About
Research & Development
Adaptive AI: How Models Learn to Choose the Right Way to Reason Visually
TLDR: Researchers introduce Mixture-of-Visual-Thoughts (MoVT), an adaptive AI paradigm that unifies different visual reasoning modes within a single model. Their AdaVaR framework, trained in...
Research & Development
HILBERT: Combining AI Intuition and Rigor for Advanced Math Proofs
TLDR: HILBERT is an agentic AI framework that significantly improves automated theorem proving by integrating the informal mathematical reasoning of large language models (LLMs)...
Research & Development
Bridging Language and Sensors: LLMs for Proactive Industrial Health Monitoring
TLDR: A new study explores the use of Large Language Models (LLMs) for fault diagnosis in industrial settings, specifically within a simulated HVAC system....
Research & Development
MathBode: A Dynamic Lens on LLM Mathematical Reasoning
TLDR: MathBode introduces a novel frequency-domain diagnostic for evaluating LLM mathematical reasoning. Instead of just final answers, it measures how models track sinusoidal changes...
Research & Development
Unpacking Transformer Reasoning: New Insights into Multi-step Logic Limits
TLDR: This paper investigates how many logical steps a Transformer model can perform in a single pass. By proposing new information propagation rules and...
Research & Development
Unlocking Data’s Hidden Paths: A New Approach to Understanding Complex Trajectories
TLDR: A new research paper introduces Energy Guided Geometric Flow Matching (EGGFM), a method that uses score matching and annealed energy distillation to learn...