Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
RLE-Bench has been introduced as a qualifying exam that evaluates coding agents as robot learning engineers, according to the benchmark's announcement. Existing robotics benchmarks primarily focus on the performance of i…
Amazon Web Services detailed Amazon SageMaker AI multi-turn reinforcement learning (MTRL), a fine-tuning capability that trains LLM search agents across full multi-turn trajectories rather than single responses, using mu…
A developer built a 38-case security vulnerability detection benchmark on Kaggle with deliberate false-positive traps to test whether LLMs reason about data flow or merely pattern-match, evaluating 9 models across 5 prov…
A developer tested the calibration of Jev, TypeSafe's new "System One" classifier model that bolts a classifier onto a pretrained transformer to return typed probability distributions over multiple-choice answers. Using …
Google Research announced a next-generation Federated Learning system that uses Trusted Execution Environments (TEEs) to provide fully verifiable and auditable data anonymization guarantees, with Gboard already adopting …
A 28 Sep 2026 arXiv paper introduces ProVer, a framework that targets potentially pivotal decisions for fine-grained credit assignment in agentic reinforcement learning, addressing what the authors call GRPO's uniform as…
A new arXiv paper submitted on 1 Oct 2026 introduces ScholarCatalyst, a retrieval benchmark built from 184 lead authors of 207 recent computer science papers who labeled which prior works advanced their projects. The aut…
TypeSafe AI released Jev, the first model in its System One class, in early access with a waitlist from September 15, 2026, returning typed probabilistic decisions instead of generated text. Jev offers three decision pri…
Uber rebuilt major parts of the Uber Eats search pipeline and reports a 50% reduction in end-to-end search latency, with early product-based search testing producing more than a 50% reduction in p99 latency. The changes …
A test of 511 AI-designed antibodies found that AI can improve promising drug candidates, but laboratory experiments remain essential for discovering safe, developable antibodies, according to Forbes contributor William …
TypeSafe AI's Jev and "System One" decision models do not outperform LLM-as-a-judge or purpose-built traditional classifiers, according to an analysis of the approach. Jev returns fixed, schema-safe decisions such as a 0…
Microsoft's 2026 Digital Defense Report, covering July 2025 to June 2026, warns that threat actors are using AI to find bugs, build malware and run intrusions faster than defenders can respond, creating a near-term perio…
TypeSafe AI launched Jev, which it calls the first "System One model," in early access on September 15, 2026, alongside a $40 million seed round led by DCVC. Jev takes text plus typed questions — choice, score, or yes/no…
A benchmark on a 120,000-listing Indian AgriTech mandi dataset compared legacy pre-filtering, legacy post-filtering, and Qdrant's single-stage in-index filtering across four filter-selectivity levels, 60 queries per scen…
Epoch AI's Benchmarking Hub held about 6,800 results covering roughly 1,200 model versions of 710 models as of 1 October 2026, spanning more than 80 benchmarks from GPQA Diamond and SWE-bench Verified to FrontierMath and…
Hopsworks is hosting two self-hosted decision models, qwen and kumo, that play the procedural game Hops Run live at game.hopsworks.ai, with each move sent to a Hopsworks model deployment that returns a probability per mo…
Databricks published a reference architecture for real-time e-commerce product recommendations built on its Data Intelligence Platform, based on an implementation for an Asian fashion retailer serving over 1 million mont…
Google DeepMind's Gemini-powered AlphaProof Nexus multi-agent framework resolved 9 of 353 previously unsolved Erdős problems, two of which had resisted mathematicians for 56 years, and proved 44 of 492 conjectures in the…
Public AI benchmarks should be used as capability screens to build a shortlist rather than as deployment decisions, according to a practical evaluation framework published by deepsense.ai. The framework cites OpenAI's Ju…
A decade of official labor statistics from the UK Office for National Statistics and the US Bureau of Labor Statistics shows no net aggregate job destruction nearly four years into commercial large language model deploym…