Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers have demonstrated that retrieval-augmented generation (RAG) can be understood as an in-context optimization process, where a single linear self-attention layer performs a gradient-descent step on a unified RA…
A new study reveals that large language models hallucinate on structured knowledge tasks due to systematic internal failures, not random errors. Researchers found that attention mechanisms disproportionately focus on sho…
Researchers introduced EnterpriseMem-Bench, a multi-turn Text-to-SQL benchmark of 300 sessions and 1,400 queries built from three enterprise domains, to evaluate how five frontier AI models handle memory across conversat…
Researchers have developed Token-to-Mask (T2M) remasking, a training-free replacement for Token-to-Token (T2T) editing in discrete masked diffusion language models like LLaDA. The method resets suspected erroneous tokens…
Researchers introduced Anchor, a task-generation pipeline that prevents artifact drift in AI agent benchmarks by formalizing business workflow specifications into constraint optimization programs. The system jointly prod…
A new survey reframes alignment tuning for large language models as a data pipeline design problem, decomposing data construction into three stages: response synthesis, preference evaluation, and preference instantiation…
Researchers introduced AgingBench, a longitudinal benchmark designed to measure how AI agents degrade over time after deployment. The study found that agent reliability declines through four distinct aging mechanisms—com…
Researchers have developed POLAR, a multimodal memory-augmented framework that enables embodied AI agents to personalize assistance over long-term user interactions by organizing prior experiences into a knowledge graph.…
A new study reveals that language model reasoning trajectories during test-time sampling cluster into "reasoning basins," causing majority votes to favor stable but potentially incorrect answers. The researchers introduc…
Researchers have proven a finite-sample bound on the approximate max-information of differentially private stochastic gradient descent (DP-SGD) that scales at most linearly with dataset size, matching classic results for…
Researchers have developed AIMS-Fold, a guided-diffusion framework that integrates structural proteomics data from XL-MS and HDX-MS to improve protein complex structure prediction. The model outperforms purely computatio…
Researchers have developed a neural rule evaluator that detects anomalies by identifying violations of logical constraints, even when such violations are absent from training data. The system compiles rules into directed…
Researchers have introduced FEPoID, a training-free method for automatically selecting intermediate layers in large language models to improve hallucination detection. The approach outperforms existing criteria and detec…
Researchers have developed Geometry-Aware Representation Denoising (GARD), a framework that uses diffusion-based restoration within a feed-forward 3D reconstruction model's feature space to improve multi-view 3D reconstr…
Researchers have developed CE-FedGNN, a federated graph neural network framework that enables organizations to collaboratively learn from distributed graph data without sharing raw data or frequent embedding exchanges. T…
Researchers have developed Lie group embedded dynamical neural networks (LieEDNN) that enable stable, learnable dynamics on smooth manifolds by treating Lie groups as intrinsic representations of continuous symmetry. The…
Researchers have developed PaGeR, a framework that adapts existing 3D foundation models designed for perspective images to reconstruct full 360-degree scenes from single panoramic images. The model predicts scale-invaria…
Researchers have developed DecoupleGen, a method that uses personalized text-to-image diffusion models to generate training images of objects in uncommon contexts, such as a beach ball on a road. The approach aims to cor…
Researchers have developed CoAD, a new framework that unifies classification and reconstruction approaches to improve time series anomaly detection. The framework uses a classification module to generate probability-info…
Researchers have developed BEAP, a black-box adversarial prompting attack that exploits vulnerabilities in text-to-image diffusion models that have undergone machine unlearning. The attack uses a large language model to …