← 返回首页

📡 新闻雷达

AI · 安全 · 经济 · 科技 · 每 30 分钟更新

📊 共 120 条🌐 10 个信源🕐 29 分钟前 更新🛟 快照已显示 · 正在后台更新
今日重点📋 十条快看
🕐 29 分钟前 生成
  1. 1🤖为什么中国的 DeepSeek、Qwen 和 Moonshot 令美国人工智能竞争对手担忧中国AI初创公司在大型语言模型领域取得显著进展，引发美国竞争对手的担忧。Bloomberg Tech
  2. 2🔬立场：AI锁定正在进行，我们必须做好准备文章讨论了AI领域可能出现的“锁定”现象，并呼吁为应对做好准备。arXiv cs.AI
  3. 3🤖大型语言模型显示医学推理中的元认知敏感性研究发现大型语言模型在医学推理任务中表现出元认知敏感性。arXiv cs.AI
  4. 4🛠️OGX：开源、供应商中立的生成式人工智能应用服务器介绍了OGX，一个开源、供应商中立的生成式AI应用服务器。arXiv cs.AI
  5. 5🔬立场：人工智能治理需要类似ISO的互操作性协议，而不仅仅是法律作者提出AI治理应侧重于互操作性协议，而非仅依赖法律法规。arXiv cs.AI
  6. 6🤖技能：用于逻辑优化的自我修正知识引导迭代大语言模型代理提出了一种名为“技能”的LLM代理，用于逻辑优化和自我修正。arXiv cs.AI
  7. 7🤖面向安全的LLM代理：规范、验证和执行的调查探讨了如何构建安全的大型语言模型代理，包括规范、验证和执行。arXiv cs.AI
  8. 8🔬立场：AI锁定正在进行，我们必须做好准备文章讨论了AI领域可能出现的“锁定”现象，并呼吁为应对做好准备。arXiv cs.AI
  9. 9🤖幻觉雪球：将错误传播建模为多代理LLM管道中的状态转换研究将LLM管道中的错误传播建模为状态转换，形象地称为“幻觉雪球”。arXiv cs.AI
  10. 10🔬全球人工智能高风险用例中的公平和道德法规：比较审查对全球范围内AI高风险用例的公平和道德法规进行比较审查。arXiv cs.AI
持续更新

雷达新报

点击卡片，直达原始信源 ↗
📈 市场📈 Bloomberg Markets🕐 41 分钟前
Situation Awareness 将太阳诱电股份出售给 Citadel 等公司
Situational Awareness sold shares in Japan’s Taiyo Yuden Co. to Citadel and other investors when it offloaded part of its stake this month, it said in a filing.
🌐 互联网🏠 IT之家🕐 43 分钟前
消息称国际版小米 18 Ultra 取消发布，OPPO / vivo 收缩海外 Ultra 手机规划
IT之家 8 月 18 日消息，消息源卡蒂基 · 辛格（Kartikey Singh）昨日（8 月 17 日）在 X 平台发布推文，指出小米公司可能不会在国际市场上推出小米 18 Ultra 机型。IT之家翻译推文内容如下：小米已取消 18 Ultra 项目，OPPO 和 vivo 品牌也决定不在国际市场上推出下一代 Ultra 机型。在国际市场上，三星 Ultra 可能是全球唯一仍采用 Ultr
🌐 互联网🏠 IT之家🕐 47 分钟前
中汽协：7 月汽车出口 104.3 万辆，同比增长 81.3%
IT之家 8 月 18 日消息，据中国汽车工业协会分析，2026 年 7 月，出口继续保持高速增长，月度出口量连续两个月超过 100 万辆，新能源汽车出口占比连续两个月超过 50%。IT之家注意到，2026 年 7 月，汽车出口 104.3 万辆，环比增长 0.6%，同比增长 81.3%。2026 年 1-7 月，汽车出口 614 万辆，同比增长 66.8%。2026 年 7 月，乘用车出口 92
🌐 互联网🏠 IT之家🕐 48 分钟前
中科宇航力巡一号“太空摆渡车”完成重要试验，明年一季度首飞
IT之家 8 月 18 日消息，据央视新闻今日报道，由中科宇航自主研制的力巡一号上面级完成一系列动力系统试验，计划明年一季度首飞。上面级就像是一辆“太空摆渡车”，一般安装在火箭和它所运载的航天器之间。火箭把它们送入太空后，就要靠上面级来完成后续的卫星入轨的工作。2026 年 3 月，上面级 20kN 发动机完成边工况、变混合比及变冷却百分比拉偏边界试车考核，累计点火 18 次、累计点火时间 481
📦 软件🏠 IT之家🕐 50 分钟前
赛昉科技正式推出服务器级 RISC-V HPC CPU IP 昉 · 天枢-100
IT之家 8 月 18 日消息，赛昉科技 (StarFive) 今日正式宣布推出 RISC-V HPC CPU 内核设计昉 · 天枢-100 (Dubhe-100)。该 IP 已被希奥端计算 (LECARC) 选为智能边缘服务器 CPU 的基础。赛昉表示，昉 · 天枢-100 专为 HPC（高性能计算）场景设计，旨在满足 AI 智能体等新兴应用对算力的持续攀升需求，其也是国内率先达到可交付状态的服
📈 市场📺 CNBC🕐 52 分钟前
美伊停火到期可能引发长期冲突，霍尔木兹海峡发生船只袭击
U.S., Iran appeared to be digging in for a protracted conflict, hardening their positions and rejecting renewed negotiations after a temporary truce lapsed.
🤖 LLM📰 Hacker News AI🕐 53 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
Article URL: https://medium.com/freedomofthought/what-if-i-told-you-humans-are-moving-faster-than-ai-833441f3fb21 Comments URL: https://news.ycombinator.com/item?id=49341102 Points: 2 # Comments: 0
🤖 LLM📊 Bloomberg Tech🕐 53 分钟前
为什么中国的 DeepSeek、Qwen 和 Moonshot 令美国人工智能竞争对手担忧
Chinese AI models are cheaper and more adaptable than the preeminent US platforms, and studies suggest they’re now almost as proficient. How did that happen?
🛠️ AI产品📰 arXiv cs.AI🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14550v1 Announce Type: new Abstract: AI efficiency has recently taken the spotlight in both academy and industry due to massive model scales, high energy demands, and environmental costs.
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
大型语言模型显示医学推理中的元认知敏感性
arXiv:2608.14552v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly evaluated and used in medicine, but clinical usefulness depends on answer accuracy and whether confidence
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
不成文的基准：抽象感知推理中多模态机器学习的新挑战
arXiv:2608.14558v1 Announce Type: new Abstract: Current multimodal models have demonstrated remarkable proficiency in recognizing static visual and auditory content. However, their capacity for abstr
🔬 研究📰 arXiv cs.AI🕐 55 分钟前
何时进行沟通：多智能体强化学习中原则性门控的置信分布和 KL 散度
arXiv:2608.14559v1 Announce Type: new Abstract: Effective communication in multi-agent reinforcement learning requires agents to decide not only \textit{what} to communicate, but when? Existing appro
🔬 研究📰 arXiv cs.AI🕐 55 分钟前
全球人工智能高风险用例中的公平和道德法规：比较审查
arXiv:2608.14562v1 Announce Type: new Abstract: AI governance is shifting from voluntary ethics to enforceable, risk-based regulation, yet cross-jurisdictional divergence creates compliance uncertain
🔬 研究📰 arXiv cs.AI🕐 55 分钟前
立场：AI锁定正在进行，我们必须做好准备
arXiv:2608.14565v1 Announce Type: new Abstract: AI safety research has mainly focused on two areas: technical alignment (ensuring AI systems produce human-aligned outputs) and the regulation of gener
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
立场：人工智能道德推理的评价仍遗漏了一半
arXiv:2608.14566v1 Announce Type: new Abstract: Recent work on evaluating the moral competence of large language models (LLMs) has focused primarily on what we call the moral value problem, i.e., whe
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
从道尔到年度股东大会：信念改变的调查和实施路线图
arXiv:2608.14567v1 Announce Type: new Abstract: This paper presents a targeted narrative review establishing the historical and theoretical foundations for computational belief change implementation.
🔬 研究📰 arXiv cs.AI🕐 55 分钟前
立场：人工智能治理需要类似 ISO 的互操作性协议，而不仅仅是法律
arXiv:2608.14568v1 Announce Type: new Abstract: As Artificial Intelligence (AI) systems become deeply integrated into critical global infrastructure, the urgency for robust governance frameworks has
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14569v1 Announce Type: new Abstract: Neural solvers for constraint satisfaction problems have achieved remarkable in-distribution accuracy, yet they suffer from a fundamental limitation pe
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
立场：想要更好的机器学习评论？停止善意的询问，开始用信用体系进行激励
arXiv:2608.14571v1 Announce Type: new Abstract: With soaring submission counts, stricter reciprocal review policies, widespread adoption of platforms like OpenReview, and without the offsetting press
🔬 研究📰 arXiv cs.AI🕐 55 分钟前
ABCD 研究中青少年药物使用起始的纵向和图形增强预测
arXiv:2608.14578v1 Announce Type: new Abstract: Early identification of adolescent substance-use risk is an important prevention challenge, yet the relative value of baseline characteristics, longitu
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
技能：用于逻辑优化的自我修正知识引导迭代大语言模型代理
arXiv:2608.14579v1 Announce Type: new Abstract: Logic synthesis optimization poses significant challenges due to exponentially growing search spaces, sparse reward signals, and diverse logic structur
🛠️ AI产品📰 arXiv cs.AI🕐 55 分钟前
OGX：开源、供应商中立的生成式人工智能应用服务器
arXiv:2608.14580v1 Announce Type: new Abstract: OGX (Open GenAI Stack) is an open-source AI application server and Python library that implements the APIs of major frontier labs (OpenAI, Anthropic, G
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
Euclid-Omni：平面几何的统一神经符号框架
arXiv:2608.14585v1 Announce Type: new Abstract: Euclidean geometry is a compelling testbed for AI reasoning, as it demands the combination of intuitive diagram understanding, axiomatic deduction, and
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
使用规则和法学硕士嵌入和注释描述性文档布局的代理框架：植物科学用例
arXiv:2608.14587v1 Announce Type: new Abstract: Background: Recent advances in information retrieval (IR) leverage both dense and sparse representations, large language models (LLMs), and specialized
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
幻觉雪球：将错误传播建模为多代理 LLM 管道中的状态转换
arXiv:2608.14588v1 Announce Type: new Abstract: Sequential multi-agent LLM pipelines chain specialized agents without verification at handoffs, creating a structural flaw with measurable and severe c
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
走向安全的法学硕士代理人：规范、验证和执行的调查
arXiv:2608.14590v1 Announce Type: new Abstract: LLM agents increasingly perform irreversible real-world actions, including database updates, API calls, file operations, and autonomous use of tools. H
🛠️ AI产品📰 arXiv cs.AI🕐 55 分钟前
立场：医疗人工智能忽视了真正的治疗结果
arXiv:2608.14598v1 Announce Type: new Abstract: Medical AI has rapidly improved its ability to perform diagnostic and prognostic tasks that lead to treatment decisions. But understanding of treatment
🤖 LLM📰 arXiv cs.AI🕐 55 分钟前
法学硕士何时适用错误的法律？诊断法学硕士在时间法律推理中的失败
arXiv:2608.14610v1 Announce Type: new Abstract: Legal reasoning tasks such as legal judgment prediction (LJP) require identifying the temporally correct version of the law governing a case -- a capab
🔬 研究📰 arXiv cs.LG🕐 55 分钟前
使用 Cochain-Frame 等变学习物理场的离散黎曼度量
arXiv:2608.14556v1 Announce Type: new Abstract: Physical fields on meshes require a separation between topology and geometry: conservation laws are topological and should be exact, while geometry, ma
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
前向传播域适应（无跨层反向传播）
arXiv:2608.14563v1 Announce Type: new Abstract: Forward-Pass-Only MLP training (FPO) adapts large language models without a backward pass through the model body, achieving 2.7--3.2x the throughput of
🔬 研究📰 arXiv cs.LG🕐 55 分钟前
用于城市系统轨迹生成的从粗到细的多分辨率扩散模型
arXiv:2608.14570v1 Announce Type: new Abstract: Understanding human mobility is critical for a wide range of urban applications, including traffic management, epidemic control, and urban planning. Ho
🔬 研究📰 arXiv cs.LG🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14594v1 Announce Type: new Abstract: Projected Gradient Descent (PGD) is widely used to evaluate adversarial robustness, typically via final adversarial accuracy, which does not capture mo
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
DumpsterCluster：从垃圾箱潜水到在 60 美元的 GPU 上为 LLaMA-70B 提供服务
arXiv:2608.14614v1 Announce Type: new Abstract: As AI datacenters retire functional GPUs, vast quantities of still capable accelerators enter secondary markets. This paper investigates whether these
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
校准信任，而不是更清晰的预测：不确定性融合的实证检验
arXiv:2608.14617v1 Announce Type: new Abstract: A recurring proposal in legal AI is to improve case-outcome prediction by fusing uncertainty tools (evidence graphs with belief propagation, sequential
🛠️ AI产品📰 arXiv cs.LG🕐 55 分钟前
PIKFNO：基于物理通知核函数的可解释神经算子
arXiv:2608.14619v1 Announce Type: new Abstract: This work proposes a new interpretable neural operator framework, termed the Physics Informed Kernel Function Neural Operator (PIKFNO), which explicitl
🔬 研究📰 arXiv cs.LG🕐 55 分钟前
解释自适应系统中的强化学习决策
arXiv:2608.14620v1 Announce Type: new Abstract: Reinforcement Learning (RL) has been extensively used in autonomous and self-* systems, but RL policies, especially deep RL ones relying on neural netw
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
形塑性作为增量学习的自适应梯度预处理
arXiv:2608.14634v1 Announce Type: new Abstract: Biological intelligence naturally prevents catastrophic forgetting through Complementary Learning Systems (CLS) theory, a macroscopic consolidation pro
🔬 研究📰 arXiv cs.LG🕐 55 分钟前
分数优化器满足分形激活函数：神经网络多尺度优化的实证研究
arXiv:2608.14636v1 Announce Type: new Abstract: Fractional optimization methods and fractal activation functions are two independent directions for improving neural network training. Fractional optim
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
基于FlowBD-E1的铁铬液流电池早期充电轨迹生成预测及全生命周期健康管理
arXiv:2608.14637v1 Announce Type: new Abstract: Long-duration stationary energy storage requires batteries whose degradation can be detected before substantial capacity loss has accumulated. Iron-chr
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
随机初始化的自动编码器：固定点和混沌边缘
arXiv:2608.14638v1 Announce Type: new Abstract: In this paper we study autoencoders, a special class of deep neural nets (DNNs) whose performance can be characterized via their fixed points. This per
🔬 研究📰 arXiv cs.LG🕐 55 分钟前
用于文档提取的有效每字段选择性风险控制：三种失败模式、有效性阶梯以及条件何时发挥作用
arXiv:2608.14639v1 Announce Type: new Abstract: Per-field accept/review with selective risk at most alpha -- accept a field only if the error rate among accepted fields is controlled -- is the trust
🛠️ AI产品📰 arXiv cs.LG🕐 55 分钟前
BDIP-Net：用于双层材料性能预测的双交互图学习
arXiv:2608.14640v1 Announce Type: new Abstract: Stacked bilayer materials exhibit rich stacking-dependent properties driven by the interplay between strong intra-layer bonding and weak inter-layer va
🔬 研究📰 arXiv cs.LG🕐 55 分钟前
根据每集分布训练和评估道德强化学习代理
arXiv:2608.14642v1 Announce Type: new Abstract: Reinforcement Learning (RL) agents trained on a single reward signal exploit the gap between the designed reward and the intended behavior. This is par
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
DUET：通过同权重分歧实现双师政策蒸馏以实现禁令合规
arXiv:2608.14644v1 Announce Type: new Abstract: Real-world LLM deployments increasingly rely on runtime-injected prohibitions--enterprise policies, PII redlines, tool boundaries--that vary per reques
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
基于高效神经网络的高分辨率辐射传输 CO___ 检索及其在干涉测量中的应用
arXiv:2608.14645v1 Announce Type: new Abstract: Studying climate change requires reducing uncertainties in CO2 and CH4 emission estimates to better distinguish anthropogenic from natural sources, whi
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
iFuzz-Meta：一种可解释的模糊学习框架，桥接自上而下和自下而上的知识集成
arXiv:2608.14646v1 Announce Type: new Abstract: Interpretable representation learning remains a key challenge in modern neural computation, particularly when models are expected not only to perform b
🔬 研究📰 arXiv cs.LG🕐 55 分钟前
SMOPD：用于肮脏历史多轮开启策略自蒸馏的选择性令牌熵掩蔽
arXiv:2608.14647v1 Announce Type: new Abstract: Dirty-history rollouts make multi-turn on-policy self-distillation (OPSD) brittle: once a student emits an erroneous intermediate reply, later turns ar
🤖 LLM📰 arXiv cs.LG🕐 55 分钟前
离散扩散语言模型是免训练的多标签分类器
arXiv:2608.14649v1 Announce Type: new Abstract: We present dLLM-SetScore, a training-free method that uses discrete masked-diffusion language models for multi-label text classification. For each cand
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
法学硕士辅助系统审查筛选的辅助不确定性信号：八个 Cohen 药物类别审查的基准
arXiv:2608.14551v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used for title-abstract screening in systematic reviews, but their decisions lack calibrated uncertainty.
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
HarmProfile：表征前沿法学硕士中的有害分布
arXiv:2608.14577v1 Announce Type: new Abstract: Frontier large language models (LLMs) safety evaluation has largely treated harmful generation as an attack outcome rather than as an object of analysi
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
多模态生成模糊系统：模糊推理引导的大模型交互式问答框架
arXiv:2608.14584v1 Announce Type: new Abstract: In Multimodal Question Answering (MQA), models are required to jointly encode and integrate heterogeneous information from multiple modalities, includi
🛠️ AI产品📰 arXiv cs.CL🕐 55 分钟前
Wiola 13M，用于参数高效小语言模型的门控螺旋注意力架构
arXiv:2608.14604v1 Announce Type: new Abstract: Small language models in the ten to one hundred million parameter range are attractive for on device inference, rapid experimentation, and controlled s
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
AutoMem：用于自动内存架构搜索的文本梯度递归自我改进框架
arXiv:2608.14621v1 Announce Type: new Abstract: Long-term memory is increasingly central to LLM agents, yet memory design remains a highly coupled architecture problem: what to encode, how to store i
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
低资源语言的法学硕士安全调整：系统文献综述
arXiv:2608.14626v1 Announce Type: new Abstract: Large Language Models (LLMs) have achieved substantial progress in safety alignment, yet their safety guarantees remain significantly weaker in low-res
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
大型语言模型中对抗性政治偏见的推理时间缓解
arXiv:2608.14629v1 Announce Type: new Abstract: As Large Language Models (LLMs) become the mainstay for information retrieval and summarization tasks, ensuring that they are always non-partisan and i
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
用语言模型表征决策中的修辞不一致
arXiv:2608.14630v1 Announce Type: new Abstract: Human decision-making is often shaped by a range of well-documented cognitive biases. As large language models (LLMs) become increasingly integrated in
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
DeMTS：作为多元时间序列的去噪轨迹，用于扩散语言模型中的幻觉检测
arXiv:2608.14632v1 Announce Type: new Abstract: Diffusion large language models (D-LLMs) have emerged as a promising paradigm for text generation. However, similar to autoregressive LLMs, D-LLMs rema
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
自动还是受控？重复启动揭示了基础法学硕士、指导法学硕士和人类的不同处理过程
arXiv:2608.14681v1 Announce Type: new Abstract: Words recur constantly in natural language use, yet it remains unclear whether language models reactivate prior representations or re-evaluate repeated
🔬 研究📰 arXiv cs.CL🕐 55 分钟前
使用指令调整根据自然语言规则进行与领域无关的文本编辑
arXiv:2608.14693v1 Announce Type: new Abstract: With the increasing digitization of personal and corporate communication, the automatic sanitization of textual data has become a crucial component of
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
您的注意力指标回答了哪个问题？作为成分数据的注意行
arXiv:2608.14712v1 Announce Type: new Abstract: Each row of a transformer's attention matrix is a probability distribution over tokens, and in trained models most of that probability lands on a singl
🔬 研究📰 arXiv cs.CL🕐 55 分钟前
基于法学硕士的系统评价筛选中的类别不平衡和批次效应
arXiv:2608.14737v1 Announce Type: new Abstract: This study analyses LLMs in imbalanced binary classification, using study screening in systematic reviews as the application domain. An experiment was
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
提示是不够的：监督基线和泄漏控制用于衡量与法学硕士在儿科接触中的共同决策
arXiv:2608.14792v1 Announce Type: new Abstract: Objectives: To determine whether zero-shot prompting of a large language model (LLM) is sufficient to detect shared decision-making (SDM) behaviors in
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
超越代币：大型语言和视觉语言模型解码方法的调查
arXiv:2608.14797v1 Announce Type: new Abstract: Large language models (LLMs) and large vision-language models (LVLMs) have demonstrated impressive generative capabilities, yet ensuring their outputs
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14813v1 Announce Type: new Abstract: Despite a strong interest on the part of the research community in the topic of trustworthy and safe AI, the composition of the text corpora that large
🛠️ AI产品📰 arXiv cs.CL🕐 55 分钟前
写作风格相似反映学术谱系
arXiv:2608.14843v1 Announce Type: new Abstract: As authorship attribution systems are increasingly deployed to detect ghostwritten and AI-generated papers, their errors can support accusations agains
🔬 研究📰 arXiv cs.CL🕐 55 分钟前
忘却时要忘记什么？忘记语言模型的设置管理
arXiv:2608.14855v1 Announce Type: new Abstract: Machine unlearning aims to remove targeted data or behaviors from a trained model without retraining from scratch. Yet most evaluations assume that the
🔬 研究📰 arXiv cs.CL🕐 55 分钟前
哪里检索失败？评估农业咨询的 RAG 架构
arXiv:2608.14886v1 Announce Type: new Abstract: Retrieval quality in RAG systems is commonly reported as a single aggregate score, which can hide large differences across query types and language con
🤖 LLM📰 arXiv cs.CL🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14896v1 Announce Type: new Abstract: Large language models work well on English and behave in poorly understood ways on languages typologically far from it. Japanese is a clean example, wh
🔬 研究📰 arXiv cs.CV🕐 55 分钟前
Xemo-Talker：明确解锁音频驱动的说话肖像合成的情感
arXiv:2608.14700v1 Announce Type: new Abstract: Precise emotion control in audio-driven talking heads remains a challenge due to the reliance on implicit emotion regulation in existing systems, which
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
眼周软生物识别技术：多媒体取证和虚假信息检测的调查和应用
arXiv:2608.14701v1 Announce Type: new Abstract: Soft-biometric attributes such as gender, age, and ethnicity provide valuable ancillary evidence when full identity recognition is not feasible, suppor
🔬 研究📰 arXiv cs.CV🕐 55 分钟前
深度模拟：具有参考条件 3D LUT 的开放式薄膜仿真
arXiv:2608.14702v1 Announce Type: new Abstract: Film emulation reproduces the look of an analog film stock on a new digital photograph. We target its open-set form -- matching any reference film fram
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14705v1 Announce Type: new Abstract: Hyperparameter optimization (HPO) can materially affect the performance of deep learning (DL) image classifiers, but there is little empirical guidance
👁️ CV📰 arXiv cs.CV🕐 55 分钟前
平衡强迫：无需噪声调节的自适应视频生成
arXiv:2608.14706v1 Announce Type: new Abstract: Standard autoregressive video generation algorithms based on Diffusion and Flow Matching rely on rigid training objectives and static sampling schedule
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
PE-CSNet：具有可学习的基于补丁的稀疏表示的等变网络架构
arXiv:2608.14708v1 Announce Type: new Abstract: Compressive sensing (CS) enables accurate signal reconstruction from sparse measurements and is widely applied in medical imaging, remote sensing, and
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
Path2ST：空间转录组学的分层细胞组织基础跨模式翻译
arXiv:2608.14710v1 Announce Type: new Abstract: Predicting spatial gene expression from hematoxylin and eosin (H\&E)-stained images offers a cost-effective alternative to spatial transcriptomics
🔬 研究📰 arXiv cs.CV🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14717v1 Announce Type: new Abstract: A query-relation deletion can improve the edited slot while reducing the utility of the prediction set that contains it. We study this tension in two r
🤖 LLM📰 arXiv cs.CV🕐 55 分钟前
VideoGAIA：通用人工智能助理在代理视频理解方面的基准
arXiv:2608.14718v1 Announce Type: new Abstract: Video understanding is a fundamental task for evaluating the capabilities of multimodal large language models (MLLMs). However, existing leading models
🤖 LLM📰 arXiv cs.CV🕐 55 分钟前
DeCo-MIL：长尾全幻灯片图像分析的去偏反事实推理
arXiv:2608.14719v1 Announce Type: new Abstract: Multiple instance learning (MIL) is widely used for weakly supervised whole slide image (WSI) analysis. However, under long-tailed distributions, MIL-b
🔬 研究📰 arXiv cs.CV🕐 55 分钟前
AeroGround：空地协同推理的综合基准
arXiv:2608.14721v1 Announce Type: new Abstract: Vision-language models (VLMs) have been widely employed in understanding and reasoning tasks for unmanned aerial vehicles (UAVs). Existing UAV benchmar
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
用于多视图视网膜眼底成像中的中风检测的编织视觉变压器
arXiv:2608.14722v1 Announce Type: new Abstract: Stroke remains a leading cause of mortality and morbidity worldwide, emphasizing the importance of its accurate and immediate assessment. Retinal fundu
🔬 研究📰 arXiv cs.CV🕐 55 分钟前
用于跨多个临床站点基于心电图检测左心室收缩功能障碍的视觉变压器
arXiv:2608.14723v1 Announce Type: new Abstract: Reduced left ventricular ejection fraction (LVEF) is frequently asymptomatic and often detected only after advanced heart failure develops. Electrocard
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14724v1 Announce Type: new Abstract: The rapid advancement of intelligent transportation systems and autonomous driving relies heavily on multi-modal urban traffic datasets. However, curat
👁️ CV📰 arXiv cs.CV🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14725v1 Announce Type: new Abstract: We present a novel causal approach to interpretability for computer vision models that dynamically masks the input image prior to classification. The i
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
边缘的低成本两阶段织物缺陷检测
arXiv:2608.14727v1 Announce Type: new Abstract: Fabric inspection in the garment industries of low-income economies remains largely manual, and commercial vision systems are priced beyond most small
👁️ CV📰 arXiv cs.CV🕐 55 分钟前
CNN 内部表示真假图像的方式不同吗？隐藏层分析
arXiv:2608.14729v1 Announce Type: new Abstract: Fake/synthetic images are increasingly prevalent, but it remains unclear whether Convolutional Neural Networks (CNNs) process real and fake images in t
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14730v1 Announce Type: new Abstract: The rapid evolution of visual generative AI has introduced a wide range of intellectual property risks, spanning the unauthorized learning, reproductio
🛠️ AI产品📰 arXiv cs.CV🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14731v1 Announce Type: new Abstract: Worldwide, millions of senior citizens are suffering from Alzheimer disease abbreviated as AD, a well- versed form of dementia. AD is featured by amnes
👁️ CV📰 arXiv cs.CV🕐 55 分钟前
从密集预测到可视化编辑：统一图像和视频创建的结构化监督
arXiv:2608.14740v1 Announce Type: new Abstract: Unified image and video creation requires a model to follow diverse instructions while preserving identity, geometry, and temporal structure from visua
👁️ CV📰 arXiv cs.RO🕐 55 分钟前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
arXiv:2608.14713v1 Announce Type: new Abstract: Robots operating in dark or poorly lit environments rely on onboard lights, which often produce uneven illumination that degrades downstream perception
👁️ CV📰 arXiv cs.RO🕐 55 分钟前
MISTac：用于微创手术的基于视觉的触觉传感器
arXiv:2608.14772v1 Announce Type: new Abstract: Minimally invasive and robot-assisted surgery offer many advantages over traditional open surgery, but deprive surgeons of tactile feedback and the abi
🤖 LLM📰 arXiv cs.RO🕐 55 分钟前
想象恢复：视觉-语言-动作模型的推理时间反事实调整
arXiv:2608.14822v1 Announce Type: new Abstract: Vision-language-action (VLA) models have improved the flexibility and generality of robotic manipulation, yet they remain fragile to online disruptions
🤖 LLM📰 arXiv cs.RO🕐 55 分钟前
用于优化设计的鳗鱼软体机器人的建模和控制
arXiv:2608.14860v1 Announce Type: new Abstract: Anguilliform locomotion is a highly efficient swimming mode; the advent of new materials for soft robots enables the development of an eel-inspired sof
🦾 机器人📰 arXiv cs.RO🕐 55 分钟前
受鳗鱼启发的软机器人上执行器控制和健康状况 (REACH) 的实时估计器
arXiv:2608.14865v1 Announce Type: new Abstract: An actuator health estimation algorithm for a soft swimming robot that can perform anguilliform swimming is developed. Due to harsh operational environ
👁️ CV📰 arXiv cs.RO🕐 55 分钟前
用于 3D 高斯泼溅 SLAM 的几何感知在线建图
arXiv:2608.14902v1 Announce Type: new Abstract: Recent 3D Gaussian Splatting (3DGS) has enabled efficient photorealistic view synthesis and is rapidly being adopted in simultaneous localization and m
🔬 研究📰 arXiv cs.RO🕐 55 分钟前
从连续设计到延迟感知离散综合：保证 PMSM 驱动器的高带宽联合控制
arXiv:2608.14937v1 Announce Type: new Abstract: The increasing dynamic demands of modern robotic joints require current controllers to achieve high bandwidth over wide operating ranges of speed, acce
🤖 LLM📰 arXiv cs.RO🕐 55 分钟前
SkillComposer：学习自然语言机器人编程的可重用技能
arXiv:2608.14944v1 Announce Type: new Abstract: Natural-language interfaces can lower the barrier to programming robots, but existing systems struggle when users request complex tasks. While large la
🔬 研究📰 arXiv cs.RO🕐 55 分钟前
缺席的证据：当视觉失败时，跨模式溯因风险感知维持世界模型
arXiv:2608.14952v1 Announce Type: new Abstract: A structured world-state (entities, relations, context, and predictive cues) is designed to preserve prediction-critical content when perception degrad
🔬 研究📰 arXiv cs.RO🕐 55 分钟前
GaussMemory：用于长视野机器人操作的任务驱动 3D 高斯场景记忆
arXiv:2608.14986v1 Announce Type: new Abstract: Long-horizon robotic manipulation fundamentally relies on persistent spatial memory. However, existing 3D memory systems function merely as passive rec
🔬 研究📰 arXiv cs.RO🕐 55 分钟前
HP2-SLAM：自适应混合 ICP，实现稳健、高效的 LiDAR SLAM
arXiv:2608.14996v1 Announce Type: new Abstract: Achieving robustness, accuracy, and efficiency simultaneously remains a central challenge in light detection and ranging (LiDAR) simultaneous localizat
🔬 研究📰 arXiv cs.RO🕐 55 分钟前
用于机器人策略训练的冻结视觉编码器的 NPU 卸载
arXiv:2608.15002v1 Announce Type: new Abstract: When a robot policy is trained for a new task or dataset, its visual encoder can be frozen and only its action generation module trained, reducing trai
🛠️ AI产品📰 arXiv cs.RO🕐 55 分钟前
ForceU-VLA：用于实体超声扫描的力感知视觉-语言-动作模型
arXiv:2608.15009v1 Announce Type: new Abstract: Embodied intelligent ultrasound scanning enables the automation and standardization of the ultrasound examination process by integrating perception, de
🤖 LLM📰 arXiv cs.RO🕐 55 分钟前
MotionGS-SLAM：用于运动模糊鲁棒 SLAM 的事件调制高斯泼溅
arXiv:2608.15024v1 Announce Type: new Abstract: Current Vision-based SLAM systems fail catastrophically when motion blur corrupts the visual input, as they attempt the ill-posed inverse problem of re
🔬 研究📰 arXiv cs.RO🕐 55 分钟前
PACE：长期具体操纵的阶段进展意识信用
arXiv:2608.15026v1 Announce Type: new Abstract: Post-training of vision-language-action (VLA) models typically relies on expert demonstrations and policy interaction trajectories. However, in long-ho
🔬 研究📰 arXiv cs.RO🕐 55 分钟前
用于人在环在线机器人学习的 Max-Q 选择性模仿
arXiv:2608.15088v1 Announce Type: new Abstract: Human-in-the-loop (HIL) online reinforcement learning for real robots must absorb human interventions quickly while continuing to improve beyond the hu
🔬 研究📰 arXiv cs.RO🕐 55 分钟前
StructRL：基于流的 VLA 的结构化动作空间探索
arXiv:2608.15139v1 Announce Type: new Abstract: Flow-based Vision-Language-Action (VLA) models are now widely used for continuous robotic manipulation, and online reinforcement learning (RL) is emerg
🛠️ AI产品📰 arXiv cs.RO🕐 55 分钟前
用于学习世界模型中反事实推出的低阶动态有效潜在载体
arXiv:2608.15156v1 Announce Type: new Abstract: World models may predict the future without making clear which parts of their hidden state actually drive those predictions. We ask whether a small, di
🤖 LLM📰 arXiv cs.RO🕐 55 分钟前
LAPF：使用 UAVScenes 数据集的基于 LLM 代理的路径查找器
arXiv:2608.15175v1 Announce Type: new Abstract: Uncrewed aerial vehicles (UAVs) are increasingly deployed for autonomous navigation in complex outdoor environments, where dynamic conditions and missi
🤖 LLM📰 arXiv cs.RO🕐 55 分钟前
更聪明地记住：用于机器人记忆的视觉历史压缩器和双曲体验空间
arXiv:2608.15269v1 Announce Type: new Abstract: Long-horizon robot policies require compact access to recent observations and reusable experience without expanding the vision-language-action (VLA)
🤖 LLM📰 Hacker News AI🕐 1 小时前
为什么它还没有发生：强大的人工智能和恶意意图
Article URL: https://substack.norabble.com/p/why-it-hasnt-happened-yet Comments URL: https://news.ycombinator.com/item?id=49341070 Points: 2 # Comments: 0
📈 市场📈 Bloomberg Markets🕐 1 小时前
刚果埃博拉疫情：随着不信任的加剧，金融监管成为优先事项
Ebola is reviving suspicions of profiteering in eastern Democratic Republic of Congo, where distrust over how response money is handled is undermining the community cooperation needed to contain the w
🔧 硬件🏠 IT之家🕐 1 小时前
三星 Galaxy A08/A07s 手机曝光：联发科 Helio G99 芯片、水滴屏
IT之家 8 月 18 日消息，科技媒体 91Mobile 昨日（8 月 17 日）发布博文，报道称在 Google Play Console 数据库中，挖掘发现 Galaxy A08 和 Galaxy A07s 两款三星手机踪迹。根据页面信息，三星  Galaxy A08 型号为 SM-A085F。根据信息显示，这款入门级手机将搭载联发科处理器（此前曝光信息为联发科 Helio G99
🌐 互联网🏠 IT之家🕐 1 小时前
英国政府资助梅赛德斯-奔驰旗下 YASA，减少电动汽车电机稀土依赖
IT之家 8 月 18 日消息，梅赛德斯-奔驰旗下牛津大学衍生企业 YASA 当地时间本月 10 日宣布，该公司获得英国政府 DRIVE35 零排放汽车技术创新计划拨款，将作为牵头方领导 Project Resilience“韧性计划”。IT之家获悉，“韧性计划”旨在降低电动汽车电机对稀土磁性材料的依赖，强化供应链安全，同时保障电机可满足未来车辆对质量、体积、效率的要求。其包括两条互补技术路径：面
📈 市场📈 Bloomberg Markets🕐 1 小时前
由于人民币支持减弱，中国企业七月份外汇销售放缓
Net sales of foreign exchange by Chinese companies fell to a eight-month low in July, suggesting a key source of support for the yuan is fading.
🔧 硬件🏠 IT之家🕐 1 小时前
三星 Galaxy S26 FE 手机曝光：Exynos 2500 芯片，8GB 内存
IT之家 8 月 18 日消息，科技媒体 YtechB 昨日（8 月 17 日）发布博文，报道称三星 Galaxy S26 FE（型号 SM-S741U）手机现身 Google Play Console 数据库，显示其搭载 Exynos 2500 芯片以及三星 Xclipse 950 GPU。页面信息显示，三星 Galaxy S26 FE 手机配备 8GB 内存，屏幕分辨率为 1080 x 234
📦 软件🏠 IT之家🕐 1 小时前
【IT之家图赏】石头迷你洗烘一体机 Z1 Mini Pro 实拍：1.08 超高洗净比，2kg/1kg 洗烘旗舰
8 月 18 日，石头科技正式发布石头迷你洗烘一体机 Z1 Mini Pro，限时到手价 2999.7 元。作为主打母婴、贴身衣物洗护领域的“性能小钢炮”，其具备 2kg 洗涤与 1kg 烘干容量，小巧的机身内塞入了 1.08 洗净比、石头蓝氧慕斯洗 3.0 以及蓝氧 +UV 双重除菌等配置。IT之家现已拿到这款产品的云霜白版本，接下来为大家带来一组实拍图赏。在外观设计上，石头 Z1 Mini P
📈 市场📈 Bloomberg Markets🕐 1 小时前
日本五年期债券发行需求自 2025 年 6 月以来最强劲
Japan’s five-year government bond auction Tuesday saw the strongest demand since June 2025, as elevated yields underpinned buying.
📈 市场📈 Bloomberg Markets🕐 1 小时前
必和必拓首席执行官：中国铁矿石需求保持稳定​‌
China's iron ore demand remains steady, supported by manufacturing activity, according to BHP CEO Brandon Craig. He also says stronger Indian demand for metallurgical coal should support the market. H
🤖 LLM📰 Hacker News AI🕐 1 小时前
150 年前的圣地亚哥：罕见的人工智能重建视觉效果 [视频]
Article URL: https://www.youtube.com/watch?v=qWaR_nc5sKM Comments URL: https://news.ycombinator.com/item?id=49340969 Points: 2 # Comments: 0
📦 软件🏠 IT之家🕐 1 小时前
思特威推出超星光级 1/2.8 英寸 CMOS 图像传感器，主打高清夜视全彩
IT之家 8 月 18 日消息，近日，CMOS 图像传感器供应商思特威正式推出其 Star Light（SL）超星光级系列智能安防应用图像传感器新品 ——SC185SL。作为 1/2.8 英寸大靶面背照式图像传感器，SC185SL 基于思特威 SmartClarity-3 工艺技术打造，搭载 SFCPixel、Micro-Pillar 近红外增强等核心技术，具备高感度、低噪声、超低功耗等多项性能优
📈 市场📈 Bloomberg Markets🕐 1 小时前
Error 500 (Server Error)!!1500.That’s an error.There was an error. Please try again later.That’s all we know.
Caught off-guard by India’s decision to close early a special window to draw foreign-currency deposits, traders are now paring bets on gains in the rupee and short-term bonds.
分享这个页面

觉得有用，就顺手递给下一个可能会踩坑的人。

分享到：TwitterFacebookLinkedIn微博QQQQ空间