arXiv subjects

Misha Sra

Publications and source records attributed to Misha Sra.

At least 55 records · Page 3Linked to original sources

arXiv2024

GraV: Grasp Volume Data for the Design of One-Handed XR Interfaces

Everyday objects, like remote controls or electric toothbrushes, are crafted with hand-accessible interfaces. Expanding on this design principle, extended reality (XR) interfaces for physical tasks could facilitate interaction without necessitating the release of grasped tools, ensuring seamless workflow integration. While established data, such as hand anthropometric measurements, guide the design of handheld objects, XR currently lacks comparable data, regarding reachability, for single-hand interfaces while grasping objects. To address this, we identify critical design factors and a design space representing grasp-proximate interfaces and introduce a simulation tool for generating reachability and displacement cost data for designing these interfaces. Additionally, using the simulation tool, we generate a dataset based on grasp taxonomy and common household objects. Finally, we share insights from a design workshop that emphasizes the significance of reachability and motion cost data, empowering XR creators to develop bespoke interfaces tailored specifically to grasping hands.

cs.HC↗

arXiv2024

XplainLLM: A Knowledge-Augmented Dataset for Reliable Grounded Explanations in LLMs

Large Language Models (LLMs) have achieved remarkable success in natural language tasks, yet understanding their reasoning processes remains a significant challenge. We address this by introducing XplainLLM, a dataset accompanying an explanation framework designed to enhance LLM transparency and reliability. Our dataset comprises 24,204 instances where each instance interprets the LLM's reasoning behavior using knowledge graphs (KGs) and graph attention networks (GAT), and includes explanations of LLMs such as the decoder-only Llama-3 and the encoder-only RoBERTa. XplainLLM also features a framework for generating grounded explanations and the debugger-scores for multidimensional quality analysis. Our explanations include why-choose and why-not-choose components, reason-elements, and debugger-scores that collectively illuminate the LLM's reasoning behavior. Our evaluations demonstrate XplainLLM's potential to reduce hallucinations and improve grounded explanation generation in LLMs. XplainLLM is a resource for researchers and practitioners to build trust and verify the reliability of LLM outputs.

cs.CL↗

arXiv2024

AID-AppEAL: Automatic Image Dataset and Algorithm for Content Appeal Enhancement and Assessment Labeling

We propose Image Content Appeal Assessment (ICAA), a novel metric that quantifies the level of positive interest an image's content generates for viewers, such as the appeal of food in a photograph. This is fundamentally different from traditional Image-Aesthetics Assessment (IAA), which judges an image's artistic quality. While previous studies often confuse the concepts of ``aesthetics'' and ``appeal,'' our work addresses this by being the first to study ICAA explicitly. To do this, we propose a novel system that automates dataset creation and implements algorithms to estimate and boost content appeal. We use our pipeline to generate two large-scale datasets (70K+ images each) in diverse domains (food and room interior design) to train our models, which revealed little correlation between content appeal and aesthetics. Our user study, with more than 76% of participants preferring the appeal-enhanced images, confirms that our appeal ratings accurately reflect user preferences, establishing ICAA as a unique evaluative criterion. Our code and datasets are available at https://github.com/SherryXTChen/AID-Appeal.

cs.CV↗

arXiv2024

LMExplainer: Grounding Knowledge and Explaining Language Models

Language models (LMs) like GPT-4 are important in AI applications, but their opaque decision-making process reduces user trust, especially in safety-critical areas. We introduce LMExplainer, a novel knowledge-grounded explainer that clarifies the reasoning process of LMs through intuitive, human-understandable explanations. By leveraging a graph attention network (GAT) with a large-scale knowledge graph (KG), LMExplainer not only precisely narrows the reasoning space to focus on the most relevant knowledge but also grounds its reasoning in structured, verifiable knowledge to reduce hallucinations and enhance interpretability. LMExplainer effectively generates human-understandable explanations to enhance transparency and streamline the decision-making process. Additionally, by incorporating debugging into the explanation, it offers expertise suggestions that improve LMs from a developmental perspective. Thus, LMExplainer stands as an enhancement in making LMs more accessible and understandable to users. We evaluate LMExplainer on benchmark datasets such as CommonsenseQA and OpenBookQA, demonstrating that it outperforms most existing methods. By comparing the explanations generated by LMExplainer with those of other models, we show that our approach offers more comprehensive and clearer explanations of the reasoning process. LMExplainer provides a deeper understanding of the inner workings of LMs, advancing towards more reliable, transparent, and equitable AI.

cs.CL↗

arXiv2024

ConnectVR: A Trigger-Action Interface for Creating Agent-based Interactive VR Stories

The demand for interactive narratives is growing with increasing popularity of VR and video gaming. This presents an opportunity to create interactive storytelling experiences that allow players to engage with a narrative from a first person perspective, both, immersively in VR and in 3D on a computer. However, for artists and storytellers without programming experience, authoring such experiences is a particularly complex task as it involves coding a series of story events (character animation, movements, time control, dialogues, etc.) to be connected and triggered by a variety of player behaviors. In this work, we present ConnectVR, a trigger-action interface to enable non-technical creators design agent-based narrative experiences. Our no-code authoring method specifically focuses on the design of narratives driven by a series of cause-effect relationships triggered by the player's actions. We asked 15 participants to use ConnectVR in a preliminary workshop study as well as two artists to extensively use our system to create VR narrative projects in a three-week in-depth study. Our findings shed light on the creative opportunities facilitated by ConnectVR's trigger-action approach, particularly its capability to establish chained behavioral effects between virtual characters and objects. The results of both studies underscore the positive feedback from participants regarding our system's capacity to not only support creativity but also to simplify the creation of interactive narrative experiences. Results indicate compatibility with non-technical narrative creator's workflows, showcasing its potential to enhance the overall creative process in the realm of VR narrative design.

cs.HC↗

arXiv2024

EntangleVR++: Evaluating the Potential of using Entanglement in an Interactive VR Scene Creation System

Interactive digital stories provide a sense of flexibility and freedom to players by allowing them to make choices at key junctions. These choices advance the narrative and determine, to some degree, how the story evolves for that player. As shown in prior work, the ability to control or participate in the construction of the narrative can give the player a high level of agency that results in a stronger sense of immersion in the narrative experience. To support the design of this type of interactive storytelling, our system, EntangleVR++, borrows the idea of entanglement from quantum computing. Our use of entanglement allows creators and storytellers control over which sequences of story events take place in correlation with each other, initiated by the choices a player makes. In this work, we evaluated how well our idea of entanglement enables creators to easily and quickly design interactive VR narratives. We asked 16 participants to use our system and based on user interviews, analyses of screen recordings, and questionnaire feedback, we extracted four themes. From these themes and the study overall, we derived four authoring strategies for tool designers interested in the design of future visual interface for interactively creating virtual scenes that include relational objects and multiple outcomes driven by player interactions.

cs.HC↗

arXiv2024

DanceGen: Supporting Choreography Ideation and Prototyping with Generative AI

Choreography creation requires high proficiency in artistic and technical skills. Choreographers typically go through four stages to create a dance piece: preparation, studio, performance, and reflection. This process is often individualized, complicated, and challenging due to multiple constraints at each stage. To assist choreographers, most prior work has focused on designing digital tools to support the last three stages of the choreography process, with the preparation stage being the least explored. To address this research gap, we introduce an AI-based approach to assist the preparation stage by supporting ideation, creating choreographic prototypes, and documenting creative attempts and outcomes. We address the limitations of existing AI-based motion generation methods for ideation by allowing generated sequences to be edited and modified in an interactive web interface. This capability is motivated by insights from a formative study we conducted with seven choreographers. We evaluated our system's functionality, benefits, and limitations with six expert choreographers. Results highlight the usability of our system, with users reporting increased efficiency, expanded creative possibilities, and an enhanced iterative process. We also identified areas for improvement, such as the relationship between user intent and AI outcome, intuitive and flexible user interaction design, and integration with existing physical choreography prototyping workflows. By reflecting on the evaluation results, we present three insights that aim to inform the development of future AI systems that can empower choreographers.

cs.HC↗

arXiv2024

TiNO-Edit: Timestep and Noise Optimization for Robust Diffusion-Based Image Editing

Despite many attempts to leverage pre-trained text-to-image models (T2I) like Stable Diffusion (SD) for controllable image editing, producing good predictable results remains a challenge. Previous approaches have focused on either fine-tuning pre-trained T2I models on specific datasets to generate certain kinds of images (e.g., with a specific object or person), or on optimizing the weights, text prompts, and/or learning features for each input image in an attempt to coax the image generator to produce the desired result. However, these approaches all have shortcomings and fail to produce good results in a predictable and controllable manner. To address this problem, we present TiNO-Edit, an SD-based method that focuses on optimizing the noise patterns and diffusion timesteps during editing, something previously unexplored in the literature. With this simple change, we are able to generate results that both better align with the original images and reflect the desired result. Furthermore, we propose a set of new loss functions that operate in the latent domain of SD, greatly speeding up the optimization when compared to prior approaches, which operate in the pixel domain. Our method can be easily applied to variations of SD including Textual Inversion and DreamBooth that encode new concepts and incorporate them into the edited results. We present a host of image-editing capabilities enabled by our approach. Our code is publicly available at https://github.com/SherryXTChen/TiNO-Edit.

cs.CV↗

arXiv2024

Exploring AI-assisted Ideation and Prototyping for Choreography

Choreography creation is a multimodal endeavor, demanding cognitive abilities to develop creative ideas and technical expertise to convert choreographic ideas into physical dance movements. Previous endeavors have sought to reduce the complexities in the choreography creation process in both dimensions. Among them, non-AI-based systems have focused on reinforcing cognitive activities by helping analyze and understand dance movements and augmenting physical capabilities by enhancing body expressivity. On the other hand, AI-based methods have helped the creation of novel choreographic materials with generative AI algorithms. The choreography creation process is constrained by time and requires a rich set of resources to stimulate novel ideas, but the need for iterative prototyping and reduced physical dependence have not been adequately addressed by prior research. Recognizing these challenges and the research gap, we present an innovative AI-based choreography-support system. Our goal is to facilitate rapid ideation by utilizing a generative AI model that can produce diverse and novel dance sequences. The system is designed to support iterative digital dance prototyping through an interactive web-based user interface that enables the editing and modification of generated motion. We evaluated our system by inviting six choreographers to analyze its limitations and benefits and present the evaluation results along with potential directions for future work.

cs.HC↗

arXiv2023

SPOTR: Spatio-temporal Pose Transformers for Human Motion Prediction

3D human motion prediction is a research area of high significance and a challenge in computer vision. It is useful for the design of many applications including robotics and autonomous driving. Traditionally, autogregressive models have been used to predict human motion. However, these models have high computation needs and error accumulation that make it difficult to use them for realtime applications. In this paper, we present a non-autogressive model for human motion prediction. We focus on learning spatio-temporal representations non-autoregressively for generation of plausible future motions. We propose a novel architecture that leverages the recently proposed Transformers. Human motion involves complex spatio-temporal dynamics with joints affecting the position and rotation of each other even though they are not connected directly. The proposed model extracts these dynamics using both convolutions and the self-attention mechanism. Using specialized spatial and temporal self-attention to augment the features extracted through convolution allows our model to generate spatio-temporally coherent predictions in parallel independent of the activity. Our contributions are threefold: (i) we frame human motion prediction as a sequence-to-sequence problem and propose a non-autoregressive Transformer to forecast a sequence of poses in parallel; (ii) our method is activity agnostic; (iii) we show that despite its simplicity, our approach is able to make accurate predictions, achieving better or comparable results compared to the state-of-the-art on two public datasets, with far fewer parameters and much faster inference.

cs.CV↗

arXiv2022

CardsVR: A Two-Person VR Experience with Passive Haptic Feedback from a Deck of Playing Cards

Presence in virtual reality (VR) is meaningful for remotely connecting with others and facilitating social interactions despite great distance while providing a sense of "being there." This work presents CardsVR, a two-person VR experience that allows remote participants to play a game of cards together. An entire deck of tracked cards are used to recreate the sense of playing cards in-person. Prior work in VR commonly provides passive haptic feedback either through a single object or through static objects in the environment. CardsVR is novel in providing passive haptic feedback through multiple cards that are individually tracked and represented in the virtual environment. Participants interact with the physical cards by picking them up, holding them, playing them, or moving them on the physical table. Our participant study (N=23) shows that passive haptic feedback provides significant improvement in three standard measures of presence: Possibility to Act, Realism, and Haptics.

cs.HC↗

arXiv2022

Adaptive Virtual Neuroarchitecture

Our surrounding environment impacts our cognitive-emotional processes on a daily basis and shapes our physical, psychological and social wellbeing. Although the effects of the built environment on our psycho-physiological processes are well studied, virtual environment design with a potentially similar impact on the user, has received limited attention. Based on the influence of space design on a user and combining that with the dynamic affordances of virtual spaces, we present the idea of adaptive virtual neuroarchitecture (AVN), where virtual environments respond to the user and the user's real world context while simultaneously influencing them both in realtime. To show how AVN has been explored in current research, we present a sampling of recent work that demonstrates reciprocal relationships using physical affordances (space, objects), the user's state (physiological, cognitive, emotional), and the virtual world used in the design of novel virtual reality experiences. We believe AVN has the potential to help us learn how to design spaces and environments that can enhance the wellbeing of their inhabitants.

cs.HC↗

arXiv2022

Txt2Vid: Ultra-Low Bitrate Compression of Talking-Head Videos via Text

Video represents the majority of internet traffic today, driving a continual race between the generation of higher quality content, transmission of larger file sizes, and the development of network infrastructure. In addition, the recent COVID-19 pandemic fueled a surge in the use of video conferencing tools. Since videos take up considerable bandwidth (~100 Kbps to a few Mbps), improved video compression can have a substantial impact on network performance for live and pre-recorded content, providing broader access to multimedia content worldwide. We present a novel video compression pipeline, called Txt2Vid, which dramatically reduces data transmission rates by compressing webcam videos ("talking-head videos") to a text transcript. The text is transmitted and decoded into a realistic reconstruction of the original video using recent advances in deep learning based voice cloning and lip syncing models. Our generative pipeline achieves two to three orders of magnitude reduction in the bitrate as compared to the standard audio-video codecs (encoders-decoders), while maintaining equivalent Quality-of-Experience based on a subjective evaluation by users (n = 242) in an online study. The Txt2Vid framework opens up the potential for creating novel applications such as enabling audio-video communication during poor internet connectivity, or in remote terrains with limited bandwidth. The code for this work is available at https://github.com/tpulkit/txt2vid.git.

eess.IV↗

arXiv2021

Self-Supervised Knowledge Assimilation for Expert-Layman Text Style Transfer

Expert-layman text style transfer technologies have the potential to improve communication between members of scientific communities and the general public. High-quality information produced by experts is often filled with difficult jargon laypeople struggle to understand. This is a particularly notable issue in the medical domain, where layman are often confused by medical text online. At present, two bottlenecks interfere with the goal of building high-quality medical expert-layman style transfer systems: a dearth of pretrained medical-domain language models spanning both expert and layman terminologies and a lack of parallel corpora for training the transfer task itself. To mitigate the first issue, we propose a novel language model (LM) pretraining task, Knowledge Base Assimilation, to synthesize pretraining data from the edges of a graph of expert- and layman-style medical terminology terms into an LM during self-supervised learning. To mitigate the second issue, we build a large-scale parallel corpus in the medical expert-layman domain using a margin-based criterion. Our experiments show that transformer-based models pretrained on knowledge base assimilation and other well-established pretraining tasks fine-tuning on our new parallel corpus leads to considerable improvement against expert-layman transfer benchmarks, gaining an average relative improvement of our human evaluation, the Overall Success Rate (OSR), by 106%. We release our code and parallel corpus for future research.

cs.CL↗

arXiv2021

Exploratory Design of a Hands-free Video Game Controller for a Quadriplegic Individual

From colored pixels to hyper-realistic 3D landscapes of virtual reality, video games have evolved immensely over the last few decades. However, video game input still requires two-handed dexterous finger manipulations for simultaneous joystick and trigger or mouse and keyboard presses. In this work, we explore the design of a hands-free game control method using realtime facial expression recognition for individuals with neurological and neuromuscular diseases who are unable to use traditional game controllers. Similar to other Assistive Technologies (AT), our facial input technique is also designed and tested in collaboration with a graduate student who has Spinal Muscular Atrophy. Our preliminary evaluation shows the potential of facial expression recognition for augmenting the lives of quadriplegic individuals by enabling them to accomplish things like walking, running, flying or other adventures that may not be so attainable otherwise.

cs.HC↗

arXiv2021

SceneAR: Scene-based Micro Narratives for Sharing and Remixing in Augmented Reality

Short-form digital storytelling has become a popular medium for millions of people to express themselves. Traditionally, this medium uses primarily 2D media such as text (e.g., memes), images (e.g., Instagram), gifs (e.g., Giphy), and videos (e.g., TikTok, Snapchat). To expand the modalities from 2D to 3D media, we present SceneAR, a smartphone application for creating sequential scene-based micro narratives in augmented reality (AR). What sets SceneAR apart from prior work is the ability to share the scene-based stories as AR content -- no longer limited to sharing images or videos, these narratives can now be experienced in people's own physical environments. Additionally, SceneAR affords users the ability to remix AR, empowering them to build-upon others' creations collectively. We asked 18 people to use SceneAR in a 3-day study. Based on user interviews, analysis of screen recordings, and the stories they created, we extracted three themes. From those themes and the study overall, we derived six strategies for designers interested in supporting short-form AR narratives.

cs.HC↗

arXiv2021

Telelife: The Future of Remote Living

In recent years, everyday activities such as work and socialization have steadily shifted to more remote and virtual settings. With the COVID-19 pandemic, the switch from physical to virtual has been accelerated, which has substantially affected various aspects of our lives, including business, education, commerce, healthcare, and personal life. This rapid and large-scale switch from in-person to remote interactions has revealed that our current technologies lack functionality and are limited in their ability to recreate interpersonal interactions. To help address these limitations in the future, we introduce "Telelife," a vision for the near future that depicts the potential means to improve remote living better aligned with how we interact, live and work in the physical world. Telelife encompasses novel synergies of technologies and concepts such as digital twins, virtual prototyping, and attention and context-aware user interfaces with innovative hardware that can support ultrarealistic graphics, user state detection, and more. These ideas will guide the transformation of our daily lives and routines soon, targeting the year 2035. In addition, we identify opportunities across high-impact applications in domains related to this vision of Telelife. Along with a recent survey of relevant fields such as human-computer interaction, pervasive computing, and virtual reality, the directions outlined in this paper will guide future research on remote living.

cs.HC↗

arXiv2015

Design Strategies for Playful Technologies to Support Light-intensity Physical Activity in the Workplace

Moderate to vigorous intensity physical activity has an established preventative role in obesity, cardiovascular disease, and diabetes. However recent evidence suggests that sitting time affects health negatively independent of whether adults meet prescribed physical activity guidelines. Since many of us spend long hours daily sitting in front of a host of electronic screens, this is cause for concern. In this paper, we describe a set of three prototype digital games created for encouraging light-intensity physical activity during short breaks at work. The design of these kinds of games is a complex process that must consider motivation strategies, interaction methodology, usability and ludic aspects. We present design guidelines for technologies that encourage physical activity in the workplace that we derived from a user evaluation using the prototypes. Although the design guidelines can be seen as general principles, we conclude that they have to be considered differently for different workplace cultures and workspaces. Our study was conducted with users who have some experience playing casual games on their mobile devices and were able and willing to increase their physical activity.

cs.HC↗