Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Researchers propose Multi-teacher On-Policy Distillation (MOPD), a post-training paradigm that combines the capabilities of multiple domain-specific reinforcement learning teachers into a single large language model by d…
A new analysis warns that the AI industry is hitting a 'data wall' as high-quality human-generated text becomes scarce, with Epoch AI projecting exhaustion between 2026 and 2032. In response, synthetic data is surging—Ga…
A developer built Tensorless, a PyTorch-based package that automates configuration of language model training, allowing users to train text-generation models with minimal parameter setup. The tool automatically derives s…
Vector RAG is emerging as the preferred retrieval method for production LLM applications, offering 5-10x faster retrieval times and lower costs compared to classic RAG. By using dense embeddings and approximate nearest n…
A developer who built a YouTube comment classifier using regex word-matching discovered that the tool misclassified 53% of test cases, including promoting social commentary and chatter into 'needs' due to accidental keyw…
Data center operators are increasingly deploying AI to improve cooling, power optimization, and predictive maintenance, with Digital Realty reporting a 34% portfolio growth against only a 3% increase in water usage, part…
A developer has released a tool to reconstruct and visualize distributed LLM training traces, enabling users to drag and drop compute kernels and collectives to create DDP, TP, FSDP, EP, and CP traces based on real torch…
Google Cloud has introduced native vLLM TPU support for embedding inference, targeting production long-context retrieval with Qwen3-Embedding-8B and Qwen3-VL-Embedding-8B. The engineering effort addresses TPU-specific ch…
Rostam Labs released Rembed, a pure-Go text embedding inference engine that runs BERT-style and decoder-derived embedding models without ONNX Runtime or cgo, achieving statistical parity with ONNX Runtime and faster int8…
VIDRAFT, a Korean AI safety startup, has published a diagnostic framework that detects causal leakage in hybrid sequence models, identifying issues in Nemotron-H-8B and Zamba2-1.2B while catching all 192 injected faults.…
In a 2019 essay, computer scientist Rich Sutton argues that 70 years of AI research show general methods that leverage computation are ultimately the most effective, citing examples from computer chess, Go, and speech re…
Tencent's WeChat Vision team has open-sourced WeMM-Embedding, a family of multimodal embedding models for representing and matching text, images, videos and other content types, with 2B, 4B and 9B versions. The 9B model …
Flint, a self-hosted neural network that forecasts stock prices and explains its reasoning, was released on Hacker News. The tool reads live market data, predicts the next hour for a basket of stocks, and generates a dai…
MLSentinel, a new free platform for monitoring deployed machine learning models, was announced on Hacker News by developer Narasimha440. The tool tracks model health and alerts users to underlying issues before they esca…
Developer Tejas released chunk-lite, an open-source JavaScript library for token-aware text chunking in RAG pipelines that respects sentence and paragraph boundaries to avoid cutting mid-thought. The library returns star…
MoganAI released MoganBert-TR, a 149M-parameter Turkish encoder foundation model trained from scratch on 237.3B tokens with a two-stage CLM-to-MLM curriculum, achieving 78.41 on TrGLUE, the best among compared Turkish Mo…
Researchers introduced ROMNet, a hybrid reduced order modeling and machine learning approach that uses a neural network to map a reduced order model (ROM) matrix to a nearby one with simpler dependence on wave speed, red…
Researchers propose PACE (Prioritized Adaptive Coverage of Evidence), a training-free framework that combines evidence frontloading with pressure-adaptive budgeting to relieve bottlenecks in Retrieval-Augmented Generatio…
Researchers from an unnamed institution propose ProViP, a training-free progressive visual token pruning framework for Vision-Language Models (VLMs), based on the finding that only a small fraction of attention heads are…
Researchers introduced CoRE, a weakly supervised coarse-to-fine framework that learns fine-grained prediction support from coarse video supervision, achieving strong temporal localization on the DoTA benchmark and compet…