MobilityGPT: Enhanced Human Mobility Modeling with a GPT model

no code implementations • 5 Feb 2024 • Ammar Haydari, Dongjie Chen, Zhengfeng Lai, Chen-Nee Chuah

Then, we constrained the training process via a road connectivity matrix that provides the connectivity of sequences in trajectory generation, thereby keeping generated trajectories in geospatial limits.

VeCLIP: Improving CLIP Training via Visual-enriched Captions

1 code implementation • 11 Oct 2023 • Zhengfeng Lai, Haotian Zhang, BoWen Zhang, Wentao Wu, Haoping Bai, Aleksei Timofeev, Xianzhi Du, Zhe Gan, Jiulong Shan, Chen-Nee Chuah, Yinfei Yang, Meng Cao

For example, VeCLIP achieves up to +25. 2% gain in COCO and Flickr30k retrieval tasks under the 12M setting.

Retrieval Text Retrieval +1

PADCLIP: Pseudo-labeling with Adaptive Debiasing in CLIP for Unsupervised Domain Adaptation

no code implementations • ICCV 2023 • Zhengfeng Lai, Noranart Vesdapunt, Ning Zhou, Jun Wu, Cong Phuoc Huynh, Xuelu Li, Kah Kuen Fu, Chen-Nee Chuah

We then utilize CLIP's zero-shot prediction to formulate a Pseudo-labeling setting with Adaptive Debiasing in CLIP (PADCLIP) by adjusting causal inference with our momentum and CFM.

Causal Inference Unsupervised Domain Adaptation