Machine Learning News

Researchers used a developmental approach to study how neural language models learn statistical patterns, finding that Transformers first acquire abstract global statistics before local dependencies, with early over-gene…

Researchers introduced DMV-Bench, the first interactive benchmark for multimodal-agent visual memory, testing agents' ability to recall visual cues from product images in a simulated e-commerce environment. Their propose…

A study evaluating boundary-condition fidelity for bottom-hole pressure and CO2 plume prediction in geological carbon storage found that conserving corner pore volume is critical for accurate truncated-domain simulations…

Researchers introduced Smooth Maximum Mean Discrepancy (SMMD) to improve numerical prediction accuracy in large language models by aligning predicted numeric distributions with targets using distance-based kernels and gr…

Researchers introduced QDiffusion-TS, the first quantum generative diffusion model for time series synthesis, validated on the IQM quantum processor. The model reduces trainable parameters by nearly three orders of magni…

Researchers introduced PEBS, a per-rater empirical-Bayes shrinkage estimator for calibrating reward models in RLHF, which reduces within-user held-out RMSE by 8.58% on PRISM and 9.66% on PluriHarms compared to the pooled…

Researchers introduced PairSAE, a method for mechanistic interpretability of protein co-folding models that summarizes pairwise tensors into token-level features, avoiding the quadratic blow-up of standard sparse autoenc…

Researchers introduced Grounded Iterative Language Planning (GILP), a hybrid method combining a small parameterized world model with LLM-based agent reasoning, reducing hallucinated-state rates from 0.176 to 0.035 on gra…

Researchers at arXiv identified a memory-update gap in LLM agents, where accuracy drops from 92% to 77% when using bounded memory instead of full context, even with frontier models like gpt-5.4. The gap persists across m…

Researchers introduced the Agent-Native Immune System (ANIS), a biologically inspired defense architecture embedded within autonomous agents' cognitive loops to counter runtime attacks like memory poisoning and tool-chai…

Team HSA_CORAL submitted to the FinCausal 2026 shared task on extracting cause-effect relations from financial narratives via extractive question answering in English and Spanish. Their best system, GPT-4.1 Mini fine-tun…

Researchers propose a three-stage training paradigm to internalize future-aware planning in LLM agents, enabling them to simulate outcomes before acting. The approach, evaluated on search and math reasoning tasks, outper…

A new position paper argues that the term 'machine unlearning' is overused in large language model research and should be reserved for dataset-defined deletion. The authors contend that many tasks labeled as unlearning, …

Researchers at arXiv have discovered that activation patching, a key tool in mechanistic interpretability, contains hidden interaction effects (INT) that can distort causal attributions. These effects, which measure how …

Researchers have developed OverFlowLight, a real-time framework that uses multi-modal sensing from cameras and radars to detect and prevent queue overflow at urban intersections. Deployed across 43 intersections in three…

Researchers propose CoIn, a novel framework for 3D scene inpainting that bridges 2D inpainting models and Gaussian Splatting through a multi-stage consistency pipeline. The method handles both object removal and insertio…

Researchers introduced Masked Language Flow Models (MLFMs), combining masked diffusion and flow-based methods for efficient language generation. MLFMs enable conditional generation via continuous flows and support multi-…

Researchers propose LaViD, a framework that transfers semantic knowledge from a large language model to a vision-only student model using multiple-choice questions, outperforming existing distillation methods on fine-gra…

Researchers developed a distribution-based deep multiple instance learning method for tumor proportion scoring in non-small cell lung cancer, using only slide-level labels to predict zero-inflated beta parameters. The ap…

Researchers propose vMFProto, a distributional part-prototype framework that models each class as a mixture of von Mises-Fisher components on the hypersphere for interpretable classification. The method achieves state-of…