1. 大模型 · 梁赛 · 今天情人节，字节跳动正式发布通用大模型doubao2.0 海豚AI智能体刚刚第一时间接入了do…
分类：产品/发布来源：weibo_ai_trending分数：50作者：梁赛时间：2026-02-14 14:48
字节跳动正式发布通用大模型Doubao 2.0，海豚AI智能体已首批接入其Pro版并用于“海豚投研”多专家辩论场景，显示新模型在Agent推理与商业化应用上的落地价值。
- 字节跳动在情人节当天正式发布通用大模型Doubao 2.0。
- 海豚AI智能体第一时间接入Doubao 2.0 Pro版（最强且价格最高版本）。
- 发布者强调该场景高度依赖Agent推理能力，接入后体验明显提升并表现兴奋。
- “海豚投研”产品形态为5个AI投资专家同场辩论，面向投研决策辅助。
- 该话题在微博有较高互动（转发507、评论105、点赞1078），关注度较高。
#大模型 #Weibo #产品/发布 #字节跳动 #AI智能体 #Agent推理 #投研 #Agent #推理
2. Sora · 新侠客行 · 明天看看AI应用还能持续吗，其实每年年初一季度都有类似的传媒应用题材，从炒元宇宙开始，后面炒C…
分类：社交讨论/观点来源：weibo_ai_trending分数：36作者：新侠客行时间：2026-02-10 20:27
微博用户质疑“AI应用”题材行情可持续性，指出市场多年反复切换元宇宙、ChatGPT、Sora等概念炒作而业绩增长不足，反映资金偏好主题驱动的短期交易。
- 作者认为每年一季度都会出现类似传媒与科技题材轮动炒作。
- 文中列举元宇宙、ChatGPT、Sora、短剧、AIGC、DeepSeek、AI营销、Seedance2.0等概念。
- 核心观点是相关股票反复叠加概念，但基本面业绩提升不明显。
- 内容体现对AI应用板块持续性的怀疑与对资金博弈行为的观察。
- 该话题在微博有一定传播度（转发31、评论100、点赞1730）。
#Sora #Weibo #社交讨论/观点 #AI应用 #AIGC
3. AI · 雷斯林Raist · 现在ai的问题是：Ai替代了很多人，毁了很多人的就业，但却没创造更多就业，也没赚什么钱。Ai就…
分类：社交讨论/观点来源：weibo_ai_trending分数：27作者：雷斯林Raist时间：2026-02-13 08:22
一则微博观点认为AI正在替代就业却未创造足够新岗位和利润，并可能引发通缩、冲击资本主义与大规模失业，反映了公众对AI经济后果的强烈焦虑与争议。
- 作者称AI已“替代很多人”，但没有带来更多就业机会。
- 帖子认为AI并未显著创造新增利润，反而压缩了原有赚钱空间。
- 核心判断包括：AI可能导致各国通缩。
- 作者进一步推测AI最终可能动摇甚至终结资本主义形态。
- 还提出“绝大部分人会失业、工作成为一种需要”的激进判断，带有强烈价值立场。
#AI #Weibo #社交讨论/观点 #资本主义
4. eyaltoledano/claude-task-master
分类：开源项目来源：github_search分数：100作者：eyaltoledano时间：2026-02-14T21:09:27Z
Taskmaster 是一个面向 AI 驱动开发的开源任务管理系统（支持 MCP 与 CLI、多模型与多编辑器），可将 PRD 自动转为可执行任务流程并通过可选工具集显著降低上下文 token 成本。
- 项目定位为 AI 开发任务管理系统，可与 Claude/Cursor 等聊天式 AI 工具协同工作。
- 提供 MCP（推荐）与命令行两种使用方式，支持 Cursor、Windsurf、VS Code、Q CLI 等编辑器/环境接入。
- 支持多模型与多供应商 API（Anthropic、OpenAI、Gemini、Perplexity、xAI、OpenRouter），并区分主模型/研究模型/回退模型。
- 强调以 PRD 驱动任务拆解，内置解析需求、规划下一步、展开任务、研究最佳实践等常用指令流程。
- MCP 工具支持按模式加载（all/standard/core/custom），可从约 21000 tokens 降至约 5000 tokens 以优化上下文占用与性能。
#GitHub #repo #开源项目 #Taskmaster #MCP #Cursor #Claude #CLI
5. microsoft/RD-Agent
分类：开源项目来源：github_search分数：44作者：microsoft时间：2026-02-14T15:42:46Z
微软开源RD-Agent并持续发布量化与数据科学扩展，其在MLE-bench上取得当前公开领先成绩，显示多智能体自动化研发在工业级机器学习与量化策略中的实用价值。
- 项目定位为数据驱动的多智能体R&D框架，以“R(研究想法)+D(工程实现)”协同自动化研发流程。
- 官方公布在75个Kaggle任务构成的MLE-bench上处于领先，All指标约30.22%，高于此前公开基线AIDE。
- 发布RD-Agent(Q)量化版本，宣称在真实股市实验中以低成本获得更高ARR并减少因子数量。
- 生态完善：提供Live Demo、视频、文档、技术报告、PyPI发行与完整CI质量体系。
- 支持LiteLLM作为默认后端并提供DeepSeek等配置，覆盖ChatCompletion、json_mode与embedding能力。
#GitHub #repo #开源项目 #RD-Agent #MLE-bench #LiteLLM #Agent
6. nhivp/Awesome-Embedded
分类：开源项目来源：github_search分数：32作者：nhivp时间：2026-02-14T23:32:16Z
GitHub 项目 Awesome-Embedded 汇总了嵌入式开发从面试、基础知识到 MCU/RTOS/Linux 驱动等海量学习与实战资源，重要性在于为工程师提供系统化、一站式的学习导航。
- 这是一个 curated list 类型仓库，聚焦嵌入式领域优质资料的集中索引。
- 目录覆盖广泛：MCU 编程、裸机开发、编译/链接、Bootloader、USB、RTOS、Linux 内核与驱动等。
- 按平台细分了大量资源，如 MSP430、TM4C123、STM32/STM8、ESP8266、Raspberry Pi、Beaglebone。
- 包含面试题、技能成长路径、教程课程、示例工程与工具链资料，兼顾入门与进阶。
- 额外收录技术博客、FAQ、书籍与实用技巧，便于持续学习与问题排查。
#GitHub #repo #开源项目 #MCU #RTOS
7. Project Genie | How world remixing works
分类：产品/发布来源：youtube_rss分数：0作者：Google DeepMind时间：2026-01-29T16:56:03+00:00
Google 推出实验性原型 Project Genie，可通过调整环境与角色提示词重混并探索多样化世界，面向 AI Ultra 订阅用户先行开放，显示其在交互式生成世界上的产品化推进。
- Project Genie 是 Google 的实验性研究原型，主打“世界重混（world remixing）”与探索。
- 用户可基于已有世界或图库模板，通过修改环境和角色提示词生成新世界。
- 当前仅向美国地区、18 岁以上的 Google AI Ultra 订阅者开放。
- 官方表示将逐步扩大可用范围，体现分阶段发布策略。
#YouTube #产品/发布 #Google #Project Genie
8. Project Genie | How image upload works
分类：视频/演讲来源：youtube_rss分数：0作者：Google DeepMind时间：2026-01-29T16:55:59+00:00
Google DeepMind 通过视频介绍 Project Genie 的图片上传功能：用户可用自有图片作为提示生成并探索多样世界，这重要在于降低个性化世界构建门槛并展示其实验性生成式交互能力。
- Project Genie 是 Google 的实验性研究原型，可创建并探索高度多样化的世界。
- 核心流程为先选择并上传图片，再将图片加入文本提示以生成基于个人参考图的世界。
- 该能力目前向美国地区 18 岁以上的 Google AI Ultra 订阅者开放，后续将逐步扩大可用范围。
#YouTube #视频/演讲 #Project Genie
9. Project Genie | How world sketching works
分类：视频/演讲来源：youtube_rss分数：0作者：Google DeepMind时间：2026-01-29T16:55:56+00:00
Project Genie is Google’s experimental research prototype that lets you create and explore infinitely diverse worlds.
Using Project Genie starts with World Sk…
- Project Genie is Google’s experimental research prototype that lets you create and explore infinitely diverse worlds
- Using Project Genie starts with World Sketching
- Describe your environment and character and see an image preview of your world made by Nano Banana Pro
- Then, modify your prompts to fine tune your world before entering
- Project Genie is currently available to Google AI Ultra subscribers (US only, 18+), with broader availability opening up gradually
- Learn more: https://labs
#YouTube #视频/演讲
10. MamaDino: A Hybrid Vision Model for Breast Cancer 3-Year Risk Prediction
分类：研究/论文来源：arxiv_search分数：100作者：Ruggiero Santeramo时间：2026-02-14T23:56:22Z
Breast cancer screening programmes increasingly seek to move from one-size-fits-all interval to risk-adapted and personalized strategies. Deep learning (DL) ha…
- Breast cancer screening programmes increasingly seek to move from one-size-fits-all interval to risk-adapted and personalized strategies
- Deep learning (DL) has enabled image-based risk models with stronger 1- to 5-year prediction than traditional clinical models, but leading …
- g
- , Mirai) typically use convolutional backbones, very high-resolution inputs (>1M pixels) and simple multi-view fusion, with limited explici…
- We hypothesised that combining complementary inductive biases (convolutional and transformer-based) with explicit contralateral asymmetry m…
- We present MamaDino, a mammography-aware multi-view attentional DINO model
#arXiv #paper #研究/论文
11. voice2mode: Phonation Mode Classification in Singing using Self-Supervised Speech Models
分类：研究/论文来源：arxiv_search分数：98作者：Aju Ani Justus时间：2026-02-14T23:51:53Z
We present voice2mode, a method for classification of four singing phonation modes (breathy, neutral (modal), flow, and pressed) using embeddings extracted fro…
- We present voice2mode, a method for classification of four singing phonation modes (breathy, neutral (modal), flow, and pressed) using embe…
- Prior work on singing phonation has relied on handcrafted signal features or task-specific neural nets; this work evaluates the transferabi…
- voice2mode extracts layer-wise representations from HuBERT and two wav2vec2 variants, applies global temporal pooling, and classifies the p…
- Experiments on a publicly available soprano dataset (763 sustained vowel recordings, four labels) show that foundation-model features subst…
- HuBERT embeddings obtained from early layers yield the best result (~95
- 7% accuracy with SVM), an absolute improvement of ~12-15% over the best traditional baseline
#arXiv #paper #研究/论文
12. GREPO: A Benchmark for Graph Neural Networks on Repository-Level Bug Localization
分类：研究/论文来源：arxiv_search分数：95作者：Juntong Wang时间：2026-02-14T23:22:15Z
Repository-level bug localization-the task of identifying where code must be modified to fix a bug-is a critical software engineering challenge. Standard Large…
- Repository-level bug localization-the task of identifying where code must be modified to fix a bug-is a critical software engineering chall…
- Standard Large Language Modles (LLMs) are often unsuitable for this task due to context window limitations that prevent them from processin…
- As a result, various retrieval methods are commonly used, including keyword matching, text similarity, and simple graph-based heuristics su…
- Graph Neural Networks (GNNs) offer a promising alternative due to their ability to model complex, repository-wide dependencies; however, th…
- To address this gap, we introduce GREPO, the first GNN benchmark for repository-scale bug localization tasks
- GREPO comprises 86 Python repositories and 47294 bug-fixing tasks, providing graph-based data structures ready for direct GNN processing
#arXiv #paper #研究/论文
13. A Comparative Analysis of Social Network Topology in Reddit and Moltbook
分类：研究/论文来源：arxiv_search分数：92作者：Yiming Zhu时间：2026-02-14T23:20:22Z
Recent advances in agent-mediated systems have enabled a new paradigm of social network simulation, where AI agents interact with human-like autonomy. This evo…
- Recent advances in agent-mediated systems have enabled a new paradigm of social network simulation, where AI agents interact with human-lik…
- This evolution has fostered the emergence of agent-driven social networks such as Moltbook, a Reddit-like platform populated entirely by AI…
- Despite these developments, empirical comparisons between agent-driven and human-driven social networks remain scarce, limiting our underst…
- This paper presents the first comparative analysis of network topology on Moltbook, utilizing a comment network comprising 33,577 nodes and…
- To provide a benchmark, we curated a parallel dataset from Reddit consisting of 7
- 8 million nodes and 51
#arXiv #paper #研究/论文
14. Common Knowledge Always, Forever
分类：研究/论文来源：arxiv_search分数：90作者：Martín Diéguez时间：2026-02-14T22:34:27Z
There has been an increasing interest in topological semantics for epistemic logic, which has been shown to be useful for, e.g., modelling evidence, degrees of…
- There has been an increasing interest in topological semantics for epistemic logic, which has been shown to be useful for, e
- g
- , modelling evidence, degrees of belief, and self-reference
- We introduce a polytopological PDL capable of expressing common knowledge and various generalizations and show it has the finite model prop…
- The latter is shown by embedding a version of linear temporal logic with `past', which does not have the finite model property
#arXiv #paper #研究/论文
15. From Pixels to Policies: Reinforcing Spatial Reasoning in Language Models for Content-Aware Layout Design
分类：研究/论文来源：arxiv_search分数：88作者：Sha Li时间：2026-02-14T22:31:49Z
We introduce LaySPA, a reinforcement learning framework that equips large language models (LLMs) with explicit and interpretable spatial reasoning for content-…
- We introduce LaySPA, a reinforcement learning framework that equips large language models (LLMs) with explicit and interpretable spatial re…
- LaySPA addresses two key challenges: LLMs' limited spatial reasoning and the lack of opacity in design decision making
- Instead of operating at the pixel level, we reformulate layout design as a policy learning problem over a structured textual spatial enviro…
- LaySPA produces dual-level outputs comprising interpretable reasoning traces and structured layout specifications, enabling transparent and…
- Layout design policy is optimized via a multi-objective spatial critique that decomposes layout quality into geometric validity, relational…
- Experiments demonstrate that LaySPA improves structural validity and visual quality, outperforming larger proprietary LLMs and achieving pe…
#arXiv #paper #研究/论文
16. Sufficient Conditions for Stability of Minimum-Norm Interpolating Deep ReLU Networks
分类：研究/论文来源：arxiv_search分数：85作者：Ouns El Harzli时间：2026-02-14T22:20:44Z
Algorithmic stability is a classical framework for analyzing the generalization error of learning algorithms. It predicts that an algorithm has small generaliz…
- Algorithmic stability is a classical framework for analyzing the generalization error of learning algorithms
- It predicts that an algorithm has small generalization error if it is insensitive to small perturbations in the training set such as the re…
- While stability has been demonstrated for numerous well-known algorithms, this framework has had limited success in analyses of deep neural…
- In this paper we study the algorithmic stability of deep ReLU homogeneous neural networks that achieve zero training error using parameters…
- We investigate sufficient conditions for such networks to be stable
- We find that 1) such networks are stable when they contain a (possibly small) stable sub-network, followed by a layer with a low-rank weigh…
#arXiv #paper #研究/论文
17. Quantifying Normality: Convergence Rate to Gaussian Limit for Stochastic Approximation and Unadjusted OU Algorithm
分类：研究/论文来源：arxiv_search分数：82作者：Shaan Ul Haque时间：2026-02-14T21:55:57Z
Stochastic approximation (SA) is a method for finding the root of an operator perturbed by noise. There is a rich literature establishing the asymptotic normal…
- Stochastic approximation (SA) is a method for finding the root of an operator perturbed by noise
- There is a rich literature establishing the asymptotic normality of rescaled SA iterates under fairly mild conditions
- However, these asymptotic results do not quantify the accuracy of the Gaussian approximation in finite time
- In this paper, we establish explicit non-asymptotic bounds on the Wasserstein distance between the distribution of the rescaled iterate at …
- As an immediate consequence, we obtain tail bounds on the error of SA iterates at any time
- We obtain the sharp rates by first studying the convergence rate of the discrete Ornstein-Uhlenbeck (O-U) process driven by general noise, …
#arXiv #paper #研究/论文
18. Diagnosing Pathological Chain-of-Thought in Reasoning Models
分类：研究/论文来源：arxiv_search分数：80作者：Manqing Liu时间：2026-02-14T21:53:47Z
Chain-of-thought (CoT) reasoning is fundamental to modern LLM architectures and represents a critical intervention point for AI safety. However, CoT reasoning …
- Chain-of-thought (CoT) reasoning is fundamental to modern LLM architectures and represents a critical intervention point for AI safety
- However, CoT reasoning may exhibit failure modes that we note as pathologies, which prevent it from being useful for monitoring
- Prior work has identified three distinct pathologies: post-hoc rationalization, where models generate plausible explanations backwards from…
- To better understand and discriminate between these pathologies, we create a set of concrete metrics that are simple to implement, computat…
- To validate our approach, we develop model organisms deliberately trained to exhibit specific CoT pathologies
- Our work provides a practical toolkit for assessing CoT pathologies, with direct implications for training-time monitoring
#arXiv #paper #研究/论文
19. RPGD: RANSAC-P3P Gradient Descent for Extrinsic Calibration in 3D Human Pose Estimation
分类：研究/论文来源：arxiv_search分数：78作者：Zhanyu Tuo时间：2026-02-14T21:49:51Z
In this paper, we propose RPGD (RANSAC-P3P Gradient Descent), a human-pose-driven extrinsic calibration framework that robustly aligns MoCap-based 3D skeletal …
- In this paper, we propose RPGD (RANSAC-P3P Gradient Descent), a human-pose-driven extrinsic calibration framework that robustly aligns MoCa…
- RPGD formulates extrinsic calibration as a coarse-to-fine problem tailored to human poses, combining the global robustness of RANSAC-P3P wi…
- We evaluate RPGD on three large-scale public 3D HPE datasets as well as on a self-collected in-the-wild dataset
- Experimental results demonstrate that RPGD consistently recovers extrinsic parameters with accuracy comparable to the provided ground truth…
- These results indicate that RPGD provides a practical and automatic solution for reliable extrinsic calibration of large-scale 3D HPE datas…
#arXiv #paper #研究/论文
20. GSRM: Generative Speech Reward Model for Speech RLHF
分类：研究/论文来源：arxiv_search分数：75作者：Maohao Shen时间：2026-02-14T21:22:55Z
Recent advances in speech language models, such as GPT-4o Voice Mode and Gemini Live, have demonstrated promising speech generation capabilities. Nevertheless,…
- Recent advances in speech language models, such as GPT-4o Voice Mode and Gemini Live, have demonstrated promising speech generation capabil…
- Nevertheless, the aesthetic naturalness of the synthesized audio still lags behind that of human speech
- Enhancing generation quality requires a reliable evaluator of speech naturalness
- However, existing naturalness evaluators typically regress raw audio to scalar scores, offering limited interpretability of the evaluation …
- Inspired by recent advances in generative reward modeling, we propose the Generative Speech Reward Model (GSRM), a reasoning-centric reward…
- The GSRM is trained to decompose speech naturalness evaluation into an interpretable acoustic feature extraction stage followed by feature-…
#arXiv #paper #研究/论文