About
Research & Development
Enhancing AI Agents for Interactive Tool Use with Turn-Level Feedback and Diverse Training
TLDR: This research introduces Turn-level Adjudicated Reinforcement Learning (TARL) and mixed-task training to improve interactive multimodal tool-use agents. TARL uses an LLM judge to...
Research & Development
Understanding Malware Behavior with Large Language Models: Introducing BEACON
TLDR: BEACON is a novel deep learning framework for malware classification that uses large language models (LLMs) to generate dense, contextual embeddings from raw...
Research & Development
OnlineMate: AI Companions That Understand and Adapt to Student Minds in Online Learning
TLDR: OnlineMate is a multi-agent learning companion system that uses Large Language Models (LLMs) and Theory of Mind (ToM) to provide personalized cognitive support...
Research & Development
MeanFlowSE: Accelerating Generative Speech Enhancement with One-Step Processing
TLDR: MeanFlowSE is a novel generative speech enhancement model that addresses the computational bottleneck of existing multistep methods. By learning an average velocity field,...
Research & Development
Decomposing Image Classification with Multi-Agent Reasoning
TLDR: MARIC is a new multi-agent framework for image classification that uses specialized agents to collaboratively analyze images. An Outliner Agent identifies the global...
Research & Development
G2LFormer: A Global-to-Local Attention Scheme for Graph Transformers
TLDR: The paper introduces G2LFormer, a novel Graph Transformer model that uses a global-to-local attention scheme. Unlike traditional methods that prioritize local or parallel...