About
Research & Development
Unveiling AI’s Blind Spot: Why Language Models Struggle with Subjective Writing Preferences
TLDR: A new benchmark, WritingPreferenceBench, reveals that current AI models struggle to evaluate subjective writing quality (creativity, style) when objective errors are removed. Standard...
Research & Development
Future Summaries: A New Approach to Training Smarter LLMs for Long-Range Reasoning
TLDR: A new pretraining method called Future Summary Prediction (FSP) improves large language models (LLMs) by training them to predict a compact summary of...
Research & Development
Virtual Mars: AI Reconstructs Missing Data for Realistic Environments
TLDR: A new research paper introduces an unconditional diffusion model to reconstruct missing data in Martian heightmaps, creating accurate and visually coherent 3D terrains...
Research & Development
Unifying User Interests Across Diverse Online Scenarios with LLM-Enhanced Recommendations
TLDR: RED-Rec is an LLM-enhanced recommendation system that unifies user interest modeling across diverse online scenarios like feeds, search, and ads. It uses a...
Research & Development
R-Super: AI Learns to Spot Tumors from Doctor’s Notes, Not Just Drawings
TLDR: R-Super is a novel AI method that trains tumor segmentation models using readily available medical reports, significantly reducing the need for costly manual...
Research & Development
Layered Control Architecture Unlocks Robust Humanoid Locomotion
TLDR: A new research paper demonstrates that a layered control architecture (LCA) with a fast proprioceptive stabilizer and a slower perceptual policy, trained in...