Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Shrijith Venkatramana, an engineer building the AI code review tool LiveReview, explains hard negative mining — training retrieval and embedding models on examples that are almost right rather than obviously wrong. He tr…
DeepL trained and served its next-generation large language models in FP8 precision on a 544-NVIDIA H100 Tensor Core GPU DGX SuperPOD, producing translations that outperform its previous models by 1.4x for European langu…
A May 29, 2026 arXiv paper finds that post-training quantization (PTQ) of reasoning models reduces accuracy while increasing chain-of-thought length, and that in up to 52% of quantized models' failures the model reaches …
Supersonic Labs, a small AI lab in Brazil, released Julia 1, a 144.3-million-parameter decision model that selects among 2 to 20 supplied options rather than generating text, trained for roughly US$104.08 in cloud GPU sp…
Typesafe released Jev, a general-purpose classifier, with pricing on OpenRouter listed as Jev 1.13 that approaches the cost envelope of purpose-built ML classifiers while avoiding the large training cycles and training d…
A research team analyzing 769 task logs from building its own AI model found that AI agents supplied up to 55 percent of method proposals, while humans made more than 85 percent of final decisions, and a third of the tas…
A developer built systemonemodels.tech, a registry cataloguing more than fifteen System One model families that read a state and answer typed questions — multiple choice, scale scores, or probabilities — in a single forw…
Sakana AI and the University of Tokyo introduced SAIL (Scaling In-Context Imitation Learning), a method that raises the average success rate of VLM-generated robot trajectories from 25% to 73% by scaling the search budge…
Cymetica's Tuatara Vector Model scored 91.79% on the 3,080-message BANKING77 test set, a statistical tie with the pinned jev-1.13.0 model's 92.40% at roughly a thousandth of the cost, according to the company's blog post…
Master's student Hung Tran of the University of Information Technology (VNU-HCM, Vietnam) is seeking an arXiv cs.CV endorsement for his preprint 'The Clipping Trap: A Silent Failure Mode in Overfitted Neural Image Codecs…
Independent testing found Jev's calibrated-probability claims fail in practice, with Valeriy M reporting calibration failures on 7 of 8 datasets across 16,500 predictions and the author's own coin-toss, two-dice and thre…
A developer built a Simplified Chinese iPhone app, 他爱我吗, that runs a Core ML conversion of the laya-multilingual model entirely on-device to score candidate answers to relationship questions such as affection, no affecti…
Nvidia released Nemotron 3 Diarization, a roughly 100-million-parameter AI model that identifies which speaker is talking at any moment and can distinguish up to eight speakers, with weights freely available on Hugging F…
A study found that indirect aggression is common in Indian conversations on X about sports and political events, and explored whether machine learning could help identify different forms of aggression in Indian X posts a…
Developer mchinen released Tape Heads, an open-source sonification system that plays back the attention patterns of audio transformer encoders including wav2vec 2.0, MuQ, and HuBERT, converting which heads attend to whic…
A developer built an independent benchmark comparing TypeSafe's Jev model against GPT-4, Claude, and Gemini on classification tasks including spam detection, sentiment analysis, and topic classification, with reproducibl…
Google Research's MSEB (Massive Sound Embedding Benchmark), version 0.1.0, is structured in three layers — types (Sound, SoundEmbedding, Score, TaskMetadata), the MultiModalEncoder contract that user models implement, an…
A developer built "Blog vs Bytecode," a Kaggle benchmark of 28 data-science code snippets paired with blog-style claims, half of which hide real methodological errors such as scaler leakage or shuffled time-series cross-…
A developer describes a paired dual-sample / feature-dropout training pattern for LLM rerankers that injects behavioral signals like CTR, QSS and exposure sequences into prompts while forcing the model to also learn sema…
A developer implemented the FlashAttention-2 forward pass twice — first as a deliberately slow PyTorch reference with a Python double loop over tiles, then as a Triton kernel — to teach kernel optimization. The Triton ve…