Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
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…
A developer built SonicSentinel AI, a dual-model sound-event detection system that pairs a locally trained Python XGBoost classifier with a Google Teachable Machine model, using their disagreements as a signal in the ale…
Xiaomi's MiMo-V2.6 models use groupwise agentic grading to replace binary pass/fail rewards in reinforcement learning, ranking multiple rollouts per task against each other to preserve gradient signal after models pass 8…
Researchers behind the paper "Evolutionary Architecture Search through Grammar-Based Sequence Alignment" (arXiv:2512.04992) use sequence alignment to identify the edits connecting two neural architectures and then recomb…
QLoRA fine-tuning on 150 synthetic examples raised a base Qwen3.5-4B model's adversarial-prompt refusal rate from 79.5% (31 of 39) to 89.7% (35 of 39), while GPTQ and AWQ 4-bit quantization added roughly +0.15 and +0.08 …
Sarvam AI released Saaras V4, a speech-to-text model covering all 22 scheduled Indian languages plus global English accents, available today through Sarvam's API as model="saaras:v4". Sarvam reports state-of-the-art accu…
Weco AI CEO Zhengyao Jiang said in a Machine Learning Street Talk feature that the company's AID system, which applies recursive self-improvement to agentic scaffolding rather than neural weights, produced seven accepted…
A developer built an order risk detector on Laya, an open-source non-autoregressive decision model based on ModernBERT, to test whether a small specialized model can replace large generative LLMs for e-commerce fraud det…
Supersonic Labs, a small AI lab from Brazil, released Julia 1, a 144.3M-parameter open decision model that runs on a plain CPU and is published on Hugging Face under Apache 2.0. Julia 1 starts from JHU CLSP's 140M-parame…
A developer ran an exploratory study testing whether modern LLMs contain internal self-models of other LLMs, using GPT2-medium and Qwen3-4B on 12 post-cutoff news headlines. By having Qwen continue GPT2-generated prefixe…
Dhruva Karkada's paper on arXiv (2602.15029) showed that LLM embeddings of the 12 months project onto a circle with an approximately circulant Gram matrix, a result the author of the post bet a beer he could reproduce wi…
A developer rebuilt an over-engineered parity classifier after finding label leakage and a mislabeled unsupervised framing in the original experiment, moving to a stricter train/validation/test protocol over all integers…
Kuaishou's A/B Agent, an AI system that studied 310 historical recommendation strategies and iteratively proposed production configurations, delivered a 4.829% gross merchandise value increase over the production compari…
Developer zatsepin released jigor, a Rust-based zero-shot classifier runner that executes JEV-style decision models locally via ONNX or remotely through OpenRouter's Decisions API. The tool installs via npm (@zatsepin/ji…
A Towards Data Science experiment found that AI detectors flagged many genuine reviews, and filtering the flagged content made the sentiment model less accurate. The author tested three methods for spotting AI slop in tr…
A developer released FLM (Fly Language Model), a language model trained from scratch on a connectome-derived recurrent network built from a modeled subset of fruit-fly brain wiring, with no pretrained transformer supplyi…
BlueVetaUpright, a purpose-built product-photo orientation system, reached 99.02% accuracy on clean cardinal rotations, 96.67% across the full 360° range, and 84.65% on rotations exactly ±40° from any clean angle, accord…
Flytrex is rolling out a predictive fleet positioning system with its Rooftop Dock in Dallas-Fort Worth that uses machine-learning models to send delivery drones to restaurant rooftops before orders are placed, a setup t…
A paper titled "Learning to Solve Hard Problems in RL for LLMs by Never Giving Up" proposes countering the Matthew Effect in reinforcement learning, where LLMs improve far more on easy problems they already solve than on…
A tutorial demonstrates an end-to-end multimodal data augmentation and adversarial robustness workflow using Meta's AugLy library for images, text, and audio, connected directly to PyTorch datasets and DataLoaders. The w…