Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
A developer has built Echo, a system that uses open-weight models including GLM-5.2 and Kimi K2.7 to achieve results comparable to Fable at roughly one third the inference cost. Echo dynamically allocates computation and…
Grabette, an open-source system for recording robot manipulation data, bridges the gap between raw robotic movement and structured data formats required for modern machine learning. The system allows users to integrate r…
Ava Amini, principal researcher at Microsoft Research and co-lead of Project Ex Vivo, is leading a collaboration with the Broad Institute and Dana-Farber Cancer Institute to use AI and machine learning to model cancer as…
Apple's MLX team introduced distributed inference and training capabilities for machine learning workloads across multiple Macs at WWDC26, demonstrating how to scale large language models using RDMA over Thunderbolt 5 in…
A research team hand-coded weights for one-layer MLPs that memorize labels for input token sequences of length two, finding that the number of facts memorized with 90% accuracy scales roughly linearly with parameter coun…
A first-year undergraduate student in Artificial Intelligence Engineering is building a Retrieval-Augmented Generation (RAG) project from scratch, documenting the process in a series of blog posts. In the first three par…
Helion, PyTorch's high-level DSL for writing performance-portable ML kernels, partnered with Google to build a TPU backend that compiles Helion kernels to Pallas, enabling PyTorch-friendly TPU kernel authoring. On a flas…
Primate Labs released Geekbench 7 on July 23 for macOS, iOS, Android, Windows, and Linux, adding AI, media, gaming, and content-creation workloads. The benchmark uses a new 2,500-point Ryzen 7 7700 reference baseline and…
A developer replaced a 7B parameter LLM running on a GPU with a 2.4 MB Go classifier for email triage, achieving sub-millisecond CPU latency and 100% reliability on deterministic rules. The three-layer filter uses determ…
Primate Labs launched Geekbench 7, a benchmarking suite with redesigned multi-core and GPU benchmarks that focus on real-world application performance and machine learning tasks. The GPU benchmark includes tests for face…
The llm-d project introduces co-operative time-slicing to interleave independent reinforcement learning jobs on shared hardware, increasing aggregate accelerator duty cycles from ~40% to 70% without impacting model conve…
A developer outlines a 3-5 month plan to build a deep learning framework from scratch, culminating in a 'Tiny Transformer,' arguing that implementing an autodiff engine, tensor engine, and optimizer from the ground up pr…
Developer Jeongseop Byeon released s3fifo v1.0.0, a zero-allocation S3-FIFO cache for Node.js that eliminates garbage collection pressure by using pre-allocated TypedArrays. The production-ready release adds features lik…
Deep learning reverses traditional programming by fitting a mathematical function to data and answers rather than hand-writing rules, according to an explanatory article. The approach powers systems like ChatGPT, Claude,…
Triton-blackhole, a deterministic numerical debugger for Triton kernels, identifies where output diverges from a PyTorch reference and classifies drift as benign fp16/bf16 drift or a real bug without using tl.device_prin…
In-context learning (ICL) allows AI models to adapt temporarily using examples in a prompt, but it is not true generalization, according to an analysis of the mechanics and constraints of ICL. ICL relies on pattern compl…
PageWatch.tech has developed a computer vision diff pipeline for web change monitoring that combines SSIM, ORB feature matching, and a Siamese ResNet-18 neural network to handle layout shifts and dynamic content. The app…
A developer launched bike.broker, a tool that uses a multimodal deep learning model (BERT plus tabular features) to predict used bike prices by factoring in custom parts and upgrades, achieving an ~18% mean absolute perc…
A developer built raven-retrieval, an open-source framework that benchmarks 19 retrieval pipelines on BEIR datasets with Bonferroni-corrected significance tests. The key finding is that top pipelines like HyDE and Naive …
Geospatial code can execute without errors and produce clean-looking results while being methodologically wrong, a problem amplified by AI coding agents generating GIS workflows from natural-language instructions. Develo…