1
我国首版全国几何基准影像成果发布，定位精度提升 1 倍以上
IT之家 8 月 28 日消息，据自然资源部官方公众号，在 8 月 27 日自然资源部召开的例行新闻发布会上，我国首版全国几何基准影像成果正式发布。该项成果是自然资源部积极服务数字中国建设、推动测绘地理信息事业转型升级的重大标志性成果，也是面向社会服务的测绘地理信息新型公共产品。 IT之家从官方介绍…
2
The Geometry of Low-Resource Language Representations
揭示低资源语言在向量空间中的几何结构，为多语言模型优化提供全新视角。
arXiv:2608.23358v1 Announce Type: new Abstract: The performance gap between low- and high-resource languages in LLMs is widely known, but it remains u…
3
Topology of a Smile: Persistent Homology in Dental Imaging
用拓扑数学捕捉微笑的几何纹理，给牙齿影像分析带来全新视角
arXiv:2608.21422v1 Announce Type: new Abstract: CBCT (Cone Beam Computed Tomography) scans provide detailed three-dimensional images, widely used in d…
4
RefusalGuard: Geometry-Preserving Fine-Tuning for Safety in LLMs
微调会破坏大模型拒答防线？这项几何保持微调技术让安全对齐不缩水。
arXiv:2605.01913v2 Announce Type: replace-cross Abstract: Fine-tuning safety-aligned language models for downstream tasks often leads to substantial d…
5
Encoded but Not Actionable: Auditing the Decode-Generate-Steer Gap in Frozen LLMs for Geometric Constraints
冻结大模型能“看懂”几何约束却不会用，审计揭示解码-生成-引导三环断裂，值得关注。
arXiv:2608.17843v1 Announce Type: cross Abstract: Large language models (LLMs) have demonstrated strong performance on structured reasoning tasks, but…
6
Geometry of Forgetting: Representation Flux in Continual Learning
从几何视角剖析持续学习中的表征漂移，揭示灾难性遗忘的内在机制，为构建不退化的智能模型提供新思路。
arXiv:2608.15854v1 Announce Type: new Abstract: Catastrophic forgetting remains a fundamental obstacle to continual learning, where neural networks lo…
7
Geometric Filtering of LLM-Generated Samples for Few-Shot Text Classification
用几何过滤筛掉低质量LLM生成样本，少样本分类精度立竿见影。
arXiv:2608.13866v1 Announce Type: new Abstract: Large language models (LLMs) can generate synthetic training data for text classification, but the qua…
8
Marionette: Predicting World States, Rendering Geometry, Painting Appearance
一个模型同时预测世界状态、生成几何并赋予外观，开启统一3D生成新范式。
arXiv:2608.14530v1 Announce Type: cross Abstract: Interactive game world models typically autoregress visual observations directly in pixel or latent …
9
GCPO: Diagnosing and Constraining Subspace Geometry in Rollout RL for LLMs
揭秘大模型rollout强化学习中的子空间几何，提出GCPO约束方法，助你理解训练背后的数学本质。
arXiv:2608.11674v1 Announce Type: cross Abstract: On-policy rollout methods such as GRPO are central to post-training of large language models, yet th…
10
几何未来 Eskimo Ultra 42 液冷发售：420 旗舰规格，799 元
420超大冷排旗舰液冷降至799元，大机箱玩家性价比新选。
IT之家 8 月 12 日消息，几何未来 (Geometric Future) 今日宣布其 420 规格处理器一体式 (AIO) 液冷散热器 Eskimo Ultra 42 正式发售，定价 799 元。这一产品配备 3 颗 140mm 风扇，拥有较 360 液冷更大的冷排端换热面积。 Eskimo …
11
Matching Accuracy, Different Geometry: Evolution Strategies vs GRPO in LLM Post-Training
进化策略竟能与GRPO精度持平？揭秘两者在LLM后训练中的不同几何路径。
arXiv:2604.01499v2 Announce Type: replace Abstract: Evolution Strategies (ES) have emerged as a scalable gradient-free alternative to reinforcement le…
12
SG-WAM: Self-Guided World Modeling in Geometry-Aware Policy Space
自监督世界模型与几何感知策略空间结合，让智能体在复杂环境中高效自主决策。
arXiv:2608.01397v1 Announce Type: cross Abstract: World Action Models (WAMs) couple action generation with prediction of future states. Their effectiv…
13
Show HN: GJK Collision Detection in Any Dimension
专为开发者打造的高维碰撞检测库，将GJK算法推广到任意维度，统一单纯形距离计算，助力复杂物理仿真。
Hey HN! I created a library that extends GJK-based collision to higher dimensions. This approach unifies the math for computing the distance to the or…
14
SG-Layout: Structured Scene Graph-Guided Layout Generation with LLMs
利用场景图引导大模型生成空间布局，破解几何关系建模难题，AI布局新思路。
arXiv:2608.01106v1 Announce Type: cross Abstract: Understanding and generating spatially coherent layouts from natural language remains a fundamental …
15
OpenAI Reports Internal Model Disproved an 80-Year-Old Geometry Problem
OpenAI内部模型推翻80年几何难题，还解决10大开放数学问题，AI推理能力迎来新突破。
OpenAI has publicly reported that an internal general-purpose reasoning model autonomously disproved the Erdős unit distance problem , an open questio…
16
A Graph Neural Network approach to zero-shot Digital Twins
零样本数字孪生新框架，用图神经网络突破传统几何刚性，实时视觉与几何无缝耦合
arXiv:2607.20535v1 Announce Type: cross Abstract: Traditional Predictive Digital Twins often remain geometrically rigid, requiring extensive retrainin…
17
Geometric Configurations of Perturbed Jailbreak Prompts
揭秘大模型安全软肋：扰动型越狱提示的几何结构，小模型内部表征泄露攻击机制。
arXiv:2607.20581v1 Announce Type: cross Abstract: Perturbation techniques that turn unsuccessful jailbreak prompts into successful ones are continuous…
18
36氪首发 | 同济博士做几何、物理AI重构设计制造，累计获超3亿元融资
同济博士团队用几何与物理AI颠覆工业设计制造，累计融资超3亿元，硬核科技获资本追捧。
图源/企业 本文约 1400 字，建议阅读 3 分钟 作者丨欧雪 编辑丨袁斯来 硬氪获悉，工业AI设计研发解决方案供应商「设序科技」于近日正式完成B轮超亿元融资，累计获超3亿元融资，投资方包括深产投、合鼎共及老股东涌铧投资等。融资将用于市场开拓（含出海）及核心模型技术研发。 设序科技成立于2020年…
19
Riemannian Deep Learning:Modules, Networks, and Geometries
突破欧式空间局限，探索黎曼流形上的深度学习框架，为几何学习提供新范式。
arXiv:2607.19305v1 Announce Type: cross Abstract: Deep neural networks on manifold-valued representations have attracted growing interest, but many ba…
20
Geometry-Enhanced Portion Estimation for Multimodal LLMs
用几何信息增强多模态大模型的食物份量估算能力，零样本识别大幅提升
arXiv:2607.16514v1 Announce Type: cross Abstract: Image-based dietary assessment promises to replace costly, bias-prone manual recalls, but portion es…