Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers propose a fuzzy-logic extension of Answer Set Programming (ASP) to bridge numerical data and qualitative reasoning, using learning-based membership functions and semantically enriched predicates. The framewor…
Researchers propose EmCom-Diffusion, a framework that measures visual reflection in emergent languages by reconstructing images from emergent messages using a finetuned diffusion model and comparing them to originals. Va…
Researchers propose R²PO, a method that decouples rollout and inference policies for LLM reasoning, achieving accuracy gains of 3.4% on MATH-500 and 1.3% on APPS. The approach attaches a lightweight Residual Rollout-Head…
Researchers propose FlashBlock, a cached block-external attention mechanism that reduces attention computation and KV cache access in block diffusion models for long-context generation. The method achieves up to 1.44× hi…
Researchers introduced Sparse-Reslim, a plug-in routing module that reduces computational cost in ViT-based weather forecasting by processing only 25% of spatial tokens through expensive transformer blocks while maintain…
Researchers compared process and outcome reward structures for reinforcement learning with verifiable rewards (RLVR) in small language models, finding that process-only supervision achieved 63.73% accuracy on GSM8K versu…
Researchers introduced the Lifting Neural Operator (LiNO), a multiresolution neural operator built on the second-generation wavelet lifting scheme, to learn solution operators of differential equations. LiNO adaptively d…
Researchers introduced a post-generation curation method for synthetic images that selects informative subsets from a fixed pool of generated data. The approach splits each real class into homogeneous and heterogeneous s…
Researchers introduced QuantFlow, a federated Mamba-based foundation model for time-series forecasting that combines inverted sequence embedding, bidirectional state-space decoders, quantile regression, and federated lea…
Researchers propose an additive causal construction (ACC) framework for multi-source image fusion that addresses cross-system discrepancy and entanglement, improving out-of-distribution generalization. The method, valida…
Researchers propose Dual-Adaptive SAM3 (DA-SAM3), a framework for parameter-efficient medical image segmentation that uses a Dynamic Expert Router and Decomposed Parameterized Experts to reduce MoE overhead by over 80% w…
Researchers at leadership computing facilities developed REDI, an open-source framework that automates the transformation of large-scale scientific datasets into AI-ready training data through a five-stage pipeline. Eval…
Researchers at arXiv propose a formal model of internal pluralism showing that local pairwise comparisons fail to capture global priorities like proportionality and egalitarianism, and can cause behavioral distortions wh…
Researchers propose Trajectory-Aware Commit Gating (TACG), a training-free decoder for diffusion language models that uses trajectory-aware signals to decide when to commit tokens, improving accuracy and efficiency on co…
Researchers propose ALORE, a scalable and domain-robust estimator for the size of the largest overlap between tables, addressing limitations of the state-of-the-art Armadillo. ALORE reduces mean absolute error by up to 5…
Researchers propose Hierarchical Landmark Sparse (HiLS) Attention, a chunk-wise sparse attention mechanism that learns chunk selection end-to-end under the language-modeling loss. HiLS-Attention achieves performance comp…
Researchers developed a semi-automatic pipeline that converts CT-based cardiac segmentations into simulation-ready meshes in minutes, preserving anatomical and topological consistency. Validated on 58 healthy cardiac CT …
Researchers studied online adaptation of discrete diffusion models for molecular optimization, finding that combining acquisition, reward shaping, debiasing, replay, and validity control outperforms offline fine-tuning a…
Researchers propose a sliding-window-based reinforcement learning (SWRL) framework for dynamic assembly flow shop scheduling with multi-product delivery, addressing challenges from dynamic order arrivals and kitting cons…
Researchers introduced Raven-Agent, the first autonomous trading agent for prediction markets, which achieved the only positive return and positive risk-adjusted return among tested policies in controlled replays over ar…