Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Neural Cellular Automata (NCA) extend Conway's Game of Life by learning neural network weights to solve problems like MNIST digit classification, maze solving, and playing Pong. NCA use local receptive fields and recurre…
Researchers Du et al. published CosyVoice 2, a unified model for streaming and offline speech synthesis combining an FSQ semantic tokenizer, a Qwen2.5-0.5B text-speech language model, and a chunk-aware causal flow-matchi…
A comprehensive research report on 2026 online marketing strategies reveals that AI-driven acceleration and privacy constraints have widened the performance gap between top and bottom lead acquisition programs to 4.7x, w…
Boson AI released Higgs Audio v2, a speech synthesis model based on a Llama-3.2-3B architecture with a forked audio feedforward network that generates interleaved text and delay-patterned residual vector quantization cod…
Kokoro, an 82M non-autoregressive StyleTTS2 + ISTFTNet model, converts phonemes and a split style vector into 24 kHz audio in a single feed-forward pass without LLMs, diffusion, codecs, or tokens. Published by hexgrad in…
DeepGate's automated model design system, combining neural architecture search, a compiler, and real-hardware measurements, produces models for microcontrollers that run up to 45× faster and use up to 11× less RAM than r…
DeepGate released v0.15.0 of its edge AI compiler, which compiles quantized .tflite models into static binaries that use up to 3× less RAM and run up to 2× faster than Google's TensorFlow Lite for Microcontrollers on Arm…
A developer released PaneTrans, a browser extension for drag-select region translation and OCR on video/canvas, built on Transformers.js with local processing by default. The tool uses an offscreen document for inference…
Sasha Rush released GPU Puzzles, an interactive notebook teaching beginner GPU programming through coding exercises that map Python to CUDA kernels using NUMBA. The puzzles aim to build intuition for GPU architectures cr…
Researchers from Renmin University of China and Microsoft Research introduced Arbor, a persistent hypothesis tree that helps AI coding agents remember and refine learnings over long research sessions. In tests, Arbor del…
Researchers published a preprint on ResearchGate detailing a machine-learning model to predict hepatic steatosis severity in nonobese individuals, addressing a screening gap where steatotic liver disease affects 40% of t…
Researchers from Sina Weibo Inc (China) released VibeThinker-3B, a 3-billion-parameter dense reasoning model built on Qwen2.5-Coder-3B using the Spectrum-to-Signal post-training pipeline. The open-source model matches th…
AI progress is driven primarily by massive amounts of domain-specific human expert data and reinforcement learning, not by architectural breakthroughs, creating a data black hole that powers model capabilities. Open-sour…
Meta's LLMs have evolved from simple Transformer stacks to complex architectures with multiple attention variants, mixture-of-experts, and multimodal encoders, mirroring the complexity of recommendation systems. The indu…
Nobel Prize-winning chemist John Jumper is leaving Google DeepMind for Anthropic, as reported by Reuters on June 19, 2026. Jumper, known for leading the AlphaFold project that predicted protein structures, joins Anthropi…
Researchers at UC Berkeley have developed a 16-element electronic nose chip that combines gas sensors and machine learning to identify food types, spoilage, and allergens, detecting as little as 0.05 grams of walnut mate…
The open-source NVIDIA NVK Vulkan driver merged support for Deep Learning Super Sampling (DLSS) into Mesa 26.2-devel, enabling DLSS on Linux/Steam Play for modern games. The implementation relies on the VK_NVX_binary_imp…
OpenAI announced a new AI tool designed to assist physicians in diagnosing rare genetic diseases in children, leveraging machine learning to analyze medical data and suggest potential conditions. The tool aims to reduce …
Dropout, a regularization technique that randomly removes neurons during neural network training, helps prevent overfitting by forcing the network to distribute learning across multiple pathways, leading to better genera…
SunPy, a Python library for solar physics, enables robust analysis of solar flares by handling massive data volumes from instruments like the Solar Dynamics Observatory. Its Map object binds image data with coordinate me…