intelligence. ZT has recently gained momentum in the industry to defend against lateral movement of malicious actors in today’s borderless networks. The United States 2021 President Executive Order requires the federal government must adopt security best practice and advance toward a zero trust architecture (ZTA). However, it is not a trivial task to implement a ZTA due to its novelty and complexity. We need to understand what ZTA is to take the advantage of it. Therefore, there is a need to introduce the fundamental concepts, principles, and architectures of ZT in cybersecurity courses at a college to better prepare our new cybersecurity professionals for their careers. We have introduced ZT in a cybersecurity course for senior undergraduates and another course for graduate students. This article provides an overview of the materials we have used to introduce ZT in both courses, including the problems in a traditional perimeter-based security model and how these problems can be either resolved or mitigated with a ZT security model. We expect our work will serve as a good reference for educators to introduce ZT security model in a cybersecurity course.

Title: Detecting Changes in User Emotions During Bicycle Riding by Sampling Facial Images Abstract: ABSTRACT In the context of mobility as a Service (MaaS), bicycles are an important mode of transport for the first and last mile between the home and other transport modalities. Also, due to covid-19 bicycle users such as food delivery drivers and commuters to work are increasing. To investigate driving experience of bicycle users in context and improve MaaS service quality, we propose and describe a method to automatically detect changes in user emotions during bicycle riding by sampling facial images using a smartphone. We describe the proposed method and how we plan to use it in the future.

Title: Student Engagement during Virtual v.s. Face-To-Face Active Learning Activities in Three IT Courses Abstract: ABSTRACT The post COVID-19 landscape of higher education has accelerated the adoption of flexible instructional modalities that blend online synchronous and in-person face-to-face teaching. Building a virtual active learning classroom that emphasizes higher-order thinking, problem-solving and collaborative programming through group work is particularly challenging. This study discusses active learning strategies in three IT courses delivered both in-person and virtually through online synchronous video-conference. Using various software tools and a dynamic breakout room strategy, we were able to create an effective virtual active learning environment. We also surveyed the students to measure two engagement factors - Value of Activity and Personal Effort, to compare their experiences in the virtual and face-to-face collaborations. The quantitative results showed no significant difference in either of the two factors between the two modalities. The qualitative results also confirmed that most students enjoy active learning in both settings, although some found in-person group work more interactive and fun.

Title: Real Talk, Real Listening, Real Change Abstract: ABSTRACT In an era of rising toxic polarization, political intolerance, and plummeting social trust, the need to listen to one another could not be greater. For all the promise of social media to give everyone a voice, in reality it is the loudest and most polarizing voices that tend to dominate. Our research and development teams at the MIT Center for Constructive Communication and Cortico are developing tools and methods designed to foster authentic conversation and scalable deep listening by merging age-old human practices of facilitated dialogue with modern methods of digital design, speech and language processing, and AI-powered data science. We have partnered with field organizations – ranging from small community organizations to municipalities to global nonprofits – to make sense of the conversations they collect, amplify typically underheard voices, inform public understanding, drive better policy and decisions, and enable unforeseen connections across the ideological spectrum. In this talk I will provide an overview of the approach, highlight some case studies, and sketch open research questions motivated by the work.

Title: What is Multimodal? Abstract: ABSTRACT Our experience of the world is multimodal: we see objects, hear sounds, feel texture, smell odors, and taste flavors. In recent years, a broad and impactful body of research emerged in artificial intelligence under the umbrella of multimodal, characterized by multiple modalities. As we formalize a long-term research vision for multimodal research, it is important to reflect on its foundational principles and core technical challenges. What is multimodal? Answering this question is complicated by the multi-disciplinary nature of the problem, spread across many domains and research fields. This talk is based on a recent review of 700+ research papers, to study computational and theoretical foundations for multimodal research, with a focus on multimodal machine learning. Two key principles have driven many multimodal innovations: heterogeneity and interconnections from multiple modalities. Historical and recent progress will be synthesized in a research-oriented taxonomy, centered around 6 core technical challenges: representation, alignment, reasoning, generation, transference, and quantification. The talk will conclude with open questions and unsolved challenges essential for a long-term research vision in multimodal research.

Title: HomeView: Automatically Building Smart Home Digital Twins With Augmented Reality Headsets Abstract: ABSTRACT Digital twins have demonstrated great capabilities in the industrial setting, but the cost of building them prohibits their usage in home environments. We present HomeView, a system that automatically builds and maintains a digital twin using data from Augmented Reality (AR) headsets and Internet of Things (IoT) devices. We evaluated the system in a simulator and found it performs better than the baseline algorithm. The user feedback on programming IoT devices also suggests that contextual information rendered by HomeView is preferable to text descriptions.

Title: Pressure Test: Finding Appropriate Data Size for Practice in Data Science Education Abstract: ABSTRACT Data science, such as data analytics, data mining, machine learning, became one popular curriculum in information technology educations. The lectures on these topics cannot stand alone without coding practice on real-world data sets. Some instructors prefer to utilize small data sets for practice in classroom or assignments, which limits experimental experiences and may even bring misleading experiences to students. Others may try to assign large data sets to students, but students may not be able to bear with the running time due to the efficiency issue raised by several factors (e.g., data size, algorithm complexity, computing power, etc.). In this paper, we first learned students’ preferences on the scalability of data sets for practice in data science courses, and performed experimental analysis by running different data science algorithms over both student laptops and personal/office computers, in order to deliver a suggestion about the appropriate data size for practice in multiple scenarios (e.g., in-class practice, assignments, class projects, research projects, etc.). We believe that our findings are valuable to help instructors prepare and assign real-world data sets to students in data science curriculum.

Title: Self-Supervised Approach for Few-shot Hand Gesture Recognition Abstract: ABSTRACT Data-driven machine learning approaches have become increasingly used in human-computer interaction (HCI) tasks. However, compared with traditional machine learning tasks, for which large datasets are available and maintained, each HCI project needs to collect new datasets because HCI systems usually propose new sensing or use cases. Such datasets tend to be lacking in amount and lead to low performance or place a burden on participants in user studies. In this paper, taking hand gesture recognition using wrist-worn devices as a typical HCI task, I propose a self-supervised approach that achieves high performance with little burden on the user. The experimental results showed that hand gesture recognition was achieved with a very small number of labeled training samples (five samples with 95% accuracy for 5 gestures and 10 samples with 95% accuracy for 10 gestures). The results support the story that when the user wants to design 5 new gestures, he/she can activate the feature in less than 2 minutes. I discuss the potential of this self-supervised framework for the HCI community.

Title: Building a Growth Mindset Toolkit as a Means Toward Developing a Growth Mindset for Faculty Interactions with Students In and Out of the Classroom: Building a Growth Mindset Toolkit for Faculty Abstract: ABSTRACTDuring the spring semester of the 2021 academic year, a group of faculty gathered as part of a Growth Mindset Faculty Community of Practice (GM-FCoP) to understand how to use a growth mindset to positively impact students in their courses, through mentoring and in daily conversations. Grounded in Carol Dweck's seminal works on theories of intelligence, a growth mindset asserts that skills can be developed over time and views challenges as opportunities for growth and future success. This contrasts with a fixed mindset which views skills as set at birth with little hope for development. This notion of a fixed mindset also contrasts with the essence of our work as faculty and educators where we strive daily to positively influence our students’ successes, learning and skill development. Yet, embracing a growth mindset over a fixed mindset can prove challenging. Learning about a growth mindset serves as an effective starting point for faculty, with next steps revolving around actively generating their own knowledge toward an overarching goal of applying the growth mindset concepts in their coursework as well as while mentoring students. This experience paper outlines the GM-FCoP's creation of a Growth Mindset Toolkit to serve as a resource for faculty as they foster and promote a growth mindset with students in formal settings such as in the classroom and in mentoring sessions, as well as informal settings such as office hours and general conversations and interactions. Faculty developed approaches for a growth mindset are highlighted along with leadership reflections and next steps.

Title: Straight From the Human Factors Professionals’ Mouth: The Need to Teach Human Factors in Cybersecurity Abstract: ABSTRACTColleges and universities are vital for integrating the human factors discipline in cybersecurity courses. Human errors, limitations, and weaknesses contribute to data breaches, ransomware attacks, and cyber-attacks. The cybersecurity community struggles to leverage human factors as a scientific discipline to eradicate human-related issues. Problems regarding the human element in cybersecurity require curricula focusing on the scientific aspects of human factors to teach human factors principles. As people and technology are increasingly interdependent, students interested in pursuing careers in technology must have a foundational understanding of people and their interactions with technology. While human factors is the discipline, human factors engineering is the work of leveraging human factors principles to improve the integration between humans and systems. Through various venues, educational initiatives, and research, academia can systematically link the science of human factors with cybersecurity. Colleges and universities are a centric node to educate industry, academia, and government leaders on the value of human factors engineering in cybersecurity. Through scholarly research, partnerships with government and industry, and a developing human factors curriculum, academia can influence business decision-makers to leverage human factors engineering with the same rigor and affinity for cybersecurity, software, and network engineering. This panel serves as a platform to increase awareness and the significance of integrating human factors courses into cybersecurity curricula that are taught by faculty members with proper credentials and industry experience.

Title: Fringer: A Finger-Worn Passive Device Enabling Computer Vision Based Force Sensing Using Moiré Fringes Abstract: ABSTRACT Despite the importance of utilizing forces when interacting with objects, sensing force interactions without active force sensors is challenging. We introduce Fringer, a finger sleeve that physically visualizes the force to allow a camera to estimate the force without using any active sensors. The sleeve has stripe-pattern slits, a sliding paper with stripe pattern, and a compliant layer that converts force into sliding paper movements. The patterns of the slit and the paper have different frequencies to create Moiré fringes, which can magnify the small displacement caused by the compliant layer compression for webcams to easily capture such displacement.

Title: SomaFlatables: Supporting Embodied Cognition through Pneumatic Bladders Abstract: ABSTRACTApplying the theory of Embodied Cognition through design allows us to create computational interactions that engage our bodies by modifying our body schema. However, in HCI, most of these interactive experiences have been stationed around creating sensing-based systems that leverage our body’s position and movement to offer an experience, such as games using Nintendo Wii and Xbox Kinect. In this work, we created two pneumatic inflatables-based prototypes that actuate our body to support embodied cognition in two scenarios by altering the user’s body schema. We call these ”SomaFlatables” and demonstrate the design and implementation of these inflatables based prototypes that can move and even extend our bodies, allowing for novel bodily experiences. Furthermore, we discuss the future work and limitations of the current implementation.

Title: Frisson: Leveraging Metasomatic Interactions for Generating Aesthetic Chills Abstract: ABSTRACT Opportunities to evoke emotional experiences and modulate cognitive processes hold great importance in studying relations between emotion, cognition and behaviour as well as building technologies for maintaining positive mental health. We present the concept of Metasomatic Interactions as a new way of generating and controlling previously untapped embodied emotions using illusory sensations tuned to the prior bodily experiences. We present Frisson, a metasomatic interface built to elicit the sensations underlying the embodied emotion of aesthetic chills (i.e., goosebumps, psychogenic shivers). We present a user study (N = 14) in which the device evokes the psychogenic experience of aesthetic chills by simulating traversing thermal sensations across the spine when the illusory sensations overlap with the prior impression of aesthetic chills. The results encourage further testing of metasomatic interfaces for inducing emotions from the body. Using theories of embodied cognition, experimental results and this device, we present a discussion on characteristics of metasomatic interfaces, which we hope will inspire new categories of emotional prostheses in HCI.

Title: Sharing Heartbeat: Toward Conducting Heartrate and Speech Rhythm through Tactile Presentation of Pseudo-heartbeats Abstract: ABSTRACTCurrently, the ongoing COVID-19 pandemic makes physical contact, such as handshakes, difficult. However, physical contact is effective in strengthening the bonds between people. In this study, we aim to compensate for the physical contact lost during the COVID-19 pandemic by presenting a pseudo-heartbeat through a speaker to reproduce entrainment and the synchronized state of heartbeats induced by physiological synchronization. We evaluated the effects of the device in terms of speech rhythm and heart rate. The experimental results showed that a presentation of 80 BPM significantly reduced the difference in heart rate between the two participants, bringing them closer to a synchronized heart rate state. The heart rates of participants were significantly lower when 45 BPM and 80 BPM were presented than when no stimulus was given. Furthermore, when 45 BPM was presented, the silent periods between conversations were significantly more extended than when no stimulus was given. This result indicates that this device can intentionally create the entrainment phenomenon and a synchronized heart rate state, thereby producing the same effect of physical contact communication without contact.

Title: Thumble: One-Handed 3D Object Manipulation Using a Thimble-Shaped Wearable Device in Virtual Reality Abstract: ABSTRACT Conventional controllers or hand-tracking interactions in VR cause hand fatigue while manipulating 3D objects because repetitive wrist rotation and hand movements are often required. As a solution to this inconvenience, we propose Thumble, a novel wearable input device worn on the thumb for modifying the orientation of 3D objects. Thumble can rotate the 3D objects depending on the orientation of the thumb and using the thumb pad as an input surface on which the index finger rubs to control the direction and degree of rotations. Therefore, it requires minimal motion of the wrist and the arm. Through the informal user study, we collected the subjective feedback of users and found that Thumble has less hand movement than a conventional VR controller.

Title: Early Usability Evaluation of a Relational Agent for the COVID-19 Pandemic Abstract: ABSTRACT Relational agents (RAs) have shown effectiveness in various health interventions with and without healthcare professionals (HCPs) and hospital facilities. RAs have not been widely researched in COVID-19 context, although they can give health interventions during the pandemic. Addressing this gap, this work presents an early usability evaluation of a prototypical RA, which is iteratively designed and developed in collaboration with infected patients (n=21) and two groups of HCPs (n=19, n=16) to aid COVID-19 patients at various stages about four main tasks: testing guidance, support during self-isolation, handling emergency situations, and promoting post-infection mental well-being. The prototype obtained an average score of 58.82 on the system usability scale (SUS) after being evaluated by 98 people. This result implies that the suggested design still needs to be improved for greater usability and adoption.