User-in-the-loop Evaluation of Multimodal LLMs for Activity Assistance
Our research investigates the capability of modern multimodal reasoning models, powered by Large Language Models (LLMs), to facilitate vision-powered assistants for multi-step daily activities. Such assistants must be able to 1) encode relevant visual history from the assistant's sensors, e.g., camera, 2) forecast future actions for accomplishing the activity, and 3) replan based on the user in the loop. To evaluate the first two capabilities, grounding visual history and forecasting in short and long horizons, we conduct benchmarking of two prominent classes of multimodal LLM approaches -- Socratic Models and Vision Conditioned Language Models (VCLMs) on video-based action anticipation tasks using offline datasets. These offline benchmarks, however, do not allow us to close the loop with the user, which is essential to evaluate the replanning capabilities and measure successful activity completion in assistive scenarios. To that end, we conduct a first-of-its-kind user study, with 18 participants performing 3 different multi-step cooking activities while wearing an egocentric observation device called Aria and following assistance from multimodal LLMs. We find that the Socratic approach outperforms VCLMs in both offline and online settings. We further highlight how grounding long visual history, common in activity assistance, remains challenging in current models, especially for VCLMs, and demonstrate that offline metrics do not indicate online performance.
Code (0)
등록된 구현이 없습니다.
Tasks
Action AnticipationBenchmarkingMultimodal ReasoningMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
ChatSpot: Bootstrapping Multimodal LLMs via Precise Referring Instruction Tuning
Human-AI interactivity is a critical aspect that reflects the usability of multimodal large language models (MLLMs). However, existing end-to-end MLLMs only allow users to interact with them through language instructions…
Instruction FollowingLanguage ModelingLanguage ModellingLarge Language Model+1Affective Multimodal Agents with Proactive Knowledge Grounding for Emotionally Aligned Marketing Dialogue
Recent advances in large language models (LLMs) have enabled fluent dialogue systems, but most remain reactive and struggle in emotionally rich, goal-oriented settings such as marketing conversations. To address this lim…
User Feedback Alignment for LLM-powered Exploration in Large-scale Recommendation Systems
Exploration, the act of broadening user experiences beyond their established preferences, is challenging in large-scale recommendation systems due to feedback loops and limited signals on user exploration patterns. Large…
DiversityRecommendation SystemsWorld KnowledgeGPT-4o: Visual perception performance of multimodal large language models in piglet activity understanding
Animal ethology is an crucial aspect of animal research, and animal behavior labeling is the foundation for studying animal behavior. This process typically involves labeling video clips with behavioral semantic tags, a …
Activity RecognitionMMR totalSemantic correspondenceVideo Understanding+1What's Missing in Screen-to-Action? Towards a UI-in-the-Loop Paradigm for Multimodal GUI Reasoning
Existing Graphical User Interface (GUI) reasoning tasks remain challenging, particularly in UI understanding. Current methods typically rely on direct screen-based decision-making, which lacks interpretability and overlo…