Deep Learning

S*: Hybrid test-time scaling for code generation, boosting both coverage and selection accuracy.

NExT-Mol: Combines 1D language models with 3D diffusion for molecule generation, achieving state-of-the-art performance and validity.

ThinkPO improves LLM reasoning by preferring longer CoT, boosting performance without new data.

Small language models struggle to learn complex reasoning from large models, but a novel ‘Mix Distillation’ method balances complexity for effective capability transfer.

AdaPTS effectively adapts pre-trained univariate time series models to probabilistic multivariate forecasting, improving accuracy and uncertainty quantification.

ProbeLog: Zero-shot model search directly from weights, boosting efficiency and accuracy!

W2SD: A novel framework boosts diffusion model quality by using the difference between weak and strong models to refine sampling trajectories, achieving state-of-the-art performance.

Physics-Informed Gaussians (PIGs) revolutionize PDE solving by using adaptive, learnable Gaussian functions for superior accuracy and efficiency.

Hybrid Graph Sequence Model (GSM++) outperforms existing models by using hierarchical sequences and a hybrid architecture of Transformers and recurrent models, effectively capturing both local and glo…

Enriched dataset TESORO improves technical debt detection by combining self-admitted comments and Java source code, advancing state-of-the-art models.

SambaMixer: A novel state-space model accurately predicts Li-ion battery health using efficient Mamba architecture and innovative resampling techniques.