Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Apple Machine Learning Research published a paper proposing that machine unlearning skip training points with negligible influence on model outputs, reporting computational savings of up to approximately 50 percent on re…
Independent researcher Taotuner has proposed a methodological framework called History-Dependent Dynamics (HDD) to disentangle history dependence, recurrence, self-reference, and self-modeling in dynamical systems. The f…
Meta Platforms Inc. has launched Scam Alert, an optional AI-powered feature in WhatsApp that uses on-device machine learning to flag suspicious messages from non-contacts without sending chat data to Meta's servers. The …
A developer overhauled their hybrid retrieval pipeline for technical documentation, combining Qwen embeddings, BM25 sparse vectors, and reciprocal rank fusion with a FastEmbed reranker. The upgrade improved retrieval acc…
Recordari, a custom memory graph built by developer Corby, scored 0.831 on the LoCoMo benchmark, outperforming Microsoft Research's Memora, which published a score of 0.801. The benchmark, run with the same model and dat…
Meta is launching an optional Scam Alert feature on WhatsApp that uses on-device machine learning to flag suspicious messages, rolling out in a limited beta. The feature shows users a warning if a chat seems like a scam,…
A layoff can be the best thing for a creative career, according to a developer who observed a coffee shop owner's method of documenting customers. The owner keeps a hidden digital guest book, recording names, dates, drin…
A study published in Translational Psychiatry found that early changes in symptom scores after a single 20-milligram dose of methylphenidate accurately forecast long-term medication success in adults with attention-defic…
AMD and PyTorch upstreamed FP8 training optimizations for AMD Instinct GPUs into TorchAO and TorchTitan, delivering a 13.4% throughput gain over BF16 on Llama3-8B dense models and recovering 89% of FP8 quantization overh…
Liquid AI released LFM2.5-VL-3B, a 3.1B-parameter vision-language model for on-device deployment, averaging 69.4 across 28 vision benchmarks, matching InternVL-3.5-4B and 0.7 points behind Qwen3.5-4B. The model reads scr…
Rust compiled to WebAssembly is replacing JavaScript for heavy AI workloads in the browser, with local inference of 1B+ parameter models via WebGPU and WASM SIMD becoming standard in 2026. Benchmarks show Rust WASM SIMD …
Backblaze has published daily health readings of every hard drive in its data centers since April 2013, and the open dataset has been cited in more than 227 papers and articles since 2018, powering research in AI/ML incl…
AMD released MiniDXNN v0.4.0, an open-source library for GPU-accelerated MLP inference and training on DirectX 12, adding an interactive GUI application for neural texture compression. The update includes a windowed app …
Chinese researchers have developed a lightweight AI system that identifies heat signatures of F-22 and F-35 stealth fighters with over 90% accuracy in laboratory tests, potentially making flares ineffective in future air…
An audit by computer science researchers finds that NVIDIA's Blackwell Ultra GPU (B300) nominally supports INT8 tensor-core compute at a 30:1 FP8-to-INT8 ratio, but the format is undeployable by default due to a layered …
MongoDB shipped three retrieval features this week: context-aware embeddings, hybrid search, and a native reranker. The developer behind csvql argues that while these features address unstructured data, structured data l…
Voyage AI released voyage-code-4, a code embedding model for coding agents, which outperforms Cohere Embed v4 and Gemini Embedding 2 by an average of 28.25% and 31.03% on a new agentic code retrieval benchmark, and by 19…
DeepSWE, a new long-horizon software engineering benchmark, reports that Grok 4.6 /medium outperforms /high effort on its leaderboard, which measures frontier coding agents on original tasks across 91 repositories and 5 …
Ryanair has added Google Cloud to its existing Microsoft Azure cloud setup, creating a dual-cloud strategy to enhance operational resilience and scalability for its digital services. The airline will use Google Cloud's i…
VectorWare has enabled Rust's portable SIMD (core::simd) to run natively on GPUs, allowing the same SIMD code to execute on x86 CPUs, ARM, and NVIDIA GPUs without modification. This breakthrough maps GPU warps directly t…