Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Fulcrum Research's Fable AI agent achieved a 7.6% improvement over the human state-of-the-art for CIFAR-10 training speed, reaching 94% accuracy in 1.828 seconds on a single A100 GPU, while Claude Opus 4.8 and GPT 5.5 fa…
Berlin-based reverse.fashion, a TU Berlin spin-off using AI for used-textile sorting, secured a seven-figure pre-seed extension from High-Tech Gruenderfonds. The funding will support commercial expansion of its computer …
A new project turns Ilya Sutskever's famous list of 30 papers for John Carmack into AI-narrated audio episodes, explaining key insights from foundational deep learning works. The list, which Sutskever said covers 90% of …
Google released TabFM on July 1, a zero-shot foundation model for tabular data that generates predictions without training on the user's dataset, using in-context learning. The model achieves state-of-the-art performance…
Hussein Mahdi published a review of the 2009 paper on graph neural networks, covering their theoretical foundations and core mechanisms. The article explores the history, concepts, and philosophy of GNNs.
Google announced LiteRT.js, a JavaScript binding of its LiteRT inference library that enables high-performance AI model execution directly in web browsers via WebAssembly. The tool supports hardware acceleration across C…
Microsoft Research released Aurora 1.5, an updated Earth-system foundation model for weather forecasting, adding 22 weather variables, hourly temporal resolution, and probabilistic ensemble forecasting. The model outperf…
A developer built hermes-harness, a CLI tool that mines git history for recurring failure and fix patterns, compressing them into a searchable 128-dim vector space. Inspired by Yann LeCun's world model concepts, the tool…
A developer explains how Knowledge Distillation (KD) enables high-performance AI on Android by transferring 'dark knowledge' from large Teacher models to compact Student models. The technique uses soft targets and temper…
Researchers at Lighthaven tinkered with Natural Language Autoencoders (NLAs), finding that while they reconstruct activations well and enable auditing of model organisms, the architecture or training process may need cha…
Google Research introduced SensorFM on July 9th, a wearable health foundation model trained on over one trillion minutes of sensor data from five million consented Fitbit and Pixel Watch users. The model uses self-superv…
Trelix evolved from v1.0 to v2.7 in thirteen days, fixing a critical config bug and adding a knowledge graph, seven retrieval legs, agentic loops, federated multi-repo search, and a GitHub Actions bot. The knowledge grap…
A developer has compiled a free, open-source syllabus for mastering AI and data engineering in 2026, arguing that traditional bootcamps are unnecessary. The curriculum covers data architecture, analytics, and AI engineer…
Sportradar is using Wimbledon to showcase AI-driven sports data, with real-time match feeds powering broadcast graphics, fan personalization, coaching analytics, and commercial data-rights products. The company extended …
PLOS Computational Biology published a study on July 9, 2026, presenting an embedding-guided method that uses protein-language-model embeddings with Monte Carlo sampling to navigate enzyme sequence space while avoiding i…
Neurotech startup Aleph published a silent-speech prototype that uses submental ultrasound to image tongue motion and convert it into text, achieving a 15.6% word error rate on an open-vocabulary test. The system, traine…
HubSpot published a guide comparing top sales analytics platforms for 2026, contrasting CRM-native reporting with general BI tools and highlighting AI-powered forecasting and integrations with HubSpot's Sales Hub and Bre…
A researcher developed a physics-informed AI method that integrates a PDE residual into the training loop of a language model by using a differentiable numerical head, enabling the model to learn physics from text prompt…
A developer outlines inference optimization strategies for small-to-mid-scale deployments, focusing on KV cache management, quantization, and latency tradeoffs. The post provides practical guidance for single-GPU or CPU …
A developer introduced Lemma, an open-source MCP server that integrates coding agents with live Jupyter kernels to enforce reproducible data science and machine learning workflows. The tool configures multiple agents in …