Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers have developed Latent Cache Flow (LCF), a method enabling direct model-to-model communication by translating and compressing key-value cache data instead of using text. The approach uses an adapter only 4% th…
Researchers have developed SciAtlas, a large-scale knowledge graph integrating over 43 million papers across 26 disciplines, 157 million entities, and 3 billion triplets to serve as a structured cognitive map for automat…
Researchers have introduced Human-Centered Learning Mechanics (HCLM), a dynamical and information-theoretic framework that treats deep learning as an open system under uncertainty, resource constraints, and human feedbac…
Researchers have developed FederatedRSF, a Python package enabling multi-center survival prediction using federated random survival forests without sharing patient-level data. The method aggregates locally trained surviv…
A study of fMRI data from 112 English, Chinese, and French speakers found that brain-LLM alignment is driven by training-language dominance, not an inherent property of English. A Chinese-dominant model reversed the alig…
Researchers introduced HawkesLLM, a framework that models how uncertainty propagates through sequential text generation in agentic simulation systems. The framework uses a multivariate Hawkes process to track temporal in…
Researchers tested seven large language models on their ability to infer employee expertise from 27,188 Slack messages, finding that Gemini 2.5 Flash achieved the lowest estimation error at 21.13% mean absolute error. Th…
Researchers have found that chain-of-thought reasoning in large language models is only beneficial when early-stage entropy dynamics show consistent reduction, according to a new study on arXiv. The team introduced EDRM,…
A new online calibration method called MARGIN corrects systematic mis-calibration in foundation model confidence scores, achieving 3-6x lower calibration error than existing design-time methods under distribution shift. …
Researchers have developed CoMoGen, a controllable video generation framework that creates realistic motion dynamics and interactions from a single binary mask sequence and an input image. The system uses a lightweight M…
A new study of small language models reveals that chain-of-thought prompting for arithmetic relies on a positional shortcut: the model copies whichever number appears last before the answer delimiter, regardless of the l…
A developer has published a conceptual blueprint for AURORA, a low-energy cognitive organism architecture that operates without transformers or token prediction. The system uses sparse predictive microfields, episodic me…
The discrete GPU market in May 2026 is defined by three distinct strategies, with NVIDIA dominating the high-end and AI/ML segments via its RTX 50-series Blackwell architecture, while AMD offers the best overall gaming v…
A new glossary from Hugging Face aims to clarify the terms "harness" and "scaffold" in AI agent development, following confusion among researchers at the ICLR 2026 conference. The glossary defines a "harness" as the exec…
Researchers at HASH built a 25M-parameter transformer decoder, vec2slug, that generates URL slugs directly from vector embeddings, achieving 14-19x faster and 85x cheaper performance than a Haiku-class LLM. The model ext…
Memory's share of AI chip component costs rose from 52% to 63% between Q1 2024 and Q4 2025, according to an analysis of Nvidia, AMD, Google, and Amazon chips. Total component spending on AI chips grew from approximately …
Researchers at Apple have developed PICO (Perceptual Image Codec), the first learned image compression model optimized for human visual perception and practical on-device deployment. The codec achieves 2.3-3× bitrate sav…
The author describes being inspired by Google I/O 2026 to create a Telugu-language AI assistant, similar to "Jarvis." They argue that AI should be accessible in regional languages like Telugu to help millions of Indian s…
Here is a factual summary of the article: The article details a stress test of the Gemma 4 E4B model's 128K context window on a laptop GPU (RTX 5050). The test found that while the model's recall of information within t…
According to a May 11, 2026 research preview from Thinking Machines, the company has developed a new class of "interaction models" that process audio, video, and text in continuous 200-millisecond ticks, eliminating the …