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.
10/3Sat
  1. AI at Meta · X66

    Meta publishes six papers on mathematicians collaborating with Muse Spark

    Meta said that after AI models achieved gold-medal-level performance in five math, physics, and chemistry competitions, it had mathematicians work with Muse Spark 1.1 and 1.2 in Thinking Mode through the meta.ai chat interface to solve open problems with no existing solution path and no custom research scaffolding.

    Why it matters: Meta made six papers from mathematicians' collaboration with Muse Spark public and described the division of work between people and AI and the review process.

  2. Hugging Face Blog62

    Ai2 open-sources AstaBrief 8B, a scientific report-generation model

    Ai2 open-sourced AstaBrief 8B, a model that turns research questions and retrieved literature excerpts into reports with citations. It was trained from Qwen3-8B using SFT and DPO and is now available as Asta's Fast mode.

    Why it matters: The source describes AstaBrief's training-data construction and citation-filtering details, which can be adapted to deploy other long-form generation models.

10/2Fri
  1. Hugging Face · X74

    Hugging Face publishes multi-harness RL guide: the same model scores 62% versus 33% with different harnesses

    The Hugging Face team published a multi-harness RL guide, showing that the same model with identical weights can score 62% with one agent harness and 33% with another.

    Why it matters: Scores for the same model differ by nearly half across agent harnesses. This open guide gives a multi-harness RL training path without changing the harness, along with measured comparisons.

  2. Hugging Face Blog42

    ServiceNow CoreAI launches AutoSynthData to generate training data for enterprise agents

    ServiceNow CoreAI built AutoSynthData to find capability gaps from a target model's failures and a stronger teacher's successful trajectories, then generate and validate new tasks for post-training. As the model improves, the curriculum shifts to tasks that remain difficult. The process was demonstrated on the EnterpriseOps Gym dataset. Generated tasks must meet feasibility, realism, and difficulty requirements; validators must provide consistency, reliability, and completeness.

10/1Thu
  1. Hugging Face Blog62

    AI2 releases Olmo-core 3, open and scalable training infrastructure for large MoE models

    AI2 released Olmo-core 3, a major upgrade to its large language model development framework. It redesigns the open MoE training system to scale MoE training to trillions of parameters while maintaining compute efficiency.

    Why it matters: The source provides throughput and memory comparisons for the MoE training stack, helping readers assess changes in the cost of large-scale MoE training.

  3. NVIDIA Blog22

    How NVIDIA AI factories maximize ROI through high output, long life, and broad applicability

    NVIDIA says AI factory return on investment depends on three factors: output capacity, useful life, and demand. Its platform uses full-stack co-design to maximize throughput per megawatt. SemiAnalysis AgentX data show Vera Rubin NVL72 delivers more than 30 times the throughput per megawatt of GB300 NVL72 and, on DeepSeek V4 Pro, can reduce cost per million tokens by up to 45 times.

  4. Perplexity · X42

    We built a new way to train contextual embedding models, which encode each chunk of a document with the whole document in view. pplx-embed-v2-context-9b-preview sets a new state of the art on ConTEB and @turbopuffer's new, privately held context-bench. https://www.perplexity.ai/hub/blog/contextual-embedding-beyond-the-gold-passage

    We built a new way to train contextual embedding models, which encode each chunk of a document with the whole document in view. pplx-embed-v2-context-9b-preview sets a new state of the art on ConTEB and @turbopuffer's new, privately held context-bench. https://www.perplexity.ai/hub/blog/contextual-embedding-beyond-the-gold-passage

9/30Wed
  1. Google DeepMind62

    Google DeepMind launches SynthID Bio to embed watermarks in AI-generated proteins

    Google DeepMind launched SynthID Bio, bringing watermarking to synthetic biology. It embeds imperceptible signatures directly into biological code so that generated content can be verified both in digital models and in synthesized physical proteins, while preserving their biological function.

    Why it matters: SynthID watermarking has expanded from images to synthetic biology, showing how AI-generated proteins can be marked and verified.

  2. NVIDIA Blog62

    NVIDIA and CoreWeave put Vera Rubin NVL72 into production; Cognition is first customer

    CoreWeave announced that NVIDIA Vera Rubin NVL72 systems and Spectrum-X 102.4T Ethernet are available on CoreWeave Cloud, making it one of the first cloud providers to deliver the platform. Cognition is the first customer to run Vera Rubin in production.

    Why it matters: CoreWeave launched Vera Rubin NVL72 and published measured Devin throughput, offering a view of compute costs for agentic coding workloads.

  3. 雷锋网22

    Gotion High-Tech unveils Xingchen low-altitude flight battery, China-Europe-Africa logistics solution, and zero-carbon energy platform

    At a special launch event at the 2026 World Manufacturing Convention, Gotion High-Tech introduced three offerings: the Xingchen high-power battery for low-altitude flight, an integrated new-energy logistics solution for China, Europe, and Africa, and the GotionGuard-ZERO zero-carbon digital energy cloud platform.

  4. 雷锋网40

    Why do Kimi and DeepSeek appear in the same model-architecture diagram?

    Of the 45 layers in GLM-5.3-Flash, 34 use Kimi's KDA linear-attention approach and 11 use DeepSeek's KPool-DSA sparse attention. Linear and Sparse attention appear together in one model for the first time. The combination reflects a shift away from finding a single best mechanism toward assigning different roles: Linear maintains long histories at low cost, while Sparse retrieves specific information as needed.

  6. 雷锋网28

    From 3D vision to world models: why spatial intelligence is a common thread in AI's move into the physical world | ECCV 2026 recap

    ECCV 2026 received 10,473 submissions and accepted 2,834. The 3D topic fell from a leading area to third place, but questions around 3D representation, geometry, and physical consistency are spreading into video generation, embodied AI, and world models. Of 93 workshops, about 10 were directly related to 3D, 11 to spatial intelligence, nine to physical AI, and five to world models. Best-paper and honorable-mention awards also focused on 3D surface representation, object reconstruction, and geometry perception.

  9. 雷锋网66

    Zibian Robotics releases TwinDEX: body-free data nearly matches real-robot teleoperation data

    Zibian Robotics released TwinDEX, combining a wearable body-free data-collection system with a compatible robot end effector and a data-processing and policy-training pipeline. Evaluations found that a few hundred body-free examples alone let a robot perform a 24-subtask chemistry experiment in a single continuous run. Body-free data can nearly 100% replace real-robot teleoperation data in training effectiveness, and the task's data collection was 5.3 times as efficient as real-robot teleoperation.

  10. 雷锋网24

    Shutu Technology raises nearly RMB 100 million in Pre-A funding to bring SynaTac tactile product line to market

    Shutu Technology raised nearly RMB 100 million in a Pre-A round co-led by CDH Baifu and Kunpeng Fund, with Guangzhou Industrial Investment and Yihe Capital participating. Existing investor Trend Investment also made an oversubscribed follow-on investment. This is the company's third funding round in one month. The money will support R&D and validation for tactile-data representation training, iterate the SynaTac product line, and accelerate applications in robot contact-policy training and on-device deployment.

9/29Tue
  1. Hugging Face Blog62

    NVIDIA releases Kumo Tabular, a foundation model for tabular prediction

    NVIDIA has released Kumo Tabular, an open-source tabular foundation model that predicts labels for new rows in a single forward pass from labeled table rows, supporting classification and regression without training, tuning, or feature engineering. The model comes in three sizes from 28M to 215M parameters, was pretrained entirely on synthetic tables, is commercially usable under the OpenMDW-1.1 license, and ranks first on four benchmarks: TabArena, BeyondArena, TALENT, and ScoringBench.

    Why it matters: The source reports model sizes, training-data provenance, and rankings on four benchmarks, helping readers assess the accuracy and efficiency tradeoffs of tabular foundation models.

  2. Microsoft Research62

    Microsoft Research releases Quine, an AI research system for biological complexity

    Microsoft Research has released Quine, an AI research system designed for biological complexity. It combines a biological world model with an interactive harness connecting models, scientific tools, literature, and experimentalists. In work with the Broad Institute on pancreatic ductal adenocarcinoma, it predicted and prioritized thousands of compounds; the top-ranked compound produced the largest classical-to-ground-state transition in wet-lab experiments. The process from narrowing the search space to selecting a candidate for validation took one weekend.

    Why it matters: Microsoft Research describes Quine's multimodal biological world model and experimental loop, including wet-lab validation in pancreatic cancer compound screening.

  3. 量子位36

    Knowin Intelligence raises hundreds of millions of yuan in an Angel+++ round as GLOW model and KNOWIN-X1 move toward mass production

    Consumer embodied-AI company Knowin Intelligence has raised hundreds of millions of yuan in an Angel+++ round led by a JD.com-affiliated fund, with follow-on investment from Zhengxingu Capital, Nanshan Strategic Emerging Industries Investment, and Walden International. Total funding since founding exceeds RMB 1 billion. Its GLOW generative embodied model achieved a 62.2% average success rate across 18 RoboDojo simulation tasks, 40 percentage points above the GPT-6 (Robocurve) baseline; it ranked first among the listed single-policy models on LIBERO-Pro with 86.7%.

  4. 量子位82

    AMD to acquire Fei-Fei Li's World Labs for $8.2 billion in an all-stock deal; Li to become chief scientist

    AMD announced an all-stock acquisition of Fei-Fei Li's World Labs for $8.2 billion. The deal is expected to close by the end of 2026, subject to regulatory approval. After closing, Li will join AMD as executive vice president and chief scientist, reporting directly to Lisa Su; co-founders Justin Johnson and Ben Mildenhall will continue leading the team as it joins AMD.

    Why it matters: The full deal details and the companies' year-and-a-half collaboration are laid out, helping readers understand why a chipmaker would acquire a spatial-intelligence team.

  5. Ars Technica · AI60

    Experts raise concerns about Nvidia AI chip sales to China and its influence on Trump

    China's Ministry of Industry and Information Technology is considering easing restrictions by asking Alibaba and ByteDance to submit plans to buy Nvidia RTX Pro 5500 chips; ByteDance is reportedly planning an order of one million. The report says Jensen Huang has become Trump's most influential adviser on technology issues, and U.S. Treasury Secretary Scott Bessent said Trump and Huang are fully aligned. Some experts worry Nvidia's profit motive is influencing U.S. AI safety policy. Nvidia says these are outdated gaming chips released nearly five years ago and that easing restrictions would only benefit U.S. companies burdened by the limits.

9/28Mon
  1. 量子位40

    Huawei unveils Source-Grid-Load-Storage AIDC 1.0, using tokens per unit of energy to rethink AI data-center efficiency

    At a concurrent AIDC infrastructure summit during HUAWEI CONNECT 2026, Huawei presented its Source-Grid-Load-Storage AIDC 1.0 solution and updates to power supply, energy storage, and liquid cooling. The Taishan UPS reaches 1,280 kVA per cabinet with up to 98% double-conversion efficiency; one Hengshan DC UPS hardware platform supports 270V, 400V, and 800V systems.

9/27Sun
  1. 量子位40

    Fermi Universe launches FermiQLLM 1.0, a fully quantum-enhanced foundation model, with a RMB 1 billion valuation

    Tsinghua-affiliated startup Fermi Universe has launched FermiQLLM 1.0, which it calls the world's first fully quantum-enhanced foundation model. Adapted from the open-source Qwen base model, it applies quantum enhancement across data representation, model architecture, training, reinforcement, and evaluation. It can be trained and deployed on existing GPUs without fault-tolerant quantum computers.

9/26Sat
  1. Jeff Dean · X60

    Jeff Dean Highlights Waymo Safety Data Showing 20X Better Serious-Injury Crash Rate

    Jeff Dean shared Waymo's latest safety data, noting the rate of crashes with serious injury compared with human drivers has improved to 20X better at 270M miles, up from 13X better at 170M miles in March 2026 and 10X better earlier. Waymo's quoted post says the Waymo Driver reduced injury-causing crashes by 82% and serious injury crashes by 95% across 5 territories, with 841 fewer injury-causing crashes.

  2. Ars Technica · AI48

    Tesla workers reluctant to collect training data for Optimus humanoid robots that could replace them

    Tesla factory workers were asked to wear special motion-recording suits to collect imitation-learning data for Optimus humanoid robots. Some resisted because they “know the robots are ultimately designed to replace them.” Tesla later shifted data collection to dedicated teams and established “training centers.”

  3. Ars Technica · AI50

    CESifo study finds no evidence AI has significantly hurt employment for recent graduates

    CESifo Munich economists Robert Fairlie and Jane Wu published a working paper saying there is “no evidence of any significant, widespread displacement or reduction in hiring of recent college graduates.” The finding conflicts with a Stanford study last month that found entry-level employment in “AI-exposed” occupations lagging substantially behind other fields. The researchers suggest employers may respond to AI taking over standardized entry-level tasks by hiring fewer people before resorting to layoffs.

  4. AWS Machine Learning Blog36

    Scaling MoE reinforcement learning on Amazon EKS with EFA and DeepEP boosts throughput by 40%

    AWS presents an architecture for scaling MoE reinforcement-learning training on Amazon EKS with EFA and DeepEP, improving throughput by 40%. It addresses resource contention between rollout generation and policy training in large-scale RL workloads such as RLHF and GRPO, high-throughput communication across hundreds of accelerators, and the cross-node all-to-all overhead introduced by Expert Parallelism.

  6. AWS Machine Learning Blog34

    NarrateAI delivers production-grade LLM quality assurance on Amazon Bedrock

    NarrateAI uses five techniques on Amazon Bedrock to achieve about 99% numerical accuracy while supporting real-time streaming responses: adaptive pipeline orchestration, cross-account multi-model failover, real-time streaming evaluation, a composite evaluation framework, and data-accuracy validation. Each addresses failure modes such as hallucinated metrics, API throttling, validation delays, and subjective language. Built on Amazon Bedrock AgentCore, the solution serves more than 4,000 AWS executives.

9/25Fri
9/24Thu
  1. Latent Space36

    Foundries vs. Navigators: how AI can lower the cost of scientific research

    Language models have greatly sped up coding and data analysis, but the bottleneck in scientific research remains physical experiments that can take days or weeks to validate: “thinking got cheaper; doing didn't.” The biotech industry is splitting into two approaches. Foundries use next-generation sequencing, multiplex assays, and lab automation to produce experimental data an order of magnitude faster; Navigators embed AI in day-to-day company workflows to accelerate decisions.

  2. NVIDIA Blog60

    NVIDIA, Google DeepMind, and partners open predicted protein-complex structures for more than 2,800 viruses

    NVIDIA, Google DeepMind, EMBL-EBI, and other organizations have made predicted 3D structures of protein complexes from more than 2,800 viruses freely available to any scientist through the AlphaFold Database.

    Why it matters: NVIDIA, Google DeepMind, and collaborators have opened predicted protein-complex structures for more than 2,800 viruses, showing a concrete approach to open science for pandemic preparedness.