Hi! I'm Menglin Jia (贾梦霖 in Chinese, KMnP in chemical elements --"jia meng lin" is the mandarin pronunciation of those chemical elements), a research scientist at Meta. I obtained my Ph.D. in Computer and Information Science at Cornell University, advised by Serge Belongie, Claire Cardie, and Bharath Hariharan.

My research interests include Generative AI and Multimodal.

Selected Publications

sphere teaser

Image Generation with a Sphere Encoder

2026

Kaiyu Yue*, Menglin Jia*, Ji Hou, Tom Goldstein (*Equal contribution)

arXiv Project Page
DIFT teaser

Emergent Correspondence from Image Diffusion

NeurIPS 2023

Luming Tang*, Menglin Jia*, Qianqian Wang*, Cheng Perng Phoo, Bharath Hariharan (*Equal contribution)

arXiv Project Page
VPT teaser

Visual Prompt Tuning

ECCV 2022

Menglin Jia*, Luming Tang*, Bor-Chun Chen, Claire Cardie, Serge Belongie, Bharath Hariharan, Ser-Nam Lim (*Equal contribution)

arXiv Code Video
VISE teaser

Exploring Visual Engagement Signals for Representation Learning

ICCV 2021

Menglin Jia*, Zuxuan Wu*, Austin Reiter, Claire Cardie, Serge Belongie, Ser-Nam Lim (*Equal contribution)

Special thanks to Marseille who is featured in the teaser photo!

arXiv Project Page Video
Intentonomy teaser

Intentonomy: a Dataset and Study towards Human Intent Understanding

CVPR 2021 (oral)

Menglin Jia, Zuxuan Wu, Austin Reiter, Claire Cardie, Serge Belongie, Ser-Nam Lim

arXiv Project Page Video
Fashionpedia teaser

Fashionpedia: Ontology, Segmentation, and an Attribute Localization Dataset

ECCV 2020 (oral)

Menglin Jia*, Mengyun Shi*, Mikhail Sirotenko*, Yin Cui*, Claire Cardie, Bharath Hariharan, Hartwig Adam, Serge Belongie (*Equal contribution)

arXiv Project Data API Model Video Challenge
Class Balanced Loss teaser

Class-Balanced Loss Based on Effective Number of Samples

CVPR 2019

Yin Cui, Menglin Jia, Tsung-Yi Lin, Yang Song, Serge Belongie

arXiv Code Slide

Activities