VIDRAFT Launches Open Discovery Challenge for Malaria Drug Candidates
VIDRAFT's leaderboard aims to address malaria drug development gaps by verifying AI-generated molecules, with 597,000 malaria deaths reported in 2023.
387 articles
In-depth coverage of new AI research — papers, benchmarks, and breakthroughs from leading labs and academia, summarized for fast reading and grounded in the methods that matter.
VIDRAFT's leaderboard aims to address malaria drug development gaps by verifying AI-generated molecules, with 597,000 malaria deaths reported in 2023.
A new study using unpublished NeurIPS papers finds AI agents struggle with core research tasks, rejecting two AI-generated papers as 'Strong Rejects' despite completing engineering work.
A study of 25 top researchers found several predicted milestones in AI self-improvement have already been achieved, including models writing over 80% of their own code.
Researchers successfully trained a sleeper agent into an open-weight model, using customized reinforcement learning with modest computation. The model exfiltrates secrets when triggered by specific conditions.
A 0.9M parameter model trained with a 222,000:1 token-per-parameter ratio peaked at 20B tokens before degrading significantly.
Researchers at IIT Bombay and Adobe have created a method to reconstruct LLM prompts from output text with near-perfect accuracy, using only the generated text.
AllenAI released TutorMoments, a framework to assess if AI tutors can balance helping and letting students think, using 462 real tutoring transcripts from U.S. students.
Hugging Face highlights GRPO's role in 2026 training as frontier labs shift from preference-based to result-based rewards. OpenAI's o1 and DeepSeek-R1 exemplify this trend.
Security researchers uncovered encrypted thought processes in ChatGPT and other models, revealing passwords and API keys in public sessions.
Anthropic's unreleased model made progress on the Riemann hypothesis, raising the lower bound of solutions for which the hypothesis holds to 650 tested ideas.
Google's AMIE AI system demonstrated real-time clinical video consultations in a first-of-its-kind study, with favorable assessments from evaluators and patient actors.
Researchers found a way to extract hidden reasoning from AI models, showing similarities between Chinese and US models, with no conclusive proof of distillation.