Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
A study by Chiara Manna, Argentina Anna Rescigno, and Eva Vanmassenhove compared the automated WinoMT annotation pipeline with the instruction-tuned LLM Qwen3-8B for annotating grammatical gender in English–Italian machi…
Vincent Vandeghinste presented a paper at the 1st International Workshop on Teaching AI-Based Translation and Technologies (TAITT 2026) describing a teaching method that replicates the historical development of neural ma…
The Federal Communications Commission is relaunching its Technological Advisory Council on October 1, 2026, with a remit covering spectrum sharing, cellular and wireless systems, and the use of machine learning and artif…
Thinking Machines Lab is hiring a research and finetuning science role in San Francisco at an annual salary range of $350,000 to $475,000, a figure the job board reports as 73% above the $238,000 median for Core ML roles…
Google, Google DeepMind, the University of Maryland, and the University of Virginia published a 2026 technical report describing Dream-RSI, a method that optimizes AI agent search strategies offline by replaying past dis…
A new research paper proposes "register tokens" as a method for bounded-state reasoning in masked diffusion language models (dLLMs), which generate text by iteratively denoising masked tokens with bidirectional attention…
Filippo De Simoni, an MSE Computer Science student at SUPSI, presented a 2026 bachelor thesis project on automatic parallelization of C programs that achieves speedups comparable to ad hoc parallelization with bit-identi…
A new class of dedicated AI judgment models, trained via Reinforcement Learning for Calibrated Decisions (RLCD), is emerging as generative AI hits economic and latency bottlenecks, according to the article. Systems such …
Ted Larson, co-founder and CEO of OLogic, will lead the session "Egocentric Robot Learning: Teaching Robots Through Human Experience" at RoboBusiness, which takes place October 20-21, 2026, in Santa Clara, Calif. The ses…
Cloudflare's Page Shield machine learning model detected all eight malicious JavaScript payloads across four client-side attacks on storefronts, while seven of the eight were entirely absent from VirusTotal and URLScan r…
Comma.ai published a blog post detailing machine learning bugs that degraded its self-driving model's training and inference, including a BF16 plan-output precision issue that limited speed representation to 0.125 m/s st…
ByteLex reported that a tokenizer-free coordinate map built from 11 tokenizer vocabularies over 3-byte windows predicted its 237M-parameter byte-level language model's per-word errors with a Spearman correlation of −0.98…
An independent research project called CetinLM, built under Me Force Technology, has trained a language model past the 2.05 billion token milestone entirely on a single 16GB NVIDIA RTX consumer GPU in a standard home des…
NVIDIA's Resiliency Extension (NVRx) can be integrated into PyTorch Fully Sharded Data Parallel training on Amazon Elastic Kubernetes Service to cut idle GPU time, with synchronous checkpointing consuming up to 40% of to…
An independent developer released a reverse-engineered, open starter model called Jev-like that scores a changing list of N text options in a single pass, matching the input and output shape of TypeSafe's unpublished com…
A team of researchers created their own contemporary model papyrus scrolls, charring them like the originals, to validate a screening method for determining which of the roughly 600 Herculaneum scrolls were written in le…
A developer built MedScribe, an offline clinical documentation pipeline that pairs FasterWhisper ASR with LLM-based medical term extraction, after finding that standard confidence scores cannot reliably gate auto-correct…
A new system called RelateAnything performs real-time open-vocabulary relation prediction from any inputs, according to the research announcement. The work targets a gap in scene-graph modeling, where models remain train…
TypeSafe AI opened early access to Jev, a non-LLM "System One Model" that returns typed probability distributions instead of text, priced at $0.042 per million input tokens with output free. Built by Diogo Almeida, a co-…
AI compilers differ fundamentally from classical compilers because they operate over a small set of recurring model families — transformers, convolutional networks, diffusion models, and mixture-of-experts — and because …