Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
A developer built a transformer from scratch and computed a full forward pass by hand to understand the architecture introduced in the 2017 paper 'Attention is All You Need' by Vaswani et al. The transformer, similar to …
Feature selection is a critical but underused step in machine learning, as irrelevant features waste compute, invite overfitting, and hide signal behind noise, according to a technical article. The article categorizes te…
Researchers from Tel-Aviv University, DisneyResearch|Studios, and ETH Zurich introduced a generative diffusion framework with a dual-objective loss function that decouples skeletal integrity from drumstick precision, ena…
A data scientist at an unnamed company cut AI API costs by 95% by analyzing six months of logs and implementing a model-routing pipeline that matches each request to the cheapest adequate model. The analysis revealed tha…
Anthropic reported that its Claude AI models, including the Mythos Preview and Opus 4.8, successfully designed protein binders against 14 of 15 targets, with 22% to 35% of individual designs binding successfully compared…
A developer has built a depression detection engine using Python and OpenSMILE, extracting acoustic biomarkers from speech to identify indicators of depression. The system leverages the eGeMAPS feature set, speech rate a…
The U.S. Department of Energy's National Laboratory of the Rockies (NLR) hosted nearly 250 undergraduate interns this summer, including Anisha Jarang from the University of Virginia, who is working on optimizing charge m…
A developer detailed a rule-of-thumb method for estimating GPU memory requirements for AI models, using Qwen2.5-7B-Instruct-AWQ as an example. The method sizes weights and KV cache from a model's spec sheet, then compare…
Inco AI released DFlash 2, a parallel speculative decoding technique that delivers over 20% more output from every verification pass with around 1% added cycle latency, achieving 2.7–3.4× throughput of autoregressive dec…
Automatic subtitling powered by AI has become viable for production companies in 2026, with OpenAI Whisper v4 achieving over 97% accuracy in clean audio conditions for major languages and AssemblyAI offering enterprise f…
Developer atomic14 released an updated AI-text detector at meatbag.atomic14.com that can now identify both Claude Opus 5 and ChatGPT output, achieving 93.5% accuracy on held-out authors for ChatGPT and 92.5% for Claude. …
Artificial intelligence is accelerating automated biolabs, or biofoundries, enabling faster design-build-test-learn cycles in synthetic biology, according to Knowable Magazine. The technology helps researchers optimize g…
ColBERT-style multi-vector models, such as colbert-ir/colbertv2.0, are now integrated into the sentence-transformers library, enabling high-precision RAG pipelines through late interaction and MaxSim matching. The approa…
An ESP32-S3 microcontroller now runs a Neural Amp Modeler A2-Lite capture natively in single-precision float, producing bit-identical output to a desktop render, with all 720,000 samples matching by hash and digital subt…
Zacharie Rodière, a recent MS ECE graduate from Georgia Tech, released a keystroke biometrics demo through his startup BehavLabs, featuring two-player and 3+ player modes for continuous authentication. The model, trained…
Darktrace's State of AI Cybersecurity 2026 report reveals that 77% of security stacks now include AI, yet trust and understanding lag behind adoption. The survey found 67% use supervised machine learning, 67% use agentic…
Snorkel AI, an enterprise AI company founded as a Stanford AI Lab research project in 2015, is hiring an AI Researcher Intern in New York City for a hybrid, up-to-six-month internship. The role involves conducting origin…
Databricks introduced Precision Mode in its AI Extract document extraction API, claiming state-of-the-art accuracy on complex enterprise documents. In benchmarks spanning roughly 9,000 documents, Precision Mode outperfor…
A kernel is a single function that runs on a GPU, and the steep memory hierarchy—with a 1,600x gap between L2 cache and HBM—makes kernel design crucial for AI performance. FlashAttention, developed by researchers includi…
DoorDash is hiring a Staff Machine Learning Scientist for Applied Causal Inference to build causal ML systems for its New Verticals business, with a salary range of $204,000 to $299,000 per year. The role, based in San F…