Latest

Every published item, in order · 644 items
Nov 3
Nov 2
PapersBuildersOfficialSelected22:49
Apple Details Online Reinforcement Learning Agent for Live iOS UI Navigation

The key shift is moving agent training from offline imitation of scripted trajectories to online learning inside real apps, which targets error recovery — the point where most SFT-trained UI agents break. Researchers and engineers building device-side automation agents can study how the planning/grounding split trades inference cost against robustness. Compared with routing every step through a frontier VLM, this points toward cheaper on-device navigation stacks.

Primary sourceApple ML ResearchSource
Papers18:15
Chinese contentNEJM AI 发表 AI-INFOCARE 随机对照试验:AI 预生成病历摘要对问诊时间与医患沟通的影响

NEJM AI 上一个正经的随机对照试验,测 AI 生成的病历摘要到底能不能让问诊更快、沟通更好,做医疗 AI 的都该看看。

NEJM AI@Elida Benker-Hasani Sven Richter Clara Buszello Sophia Willkommen Markus Georg Prem Mido Hijazi Klaus Daniel Martin Tareq Juratli Ilker Y. Eyüpoglu and Witold H. PolanskiSource
Chinese contentNEJM AI 发布 2025–2026 临床 AI 年度回顾,提炼六大主题

NEJM AI 把 2025 到 2026 年的临床 AI 研究整理成六大主题,做医疗 AI 的人可以直接当年度文献地图用。

NEJM AI@Eileen Kim Marianna V. Elia Michelle M. Li Elizabeth Healey Lauren Erdman Payal Chandak Peniel Argaw Ben Li Kameron C. Black Jayson Marwaha Hugo Morales Anna Taylor Joseph Kvedar Peter Brodeur Brett Beaulieu-Jones Emily Alsentzer Julian C. Hong and Jean FengSource
Tomorrow 08:00, the daily lands in your inboxTelegram channel or email. Same screening as this page, no ads.
Chinese contentAI 听诊模型识别左心室功能不全,多中心前瞻研究完成外部验证

医生用普通听诊器加 AI 就能筛查左心室功能不全,还在多中心做了外部验证,做医疗 AI 的可以看看验证流程怎么设计。

NEJM AI@Eiichiro Sato Yuki Izumi Yuki Saito Yoshihito Saijo Takeshi Kitai Akihiro Hayashida Azusa Murata Tomohiro Kaneko Akira Sakamoto Takahiro Sakamoto Kenya Kusunose Tohru Minamino Hiroshi Seki Koshiro Ikeda Shumpei Saito Shimpei Ogawa and Nobuyuki KagiyamaSource
Nov 1
Oct 31
Oct 30
PapersBuildersOfficialSelected00:23
Meta Deploys JiTTesting LLM Bug-Catching System to Production

JiTTesting shifts test generation from pre-planned suites to on-demand tests written exactly when a risky diff appears, with humans confirming rather than authoring. Engineering teams running large monorepos can study Meta's human-on-the-loop design as a template for adding LLM-generated checks without letting unreviewed code gate merges. The paper gives concrete deployment lessons rather than a lab-only prototype.

Primary sourceMeta Engineering BlogSource
Oct 29
Oct 28