Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers propose an end-to-end graph neural surrogate for forecasting CO2 plume migration in geological storage, evaluated on the SPE11A benchmark. The model uses an anisotropic message-passing mechanism to capture di…
Researchers propose a unified framework for learning geometry-consistent and domain-robust image representations in monocular endoscopy, combining a synthetic data pipeline with Hierarchy-Aware Geometry-Semantic Adaptati…
Researchers propose a hybrid framework for intraoperative liver registration that combines biomechanical priors with meta-learned residual corrections, achieving improved accuracy and generalization on deformable phantom…
Researchers at TerraTransfer propose a method for end-to-end autonomous driving that eliminates the need for expert demonstrations by decoupling learning to drive from learning to see. The approach uses self-play in vect…
Researchers propose RepSelect, a method for robust LLM unlearning that isolates forget-set-specific representations by collapsing top principal components of weight gradients, achieving 4-50x larger reduction in post-rel…
Researchers introduced GeoDisaster, a benchmark for operational geospatial disaster reasoning with 2,921 instances across 43 question types, and proposed an orchestrated multi-agent framework with 18 disaster-oriented to…
Researchers proposed a BiSpectral Mamba framework combining multi-scale CNN, spectral attention, bidirectional state-space modeling, and quantum-inspired learning for hyperspectral crop field analysis, achieving 84.83% a…
Researchers at arXiv have identified a conflict signature in language models that distinguishes deceptive outputs from honest errors, achieving 100% accuracy in detecting lies across multiple models including GPT-2, Qwen…
Researchers developed constrained diffusion models with primal-dual inference (PDI) to sample from optimal distributions of entropy-regularized optimization problems with average constraints. PDI jointly infers the optim…
Researchers show that grokking in deep neural networks is caused by noise-driven escape from metastable phases during first-order L2 regularization phase transitions. Using linear DNNs, they demonstrate that SGD noise he…
Researchers tested whether language models can independently discover the concept of zero through arithmetic generalization. GPT-2-sized models failed at test time regardless of language pretraining, but improved after t…
Researchers introduced SpeechDx, a large-scale benchmark for clinical speech AI spanning 12 datasets and 27 tasks across diverse health conditions, designed to evaluate generalization by structuring tasks according to di…
Researchers propose SierpinskiCam, a method for video retaking that uses Sierpinski triangle pattern cues to improve camera controllability and geometric consistency under large viewpoint changes. The approach augments g…
Researchers developed an online adaptive clinical decision support AI system that integrates treatment effect estimation, patient digital twins, and reinforcement learning to recommend treatments in real-time while ensur…
Researchers introduced a geosteering framework that combines particle filtering with reinforcement learning to optimize well trajectory decisions under geological uncertainty. The framework evaluates three decision-makin…
Researchers introduced the Simple Strands Agent (SSA) to minimize the 'intent-execution gap' between AI models and their harnesses, improving agent performance. Analyzing 138k trajectories, they found that fine-grained m…
Researchers propose negative token filtering, a method enabling stable single-rollout training for critic-free reinforcement learning in large language models, outperforming group-based techniques on agentic tasks.
Researchers propose a three-layer architecture for verbal reinforcement learning in LLM agents, addressing the retention-forgetting dilemma in non-stationary environments. The system uses rules, evidence, and skills with…
Researchers tested Word2Vec on Toki Pona, a constructed language with only ~130 words, using 1.4 million sentences. They found that non-core tokens like loanwords improved embedding quality by drawing similar words close…
Researchers developed the first principled method for unsupervised learning of statistics for a maximum entropy model of visual textures, using diffusion-based training and sampling. Despite using only 512 statistics, th…