Skip to content
K = 0.731
All news · Today3 items

#Data and training

All collected and published items, ordered by time. Picks are a selected subset.
10/9Fri
  1. Hugging Face Blog34

    How Ai2 Rebuilt Cluster Scheduling with GPU Time Budgets

    Ai2 replaced its original priority scheduler with GPU time budgets, hierarchical fair allocation, and time-slice contracts, turning debates over how much GPU time each research project should get from case-by-case operational decisions into a transparent budget allocation process. Its cluster manages thousands of NVIDIA H100, B200, and B300 GPUs, with scales ranging from 88 to 1,024 GPUs, and demand from roughly 150 internal researchers has long been 2-3 times the available compute.

  4. 量子位62

    ByteDance Seed finds why DeepSeek-V4 can be inconsistent: chunked KV-cache compression creates position sensitivity

    The ByteDance Seed team found that the DeepSeek-V4 series shows periodic fluctuations in long-context retrieval depending on where information appears, with a cycle of every 4 tokens. In the 128K needle-in-a-haystack test, moving the same information to a different position changed accuracy by as much as 40.2 percentage points.

  10. Latent Space36

    Periodic Labs' Liam Fedus and Ekin Dogus Cubuk on "synthesis superintelligence": from semiconductors to superconductors

    In an interview, Periodic Labs' Liam Fedus and Ekin Dogus Cubuk proposed a "synthesis superintelligence" approach: reinforcement learning in physical experiments, AI-driven materials characterization, density functional theory (DFT), and high-throughput labs. The aim is for models to learn the full process of doing science, rather than only studying published results.

10/8Thu
  2. 雷锋网38

    Tsinghua AIR's first PhD graduate Li Jianxiong founds BASAL Intelligence; 5Y Capital leads investment in action-native embodied models

    Li Jianxiong, the first PhD student at Tsinghua University's Institute for AI Industry Research (AIR), founded embodied AI company BASAL Intelligence. 5Y Capital led the round, with Ivy Capital, Linear Capital, and Huashan Capital participating. Tsinghua AIR associate professor Zhan Xianyuan is the chief scientist.

  5. 雷锋网32

    Dialogue with ONTIME Mobility COO Han Feng: The stronger the world model, the more real data becomes the true moat

    ONTIME Mobility COO Han Feng believes that as world model capabilities grow stronger, the importance of real data will not diminish: synthetic data is responsible for expanding known long-tail scenarios, while real data is responsible for discovering unknown problems and providing calibration and validation. ONTIME Mobility does not train large models; instead, it collects real data through routine mobility services, Robotaxi, and offline operations and maintenance networks, and delivers training materials to autonomous driving, world model, and embodied intelligence companies through a data closed loop.

  7. Microsoft Research72

    Microsoft Research open-sources Agent Lightning v1.0, a 3,500-line lightweight agent RL framework for training with real harnesses

    Microsoft Research open-sourced Agent Lightning v1.0, a lightweight agent reinforcement-learning framework of about 3,500 lines. Its Harnessed Agentic RL approach lets the same agent harness used in deployment participate directly in reinforcement learning, without reimplementing the agent inside a training framework.

    Why it matters: The source gives the framework's code size, Kubernetes support, and training recipe, helping readers assess the cost of bringing a real harness into RL.

10/7Wed
  3. Hugging Face Blog62

    Nemotron fine-tuning reaches gold-medal level at both IOI 2026 and IMO 2026

    The NVIDIA team started with Nemotron 3 and used supervised fine-tuning, reinforcement learning, and a generate-verify-improve reasoning process to build specialized models that reached gold-medal thresholds at both IOI 2026 and IMO 2026.

    Why it matters: The source gives a complete recipe for fine-tuning Nemotron 3 to gold-medal-level results at IOI and IMO, which readers can use to reproduce its data and reasoning workflow.

  4. Latent Space88

    OpenAI's internal model releases 722 math papers, claiming to solve 90 of 500 open problems

    OpenAI published mathematical results from an unreleased internal frontier model through a blog, GitHub repository, and posts. The release includes 722 manuscripts grouped into 372 related result families, drawn from an evaluation of about 4,000 research questions. Each result used an average of about three hours of ChatGPT Pro reasoning compute.

    Why it matters: OpenAI is producing mathematical results in bulk with an internal model, giving readers a view of the scale, compute cost, and outside criticism surrounding the work.

  6. TechCrunch · AI22

    Mirror Particle is building a 'world model' that simulates human behavior

    San Francisco startup Mirror Particle is building a foundation model from scratch to simulate human behavior and why it changes. It argues that using LLM role-play to predict group behavior is fundamentally ineffective. Its model combines customer data with current events, popular culture, and social media to track 'real behavior' rather than survey responses, and explains the motivations and constraints behind recommendations.

  7. AWS Machine Learning Blog22

    Management and governance best practices for Amazon SageMaker HyperPod

    Amazon SageMaker HyperPod can connect clusters to projects through SageMaker Unified Studio, letting team members launch training and fine-tuning workloads in workspaces while retaining governance controls at the infrastructure level. The article proposes a four-layer control model—organization, project, cluster, and workload—and explains how to design identity, capacity, and observability policies across those layers. Infrastructure teams retain control of cluster operations while ML teams use approved compute in project contexts.

10/6Tue
  3. The Decoder62

    Researchers extend LeCun's JEPA into JEPA-Anything, a world model spanning physical and biological domains

    Researchers at PhAI Labs, Stanford, Oxford, and Princeton proposed JEPA-Anything. It splits JEPA's predicted state into four orthogonal factors, assigns a prediction module to each, and recombines them. It outperformed standard JEPA with the same architecture and data on ten test tasks in physics, robotics, weather, and other areas.

  4. Hugging Face Blog26

    Falcon-Emirati-7B: teaching LLMs the Emirati dialect, culture, and feel for language

    Falcon-Emirati-7B is a specialized model for Emirati dialects built on Falcon-H1-Arabic. It aims to understand and generate Emirati Arabic like a native speaker. It uses the Falcon-H1 hybrid architecture, running Mamba alongside Transformer attention. The base model family comes in 3B, 7B, and 34B sizes, with a context window of up to 256K tokens.

  6. TechCrunch · AI72

    Reflection releases open-weight Beam model to compete with Chinese open models

    Reflection AI officially released its first frontier open-weight model, Beam, saying it performs comparably to leading Chinese open models on advanced reasoning benchmarks with 3–4 times less inference compute. Beam is a 501-billion-parameter, 23-billion-active-parameter text model, pretrained on 23.8 trillion tokens with a 1-million-token context window. The company says it matches Z.ai's GLM-5.2 and exceeds current leading Western open models.

  7. Simon Willison22

    Benchmark tests whether Qwen3.8 27B can add using English number words

    A benchmark tested the local Qwen3.8-27B-Q4_K_M.gguf on 5,070 cases without reasoning, asking it to add positive integers using only English words and return exact results. It achieved 23.57% numerical accuracy and 96.17% format compliance. Accuracy fell as the numbers grew, from 97.04% on one- to three-digit numbers to 6.44% on ten- to thirteen-digit numbers. In 169 paired cases with medium reasoning enabled, the model answered 167 correctly.

  9. Google AI · X62

    Gemma 4 and BOTANIC-1 rank plant mutations in minutes

    Google AI highlights work by AI research lab Living Models, which paired Gemma 4 with BOTANIC-1, a model trained on plant DNA from 320 species, to identify mutations that alter crop traits. In a melon yield gene test, the pair ranked the target mutation first out of 2,494 possibilities in under four minutes, versus years of physical breeding trials. A demo and paper are linked.

10/5Mon
  1. The Decoder62

    Aleph Alpha releases Kolibri, an open-weight 78B model

    Aleph Alpha released Kolibri, a bilingual German-English model with 78B total parameters using a mixture-of-experts architecture that activates about 3B parameters per token. The company says Kolibri sits on the Pareto frontier for quality and operating cost in both languages. It scores 71% on German benchmarks and decodes faster than comparable models including GPT OSS A5B, Qwen 3.6 A3B, and Gemma 4 A4B.

  2. MIT Technology Review · AI12

    How to bring predictive analytics into the age of agentic AI

    The focus of enterprise AI competition has shifted from whether predictive models can outperform statistical forecasts to how prediction systems can make autonomous decisions without straying from business intent. Deep learning and generative AI enable real-time training, allowing AI to keep evolving rather than wait for quarterly refreshes, and expand its data sources from structured numerical records to unstructured interaction data. Vishal Gupta, a partner at Everest Group, says the word 'analytics' is giving way to AI.

10/4Sun
10/3Sat