Machine Learning News

Researchers have developed EURO-5K, a new dataset of sentence-level reporting obligations from 136 EU legislative acts, to improve automated extraction of regulatory requirements. Benchmarking discriminative BERT-style m…

Researchers have developed WRIT (Write-Read Intensive Trajectory Synthesis), a pipeline that generates multi-turn training data for user-facing agents by creating tasks with both high write-decision counts and heavy read…

Linear probes applied to Qwen3-14B hidden states achieve 100% accuracy in distinguishing deductive, inductive, and abductive reasoning tasks, but this separation is entirely driven by format confounds such as source iden…

A systematic human audit of the NL-to-FOL benchmarks FOLIO and MALLS found that approximately 39% and 36% of entries, respectively, contain incorrect FOL formalizations, with additional errors in ambiguous natural langua…

Researchers found that transformer-based language models perform better when deeper attention layers learn context-free value vectors that preserve original token information, rather than relying on the residual stream f…

A developer built MangoGuard, an edge AI system that detects mango diseases directly on an Arduino Nano microcontroller without requiring cloud connectivity or internet access. The quantized MobileNetV1 model classifies …

Deep neural networks have enabled generative modeling of biomolecules, with models like AlphaFold3 and Chai-2 now capable of predicting and designing drug-like molecules and antibodies. However, researchers at Ligo found…

A team of University of Toronto researchers has demonstrated the first AI-powered computer worm that can adapt its strategy as it spreads from one device to another, potentially seizing control of entire networks. Built …

TradingAgents, an open-source multi-agent LLM trading framework with 82,356 GitHub stars and 15,978 forks, includes a `SqliteSaver` checkpointer that writes a separate SQLite database per ticker to save state after every…

A solo developer built a fully offline conversational AI Hydration Coach using React Native and Google's Gemma 4 E2B model, running entirely on-device with no cloud servers or API calls. The project, called Water Tracker…

A developer found that retrieval systems retrieving sensitive memories mislabeled as ordinary context produce false-certainty errors, where agents confidently act on sensitive content without verification. Testing across…

Software engineers are shifting their focus from high-level abstractions to foundational computing skills, driven by the rapid advancement of agentic coding tools like Claude Code and Codex that outperform most programme…

A developer built Sentinel, a real-time fraud detection system that processes 71,000 requests per second with p95 latency under 6 milliseconds. The system handles 7.8 million requests with zero errors by serving XGBoost …

A free, fully automated pipeline now pulls daily health data from Garmin Connect into a SQLite database, computes percentiles and trends across a user's entire history, and delivers a short coached brief via ntfy, Telegr…

A Hacker News user asked for book or article recommendations covering the modern history of machine learning and large language models, specifically how the field reached its current inflection point. The request seeks r…

Microsoft and MAI announced new frontier models and Frontier Tuning capabilities, enabling companies to customize AI against their unique data and processes rather than simply consuming generic intelligence. The initiati…

Liquid AI's LFM2 model can be fine-tuned using QLoRA and DPO in a complete open-source workflow on Google Colab. The process loads the base LFM2 checkpoint with 4-bit quantization, trains a lightweight LoRA adapter using…

A developer is building an autonomous racing agent for the TORCS (The Open Racing Car Simulator) simulator, aiming to complete a clean lap around the Corkscrew track without crashing. The project will use a local LLM, IB…

A developer built a "Medicine Box Expert" pipeline that uses YOLOv8 for object detection and OpenAI CLIP for multimodal understanding to turn photos of medicine packaging into a searchable digital database. The system em…

A RAG system's PII redaction sentinel tokens, such as [NAME] and [EMAIL], inadvertently collapsed vector index retrieval by creating strong co-occurrence signals that dominated semantic similarity, causing queries to ret…