Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers introduced Hierarchical Global Attention (HGA), a method that reduces GPU memory usage in long-context transformers by using hierarchical routing, enabling a 64K-token context on an RTX 5090 GPU with 32GB mem…
A developer describes how combining shift-left testing with machine learning can predict where bugs will appear before code is merged. By analyzing historical defect data and commit-level features such as code churn, own…
Researchers propose the Quality Representation Module (QRM), a lightweight transformer module that learns a quality-aware representation to modulate diffusion transformers (DiT) for text-to-image generation, improving im…
Researchers developed a closed-form model of Group Relative Policy Optimization (GRPO) training dynamics, predicting reward trajectories and stability thresholds from first principles. The model subsumes empirical satura…
Researchers propose an offline reinforcement learning framework for active flow control that extracts policies from existing data, eliminating the need for real-time interactions. The sensor position-conditioned architec…
Researchers introduced HyPOLE, a framework for multi-agent reinforcement learning under partial observation that uses hyperproperty specifications in HyperLTL to guide policy synthesis. The framework integrates centraliz…
Researchers developed a causal machine learning approach to estimate the impact of additional supply on transaction outcomes in two-sided marketplaces, using Airbnb as a case study. The method combines double/debiased ma…
Researchers introduced HASTE, a hierarchical multi-agent system that organizes cross-competition knowledge into three scope tiers to improve transfer efficiency in ML engineering. In tests on the MLE-Bench Lite benchmark…
Researchers introduced a Bangla event detection benchmark with 9,979 annotated sentences across clean text, ASR transcripts, and corrupted text. They found encoder models perform better on clean text but degrade under no…
Researchers introduce Depth-wise Gradient Augmentation, a new optimization paradigm that transforms layer-wise updates to exploit structured relationships across deep neural network layers. Their instantiation, Gradient …
Researchers developed a physics-informed conditional normalizing flow for angles-only cislunar orbit determination, enabling probabilistic inference of initial states from short observation arcs. The method, trained on p…
A new study from arXiv investigates how AI can retrieve simulation models using natural language queries, finding that data representation, open-source embedding models, and reranking strategies significantly impact retr…
Researchers introduced LearnStop, a cost-aware early-exit method for reasoning language models that learns when to stop computation based on online features like answer confidence and entropy. Across 18 task-model settin…
Researchers introduced MCO-PDE, a competitive optimization framework that discovers shared partial differential equations from multi-source datasets by fusing as few as 50 observations per dataset. The framework uses neu…
Researchers introduced ReactionAtlas, a machine learning framework that autonomously constructs chemical reaction networks from seed molecules without hand-crafted rules. Starting from eight pre-biotic seeds, it discover…
Researchers introduced a three-phase deep reinforcement learning system for personalized portfolio management that overcomes ticker lock-in, monolithic objectives, and static user models. Phase 1 uses a self-supervised c…
Researchers argue that AI agents should help users construct preferences rather than assume they have well-formed ones, introducing CoPref and CoShop to model and benchmark this. Testing five frontier models on CoShop, n…
Researchers introduced AgRefactor, an LLM-based multi-agent workflow that refactors software into HLS-compatible code with a self-evolving memory system. On 9 of 11 challenging benchmarks, it outperformed state-of-the-ar…
Researchers introduced BayesBench, a suite of simulation environments to evaluate how large language models update beliefs under multi-turn evidence accumulation. Testing seven LLMs from 3B to 70B parameters, they found …
Researchers submitted a system to SemEval-2026 Task 9 for detecting multilingual online polarization using transformer-based models with class weighting and threshold tuning. Their approach achieved competitive F1 macro …