About
Research & Development
A New AI System for Verifying Health Information Across Text and Video
TLDR: A new system called CER (Combining Evidence and Reasoning) has been developed to combat biomedical misinformation. It uses AI to detect claims in...
Research & Development
Slim-SC: Enhancing LLM Reasoning Efficiency Through Intelligent Thought Pruning
TLDR: Slim-SC is a novel method that improves the efficiency of Self-Consistency (SC) in Large Language Models (LLMs) by pruning redundant reasoning chains. It...
Research & Development
Enhancing Compact Object Detectors with Multimodal Semantics
TLDR: MOCHA is a novel knowledge distillation method that transfers rich, object-level multimodal semantics from large vision-language models (like LLaVa) to lightweight, vision-only object...
Research & Development
Instruction-Driven Audio Editing with RFM-Editing
TLDR: RFM-Editing is a novel framework for text-guided audio editing that utilizes rectified flow matching and diffusion models. It enables precise modification of audio...
Research & Development
TimeAlign: Enhancing Time Series Forecasting Through Distribution-Aware Representation Alignment
TLDR: TimeAlign is a new framework for time series forecasting that improves accuracy by explicitly aligning learned representations between historical inputs and future targets....
Research & Development
Unlocking Fine-Grained Video Understanding with Gated Residual Tokenization
TLDR: A new research paper introduces "Dense Video Understanding," a task for AI models to process high-frame-rate videos without losing crucial temporal details. It...