Entropy Centroids for Efficient Test-Time Scaling in LLMs
Discover how entropy centroids improve test-time scaling by using intrinsic rewards to boost large language model response quality without external models.
Co-Learning Port-Hamiltonian Systems for Optimal Energy Control
Discover a novel framework for co-learning port-Hamiltonian systems and optimal energy-shaping control using neural networks and trajectory data.
Test-Time Safety Alignment for Safer AI Outputs
Discover how test-time safety alignment optimizes AI input embeddings to ensure safer, more reliable AI-generated content without sacrificing creativity.
Neural Cellular Automata for Structural Generalization on SLOG
Discover how neural cellular automata achieve structural generalization on SLOG without hand-written rules, outperforming traditional semantic parsing mode...
Data-Centric AI for Fluorescence Imaging in Glioma Surgery
Discover a data-centric AI framework enhancing intraoperative fluorescence imaging accuracy for glioma surgical guidance and improved patient outcomes.