2026-08-21 科技日报

扫描 565 篇候选内容 · 覆盖 650 个信息源 · 纳入 552 篇

📌 今日要点

• Measuring benchmark optimization in speech recognition
• Towards On-Board Implementation of ML-Based Helicopter Weight Estimator
• Triangular Fuzzy Rescaling Distance
• Holtercare-Bench: A Multimodal Benchmark for Evaluating Long-Term Dynamic ECG Analysis
• Quantum Kernel Estimation for the Discovery of Early Lung Cancer Detection
• Improved Confidence Estimates for Black-Box Large Language Models

📝 科技简讯

• Measuring benchmark optimization in speech recognition — Hugging Face Blog
• Towards On-Board Implementation of ML-Based Helicopter Weight Estimator — ArXiv ML (cs.LG)
• Triangular Fuzzy Rescaling Distance — ArXiv ML (cs.LG)
• Holtercare-Bench: A Multimodal Benchmark for Evaluating Long-Term Dynamic ECG Analysis — ArXiv ML (cs.LG)
• Quantum Kernel Estimation for the Discovery of Early Lung Cancer Detection — ArXiv ML (cs.LG)
• Improved Confidence Estimates for Black-Box Large Language Models — ArXiv ML (cs.LG)
• Mechanistic Tomography: Designed Measurement for Control-Oriented Interpretability — ArXiv ML (cs.LG)
• Uncovering the Limits of Proof Sharing for Neural Networks — ArXiv ML (cs.LG)
• Run Muse Glimmer for Local Vibe Coding with llama.cpp, DFlash, and Pi — KDnuggets
• 5 Real-World Use Cases for AI Agents Transforming Industries — KDnuggets
• From Atari to EVE Online: Building on 15 Years of AI Research in Games — Google DeepMind
• How AI coding tools are contributing to the popularity of JavaScript — AI News (Jack Clark)
• Agentic Data Operations Platform (ADOP): Data engineering into hours — AWS Machine Learning
• Govern AI agent tool access with Amazon Bedrock AgentCore Gateway — AWS Machine Learning
• Reduce RAG costs on Amazon Bedrock with query-aware compression — AWS Machine Learning
• Accelerating aircraft IFEC diagnostics with agentic AI on AWS — AWS Machine Learning
• Anthropic’s Opus 4.6 is a smut-machine — TechCrunch AI
• Nvidia partners with data center developer Cloverleaf — TechCrunch AI
• Nvidia just showed that the harness, not the AI model, is now the real hero — TechCrunch AI
• Starcloud raises $250 million for orbital data centers as launch options dry up — TechCrunch AI
• The DOJ is investigating a16z. What does this mean for venture capital? — TechCrunch AI
• AI data startup Micro1 reaches $500M gross run rate amid AI training boom — TechCrunch AI
• OpenAI is gaining on Anthropic with business users, new data indicates — TechCrunch AI
• A Virtual Member of a Community of Practice for the Society of Petroleum Engineers: From Prototype to Deployment — ArXiv CL (cs.CL)
• Transformer Models for Text Summarization: A Comparative Study of BART, BERT, and RoBERTa — ArXiv CL (cs.CL)
• Automatic bioinformatic software named entity recognition from literature — ArXiv CL (cs.CL)
• Asymmetric Attention Heads: Structured Head-Wise Context Allocation for Transformer Attention — ArXiv CL (cs.CL)
• Hallucination as a Feature, not a Defect: Evaluating a multi-agent architecture to transform speculative language-model outputs into testable scientific hypotheses — ArXiv CL (cs.CL)
• 刚刚，多模态版 DeepSeek「长眼」了！1000 张图只要 1 块钱 — 新智元
• Cursor 祭出最强「自动打工」模式！一个/goal 完爆全场 — 新智元
• 机器人 GPT-3 时刻震动硅谷！0 行代码，看一遍秒会，黄仁勋李飞飞参投 — 新智元
• 失业小哥 7000 行代码，复刻了 AMD 重金收购的 AI 芯片！ — 新智元
• AI 向癌症发起猛攻，4000 种抗癌药虚拟试杀！谷歌 Gemma 下载破 10 亿 — 新智元
• APPROVE: Visual End-User-in-the-Loop Robot Programming with LLMs — ArXiv RO (cs.RO)
• The Missing Touch: Spatially Distributed Tactile Feedback Brings Teleoperation Closer to Human Dexterity — ArXiv RO (cs.RO)
• Learning the Right Abstraction: Neural Reduced Dynamics for Complex Robot Control — ArXiv RO (cs.RO)
• SCAPE: Scenario-Conditioned Simulation-Augmented Policy Evaluation — ArXiv RO (cs.RO)
• Hybrid Feedback Sampling for Sample-Efficient Model Predictive Control — ArXiv RO (cs.RO)
• 最大运力自动驾驶轻卡落地，来自无人车巨头 — 量子位
• 明略科技携手海康机器人亮相世界机器人大会，以“Agent+ 具身”联合进入商业机器人场景 — 量子位

📊 数据概览