Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers propose SAGE (Selective Agent Guidance via Entropy), a framework that queries a Vision-Language Model (VLM) only when a learner is uncertain and distills the guidance into a lightweight Reinforcement Learning…
A new arXiv study (2609.00879v1) proposes a two-stage training framework combining Supervised Fine-Tuning (SFT) and Direct Preference Optimization (DPO) to improve multimodal disaster damage assessment, raising accuracy …
Researchers propose CARE (Contrastive Anchor-based Rubric Evolution), a method that uses anchor responses from frontier models to dynamically evolve rubrics during LLM post-training, addressing reward hacking in open-end…
Researchers propose ARISE-RL, a self-evolution framework that couples a task/rubric Generator and a reasoning Solver through rubric-mediated co-evolution to train open-ended agents via reinforcement learning, addressing …
Researchers introduced a framework for evaluating behavioral fidelity in long-horizon activity simulations, finding that statistical priors bring activity and sequence distributions closest to real behavior but over-frag…
Researchers propose H2Table, a hierarchical hypergraph-enhanced framework for complex table reasoning, which represents tables as nested hypergraphs and uses a tailored encoder with learnable query vectors to bridge stru…
A new arXiv preprint (2609.01260v1) introduces a dual-process motion planning framework that combines symbolic solvers with learning-based modules, inspired by the 'Thinking Fast and Slow' paradigm. The architecture, fea…
A new research paper introduces SMELT, a study of scaling laws for compute-matched Mixture-of-Experts (MoE) looped Transformers, finding that looping provides architectural advantages beyond mere FLOPs. The authors compa…
AWS is updating its Machine Learning Engineer – Associate certification exam from MLA-C01 to MLA-C02, with the new version placing significant emphasis on generative AI and Amazon Bedrock. The exam now requires candidate…
A developer built Base2026, a source-backed expert-video search tool on Cloudflare, using D1, R2, Queues, Workflows, and Workers AI for transcription. The platform offers deterministic retrieval with attributable finding…
A developer published a 2026 roadmap for becoming an AI engineer, emphasizing strong programming, mathematics, and machine-learning foundations before adopting modern frameworks. The guide covers Python, math, ML, deep l…
Baseten's technical breakdown of LLM inference efficiency identifies two categories of engineering choices: those that trade latency for throughput, such as batch sizing, tensor parallelism, expert parallelism, and atten…
A 3D Spatial Transformer Network implementation for Torch, derived from the 2D version by @qassemoquab, is released with modules for affine 3D grid generation and trilinear sampling in BTHWC layout. The code includes an …
A developer has built a machine learning bot for the traditional board game Mancala using agentic coding and a Python-based game simulator. The project addresses the game's player-one bias by introducing random stone dis…
Researchers at the Norwegian University of Science and Technology (NTNU) used artificial intelligence to analyze health data from more than 43,000 participants in the Trøndelag Health Study (HUNT) and found that migraine…
Researchers introduced ReFlowSET, a representation-aligned latent flow matching method for SAR-to-EO image translation that improves reconstruction fidelity across codecs and modalities by aligning latent representations…
A developer used Antigravity CLI, powered by Gemini and the tanaike-lab framework, to analyze 29,334 Netatmo weather station records in Japan, reconstructing volcanic Lamb wave velocities with 98.5% precision and localiz…
A technical article outlines a multi-layered approach to AI alignment and security, emphasizing the need to combat reward hacking in RLHF and move toward mechanistic interpretability for deterministic safety. It details …
A developer at forge.workstyle.tech discovered that AI voice models sometimes elongate greetings like 'Kon'nichiwa~' due to a flaw in the corpus generation pipeline. The normalization process, which removes prolongation …
Meta is releasing Muse Voice Transcribe, a real-time speech model that combines streaming transcription, speaker diarization, and endpointing, capable of handling 20 or more speakers and code-switching, with API access p…