Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
RelativeDB, an open source query engine built around the Relational Transformer architecture from Stanford's Tabular and Relational (STAR) Project, can outperform XGBoost on some tasks with minimal labeled context and no…
A developer built a prompt-injection detector trained on 78,000+ attack samples, opting for a deterministic approach using regex and classical ML (TF-IDF with logistic regression) instead of an LLM-based detector. The sy…
A developer built a Genomic RAG pipeline using LlamaIndex and FAISS to interpret raw DNA data from services like 23andMe. The system retrieves clinical significance from the ClinVar database and generates an interactive …
A developer highlights Docling, an open-source document parser from IBM, that converts PDFs, DOCX, PPTX, images, and scanned documents into clean Markdown or JSON while preserving tables, reading order, and multi-column …
A full enumeration of every possible Cribbage deal, completed in under an hour using AI-generated code and Claude Opus 5.0, reveals that the pone's hand is worth more than the dealer's (8.125 vs. 7.990 hand EV) and that …
Experiential Labs launched wmo serve, an open-source tool that routes repetitive agent tasks to distilled smaller models, achieving frontier quality at 40%+ lower cost. The tool uses an OpenAI-compatible endpoint and con…
A curated index of over 30 free AI/ML textbooks has been published on GitHub by MarcosSete, with a weekly GitHub Action to validate all links and ensure they point to official author or publisher sites. The collection co…
In a head-to-head test, standard LoRA outperformed Weight-decomposed Low-Rank Adaptation (DoRA) by 2.1% in accuracy while taking 15% less training time per epoch, according to a developer's experiments on a consumer-grad…
Splitting a dataset into training and testing sets is essential to detect overfitting, and the standard Python implementation uses Scikit-Learn's train_test_split with an 80/20 split. Key pitfalls include setting a fixed…
A developer comparing C++ and Rust for AI/ML code with LLMs found that while Rust's Burn framework is competitive on GPU, it performs poorly on CPU compared to C++ with GGML. The experiments, published in the rust-vs-cpp…
Promograph uses Graph Neural Networks (GNNs) to model retail promotion waste by representing SKUs as nodes and co-purchase patterns as edges, enabling the propagation of discount impacts across the product graph. This gr…
A backend engineer with 8+ years of experience in APIs, distributed systems, and cloud infrastructure is seeking advice on transitioning to ML inference infrastructure roles, asking HN for the best hands-on projects or e…
A developer shipped a browser version of a four-in-a-row board game where the opponent is a neural network, running entirely as static files with no backend. The network, a policy+value model with under a million paramet…
Huawei Cloud ModelArts integrates data labeling, model training, and deployment, but users may encounter resource allocation errors when spinning up GPU-accelerated inference endpoints due to mismatches between requested…
Karpenter v1.14.0 can autoscale GPU inference on Amazon EKS by provisioning spot GPU nodes on demand, bin-packing a vLLM v0.25.1 model server, and deleting nodes when traffic drops, eliminating static GPU node groups. Th…
US Neural built Mycelium, a decoupled semantic tool registry that replaces LangChain's AgentExecutor for sub-10ms local tool routing. The system achieves 9.6ms cold discovery latency and 70.7% family-level intent matchin…
The UK's liver allocation algorithm appears to discriminate by age, with some younger patients seemingly unable to receive a transplant regardless of illness severity, according to researchers Arvind Narayanan, Angelina …
A technical analysis of backpropagation kernels reveals that performance walls in certain layers are caused by low occupancy, where the GPU fails to keep enough warps active on the Streaming Multiprocessor, leading to id…
A developer recounts realizing that building better AI models is not the answer; instead, great AI products are built by great systems. The developer emphasizes that the model is just one component, and the surrounding i…
Microsoft Research has introduced SkillOpt, a system that enables AI agents to self-evolve by optimizing their skill documents—markdown files that guide agent behavior—without modifying model weights. In evaluations, Ski…