About
Research & Development
Enhancing Web Agent Performance with Efficient DOM Processing
TLDR: D2Snap is a new algorithm that significantly reduces the size of Document Object Model (DOM) snapshots for LLM-based web agents, making them a...
Research & Development
Unlocking Instructional AI: A Tool for Generating Teaching Dialogues
TLDR: SIMINSTRUCT is a novel tool that uses LLM-simulated novice instructors to enable human experts to generate high-quality, multi-turn scaffolding dialogues. This method addresses...
Research & Development
Teaching AI When to Stop Thinking: A Meta-Cognitive Approach for Large Language Models
TLDR: The Meta-cognitive Reasoning Framework (MERA) addresses the 'overthinking' problem in Large Reasoning Models (LRMs) by explicitly separating their reasoning and control processes. MERA...
Research & Development
Understanding AI Assistants: A Deep Dive into OS Agents for Digital Device Control
TLDR: This paper surveys OS Agents, AI assistants powered by multimodal large language models (MLLMs) that operate computing devices like phones and computers. It...
Research & Development
Assessing AI’s Role in Guiding Interdisciplinary STEM Learning
TLDR: SID is a new benchmark for evaluating large language models (LLMs) in guiding interdisciplinary STEM education through Socratic dialogues. It features a large...
News & Current Events
AI Enables Robots to Self-Program, Paving Way for Advanced Autonomy
TLDR: A recent demonstration by computer scientist Peter Burke has shown that generative AI models can enable one robot to program the artificial intelligence...