Kai Han

@kaihan_x

Computer Vision || Asst. Prof. @HKUniversity | Ex-{Researcher @GoogleAI, Asst. Prof. @BristolUni, Postdoc @Oxford_VGG, @UniofOxford}

kaihan.org

Joined July 2018

655 Following

1,642 Followers

Replies Reposts Media

Kai Han
@kaihan_x
Sep 22
nice progress
Jonathan Roberts
@JRobertsAI
Sep 22
GPT-6 Astra (max) on ZeroBench: pass@5: 51% (prev. SOTA: 30%) pass^5: 35% (prev. SOTA: 13%) If this reflects a genuine capability improvement, it is a significant step forward in both performance and consistency
Kai Han
@kaihan_x
Aug 5
🚀 CodeBind (#ACL2026): multimodal alignment that keeps modality-unique info. A shared+specific codebook + compositional VQ aligns 9 modalities, and new modalities plug in the same way. 🔗 visual-ai.github.io/codebind/ 📄 arxiv.org/abs/2605.18257
Kai Han
@kaihan_x
Jul 8
🚀 PartCo (#ICML2026): fine-grained category discovery hinges on local parts. It mines part-level correspondence priors from frozen ViT patch tokens—a train-time regularizer that boosts many category discovery methods. 🔗 visual-ai.github.io/partco 📄 arxiv.org/abs/2509.22769
Kai Han
@kaihan_x
Jul 7
🚀 iVGR (#ICML2026): making MLLMs draw boxes at inference can hurt fine-grained visual reasoning. Dual-stream RL + a consistency reward internalizes grounding into pure text—no boxes at test time, yet SOTA on V*/HRBench. 🔗 visual-ai.github.io/ivgr/ 📄 arxiv.org/abs/2605.31096
Kai Han
@kaihan_x
Jun 12
2023: GPT-4 could sketch a partial, static Hong Kong MTR map with some effort. 2026: the full network, live, real-time train positions. 🚀
Jonathan Roberts
@JRobertsAI
Jun 11
Revisiting transport networks after ~3 years In 2023, we evaluated whether GPT-4 could recreate the Hong Kong MTR, and the result felt remarkable at the time Now, Claude Fable 5 can one-shot a live, interactive MTR map with real-time train positions

Kai Han

Kai Han’s posts