Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
A new hybrid framework called FraudShield AI, integrating LSTM networks with graph topological features, achieves robust financial fraud detection on the PaySim dataset, outperforming Logistic Regression and XGBoost base…
Researchers propose a lightweight, inference-time decoding layer for autoregressive generative recommender systems that enables multiobjective slate generation without retraining the model. The method, based on a stochas…
Researchers propose LISA (Linear-Indexed Sparse Attention), a plug-and-play attention module that reduces inference complexity from O(n^2) to O(nM) for long-context reasoning models. Experiments on DeepSeek-distilled-Qwe…
A new study from arXiv introduces AdaRoPE, a method that assigns learnable rotation frequencies and attention scaling factors to each attention head in Transformers, outperforming standard Rotary Position Embedding (RoPE…
Researchers propose Prefix-GRPO, a reinforcement learning framework that improves small language model agents by decomposing teacher trajectories into replay-aligned prefix queries and online continuations, enabling more…
A new study from researchers at arXiv shows that hybrid offline-online reinforcement learning (RL) for vision-language-action (VLA) models achieves out-of-distribution (OOD) performance close to standard RL while requiri…
Researchers introduce Tabula, a privacy-preserving foundation model for single-cell genomics that uses federated learning and explicitly models tabular data structure, along with Chiron, a decentralized AI agent platform…
A new study from arXiv (2607.19382v1) finds that explainability methods such as attention rollout and Grad-CAM can reveal how continual learning models adapt to non-stationary time series, but also expose challenges in i…
Researchers propose STN-TGAT, a model integrating a temporal Transformer with a Graph Attention Network for stock ranking and portfolio construction, which outperforms benchmarks in predictive accuracy and investment pro…
A study by researchers on arXiv (2607.20301v1) finds that PortLLM, a training-free and data-free scheme for adapting large language models after continual pretraining, maintains long-term temporal portability across 10 c…
Researchers introduced PhenSPINE, a Magnetic Resonance Imaging dataset of 16,813 images from 250 patients, to benchmark deep learning models for spine pathology diagnosis. The study found that the Sagittal T2-weighted se…
Researchers propose Thinking Checklist Reward (TCR), a process-oriented reward for reinforcement-learning-based LLM preference alignment that evaluates reasoning traces against sample-specific checklists, improving credi…
Researchers propose a cross-subject semantic decoding framework that aligns neural responses to speech perception from multiple subjects into a shared latent space using the shared response model, then trains a decoder t…
A new benchmark, Air Quality Arena (AQA), covers 6 pollutants over 3 years across 7 countries and 4 continents, with more than 14,000 station-pollutant series, and finds that time series foundation models (TSFMs) consist…
A 2.8M-parameter transformer achieves 0.01-bit entropy agreement with the Bayesian optimum for model selection in Bayesian wind tunnels using fixed-point-free involutions, but fails completely with opaque symbols when ar…
A new study from researchers using graph neural networks (GNNs) predicts groundwater arsenic concentrations across the United States, matching or outperforming gradient-boosted trees by accounting for spatial dependence.…
A new study finds that orthogonalizing the mLSTM memory matrix at read time (five Newton-Schulz iterations, trained through) improves noisy associative recall by re-conditioning the learning problem during a training pla…
Researchers extended neural operator surrogate models for neutron flux estimation from one to two dimensions, testing Fourier neural operators (FNOs) and U-shaped neural operators (UNOs) on one-group transport with isotr…
Researchers propose Spectral-LSH, a training-free prompt compression method that uses Krylov subspace methods and SimHash to reduce long-prompt inference costs by grouping similar tokens into macro-tokens. On Qwen2.5-7B …
Researchers propose Efficient Chain-of-Modality Reasoning (ECoM Reasoning), the first framework to introduce compressed reasoning into spoken language models (SLMs), improving reasoning accuracy by 21% over standard Chai…