Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Decagon deployed self-evolving AI agents that automatically improve their own prompts without human intervention, using a feedback loop that tests and promotes winning versions. The team found that the optimization step …
Shrijith Venkatramana, building git-lrc, explains the three phases of post-training that transform pretrained LLMs into helpful assistants. The process starts with supervised fine-tuning (SFT) on curated examples, follow…
A developer discovered metadata routing in scikit-learn, a feature that elegantly solves the problem of passing sample weights and groups through complex ML pipelines. The feature, enabled via set_config(enable_metadata_…
A developer built a 'Dream Protocol' for their OpenClaw AI agent that runs nightly memory consolidation, scoring 700+ recall entries against three gates and promoting only the most significant signals to long-term memory…
Researchers from Peking University developed ECG-R1, a multimodal vision-language model trained with reinforcement learning to differentiate ischemic from nonischemic T-wave inversion on ECGs. The model, accepted at ICML…
A cautious optimist argues that AI safety, interpretability, bias, and alignment are as critical as raw capability, warning that the same technology enabling vaccine research could also design pathogens. The author empha…
A new approach to memory systems for long-running AI agents transitions from episodic to procedural memory, enabling agents to accumulate knowledge and skills over days. This advancement addresses the limitations of flat…
Researchers used deep learning to screen 19.3 million fragments from 2,897 prion-related proteins, identifying 1,179 candidate antimicrobial peptides called prionins. Of 75 synthesized prionins, 59 inhibited bacterial pa…
The art collective 0rphan Drift, founded in London in 1994, has evolved from a hive-mind approach to embrace distributed, many-minded consciousness through collaborations with machine learning engineers, marine biologist…
A developer built a self-optimizing Python trading bot using reinforcement learning and the Binance API. The bot uses a custom Gym environment with a PPO agent from Stable-Baselines3 to learn trading strategies from mark…
A developer created a DKMS quirk to fix missing internal speakers on the ASUS Zenbook S16 UM5606GA (AMD Strix Point) under Linux. The BIOS incorrectly sets the audio chip to a legacy mic-only mode, preventing the SoundWi…
Microsoft, Hugging Face, Meta's PyTorch team, NVIDIA, and others launched OpenEnv, an open protocol for agent learning environments that standardizes how agents practice and improve. The protocol aims to make the learnin…
A group of Berkeley CS students built autonomous AI agents that replace chat interfaces with a system using task planning, sandboxed Python execution, SQLite storage, and email aggregation. The agents decompose high-leve…
A Hackster.io project by Marc Pous demonstrates real-time room classification using a DFRobot LiDAR Developer Kit connected to an Arduino UNO Q, running an Edge Impulse ML model for on-device inference. The 2D dToF LiDAR…
The foundations of today's AI boom, including the first Transformer variant, unsupervised pre-training, neural network distillation, and deep residual learning, were all published in 1991 by Jürgen Schmidhuber's lab at t…
Jep embeds CPython in Java via JNI, enabling Java applications to run Python code natively. The project supports Python ≥3.10 and Java ≥1.8, offering features like multiple sub-interpreters and NumPy array support.
Neo launched an MCP server that lets users offload AI tasks from Claude, reducing costs by 62% and speeding up runtime by 37% in benchmarks. The tool integrates with Claude Code and other MCP clients, choosing optimized …
A developer explores five key lessons from the architecture of intelligence, focusing on how informed search algorithms use heuristics to make efficient decisions. The post explains concepts like greedy best-first search…
Mvsep launched iOS and Android apps with auto-update and review features, introduced new instrument separation models including SATB Choir, added VibeVoice for voice cloning and TTS, and integrated Transkun for piano tra…
IEEE Quantum Week 2026 will host a workshop titled 'AI for Circuit Synthesis, Optimization, and Discovery' in Toronto, Canada, from September 13-18, 2026. The workshop focuses on applying artificial intelligence, includi…