#datasets

73 stories · 7 in 7 days · 18 in 30 days
Stories, products and papers tagged "datasets", newest first.
Nov 13
PapersBuildersMulti-source12:01
Chinese contentGHAgentCI 数据集发布:从 GitHub Actions 历史中挖掘编码智能体调用记录

搞智能体工程的朋友可以看看:研究团队从 6K+ 仓库里挖出 15K+ 条 CI 工作流历史,专门追踪编码智能体在 GitHub Actions 里的用法,抽取工具 cai-miner 开源,准确率验证到 96% 以上。想做智能体与 CI 集成研究或评估的项目直接用得上。

Event
arXiv cs.AI@Daniel Ogenrwot et al. (4)Source
ModelsBuildersSelected05:54
LMArena open-sources two web development evaluation datasets

Preference data of this scale for web app generation is scarce, and the pairing of subjective preference labels with verified functional requirements gives training and evaluation teams two complementary signal types. Model developers working on coding agents can benchmark against human judgment rather than proxy metrics. Comparison with existing web-dev benchmarks becomes straightforward since both datasets and the accompanying paper are public.

lmarena.ai@lmarena_aiSource
Nov 11
Chinese content教育知识图谱驱动的大规模语料 WoW 发布,覆盖 LLM 全生命周期训练

中文教育语料老是被吐槽碎、乱、缺体系,这批人用知识图谱当骨架重建了一遍,还开源了检索平台。做中文模型训练或垂直领域数据构建的可以看看他们的组织思路,比自己乱爬数据靠谱。

arXiv cs.AI@Tianze Sun et al. (3)Source
Nov 10
Chinese contentLUCQUE:魁北克肺癌筛查数据集,支持可靠的 AI 评测

搞肺癌筛查 AI 的朋友可以看看:魁北克团队放出了 1,678 例 LDCT 扫描加专家标注的公开数据集,标注量比现有公开数据集都大。亮点是 model-to-data 访问机制,数据不出门,代码上门跑,做模型对比评测方便多了。

arXiv cs.AI@Johanna Topalis et al. (10)Source
Chinese contentChannels4World:面向无线基础模型的环境级数据集发布

做无线通信和信道建模的人可以看看这个数据集。10万个室外射频场景、29TB 数据,覆盖 3.5 和 7 GHz 两个频段,专门为训练无线基础模型设计。论文还验证了场景多样性对泛化的作用,想做跨部署环境泛化的团队用得上。

arXiv cs.LG@Joonkyu Jang et al. (7)Source
Nov 9
Nov 4
PapersBuildersSelected22:19
MIT 6.S950 Course Notes Cover Outsourced Data Risks in AI Agents

The non-obvious point is that when an AI agent scrapes or filters data, it silently makes coverage decisions that would normally require a human's judgement, so inherited bias compounds invisibly. Practitioners building RAG pipelines or training data collections in Southeast Asia can use the four-question checklist — population, prediction target, deployment shift, record definition — to audit datasets before use. The C4 versus FineWeb-Edu comparison shows swapping filters only changes what gets excluded, not exclusion itself.

Event
shao__meng@shao__mengSource
Nov 3
Oct 30
Oct 26
Oct 23
Oct 12
Oct 9
Oct 3
Sep 28
ModelsBuildersSelected73°23:58
Monsoon speech dataset cuts Telugu WER from 92.7% to 16.1%

The takeaway is that low-resource ASR quality can be bought with better data curation rather than bigger models, since a mid-size Whisper checkpoint beat Gemini 3.1 Pro on Telugu purely through fine-tuning. Teams building speech products for Indic and other under-served language markets can use this dataset-first approach instead of waiting for larger architectures. Keeping noisy audio in training, with DNSMOS-based stratification, is the transferable technique.

rohanpaul_ai@rohanpaul_aiSource
Sep 23
Sep 18
Sep 15
Sep 1
Aug 25
Aug 22
Aug 19
Aug 18
Aug 12
Aug 6
Jul 23
Jul 15
Jul 8