Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers from arXiv identify a structured confound in Reinforcement Learning from Human Feedback (RLHF) where pairwise preference labels may reflect the rater's state during annotation rather than just output quality,…
Researchers propose DocOCR-Eval, an annotation-free framework for automatically evaluating and selecting OCR tools without ground-truth labels. The framework uses a three-stage correction and ranking strategy that aggreg…
Researchers propose W2SPO, an off-policy reinforcement learning method that uses short auxiliary segments from a weaker model to improve reasoning in large language models, achieving a Pass@1 improvement from 62.3% to 64…
Developers can demystify AI by understanding that AI, ML, deep learning, and LLMs are nested concepts, not synonyms, and that training and inference are the two halves of a model's life, with most app developers working …
A developer compares PyTorch and TensorFlow for computer vision projects, noting that PyTorch dominates research with dynamic computation graphs and recent performance improvements via torch.compile(), while TensorFlow l…
A developer published ML-pokedex, a set of four browser-based experiments training language models on RDF data from a single Pokémon knowledge graph instead of English text. The models memorize real Pokémon facts and can…
A developer discovered that calling optimize() on a ZVec collection recovered from an unclean process termination silently discards documents restored by WAL replay, reporting index_completeness: 1.0 while doc_count regr…
Researchers from MIT and other institutions have developed iTracker, a convolutional neural network for eye tracking that works on commodity mobile phones and tablets without additional sensors, achieving a prediction er…
AIBuildAI Agent 2.5 achieved #1 on OpenAI's MLE-Bench with a score of 70.7%, substantially outperforming the agents ranked 2nd through 5th. The AI agent, which automatically builds AI models, also ranked in the top 5.7% …
A Flutter developer built vector_kit, a Dart library that accelerates semantic search by storing embeddings as a packed Float32List and using SIMD dot products, achieving 11x faster top-k cosine similarity queries over 2…
An artificial neuron is a simple math function with four components: inputs, weights, a bias, and an activation function. Weights are initialized to small random numbers to break symmetry, and the bias shifts the neuron'…
Researchers Anna Manaseryan and Casey Kennington presented an online reinforcement learning framework at the 27th Annual Meeting of the Special Interest Group on Discourse and Dialogue that adapts robot emotional behavio…
Researchers from Sorbonne Université and CNRS have developed the first reinforcement learning-based dialogue manager for simulated group motivational interviewing, using a discrete Soft-Actor-Critic policy to select ther…
Researchers at an undisclosed institution introduced CalibAtt, a training-free method that accelerates text-to-video generation by up to 1.58× via calibrated sparse attention. The method identifies and skips negligible t…
The open-source Grabette system, developed by an unnamed team, enables recording robot-manipulation data using a handheld gripper with two cameras, eliminating the need for a robot or teleoperation rig. The system output…
AMD ROCm software provides an open-source stack for programming AMD GPUs, enabling high-performance video inference pipelines. A tutorial demonstrates building a pipeline using rocDecode for hardware-accelerated video de…
Prince Canuma, maintainer of the mlx-vlm library, released Nativ, an MIT-licensed macOS app that provides a polished GUI for running local MLX models and exposes both OpenAI-compatible and Anthropic-compatible API endpoi…
Embeddings are fixed-length vectors of floating-point numbers that encode semantic meaning, enabling similar sentences to be mathematically close and powering RAG, semantic search, and recommendation systems. The author …
Algorithm aversion is defined as a biased assessment of an algorithm that manifests in negative behaviors and attitudes compared to a human agent, according to Wikipedia. The phenomenon leads people to reject algorithmic…
D-FINE-seg, a framework for real-time object detection, instance segmentation, and semantic segmentation, achieves higher F1 scores than YOLO26 and RF-DETR on Cityscapes detection and instance segmentation, and leads mIo…