Jun Wang

Jun is a final-year Ph.D. student candidate in Electrical and Computer Engineering at University of Maryland, College Park, where he is honored to be co-advised by Prof. Larry S. Davis and Prof. Joseph F. JaJa. His research primarily focuses on multi-modal learning, object detection, and 3D scene understanding. Currently, he is on the job market.

Recently, Jun has been fortunate to work with Dr. Kishore Prahallad (Apple), Dr. Mingfei Gao and Dr. Ran Xu (Salesforce Research), and Dr. Siheng Chen (Mitsubishi Electric Research Labs). Prior to that, he obtained his M.Sc. degree in Electrical and Computer Engineering from University of Michigan, Ann Arbor in 2017 and B.Eng. degree from Beijing Institute of Technology, China in 2015.

Email: junwang [AT] salesforce [DOT] com

Google Scholar   /   Semantic Scholar   /   DBLP   /   LinkedIn   /   GitHub

profile photo
Selected Publications
Align and Attend: Multimodal Summarization with Dual Contrastive Losses
Bo He, Jun Wang, Jielin Qiu, Trung Bui, Abhinav Shrivastava, Zhaowen Wang
CVPR, 2023
arXiv / code / project / bibtex

Multimodal summarization that summarizes video frames and text sentences with time correspondence.

TAG: Boosting Text-VQA via Text-aware Visual Question-answer Generation
Jun Wang, Mingfei Gao, Yuqian Hu, Ramprasaath R. Selvaraju, Chetan Ramaiah, Ran Xu, Joseph F. JaJa, Larry S. Davis
BMVC, 2022
arXiv / code / poster / bibtex

The first generic text-aware question-answer generation approach for Text-related VQA.

ESSumm: Extractive Speech Summarization from Untranscribed Meeting
Jun Wang
arXiv / code / slides / bibtex

The first automatic speech summarization system with Wav2vec 2.0.

PointMotionNet: Point-Wise Motion Learning for Large-Scale LiDAR Point Clouds Sequences
Jun Wang*, Xiaolong Li*, Alan Sullivan, Lynn Abbott, Siheng Chen
* denotes equal contribution.
WAD, CVPR, 2022
arXiv / bibtex

3D motion learning with a novel point-based spatiotemporal convolution operation module.

M3DETR: Multi-representation, Multi-scale, Mutual-relation 3D Object Detection with Transformers
Jun Wang*, Tianrui Guan*, Shiyi Lan, Rohan Chandra, Zuxuan Wu, Larry S. Davis, Dinesh Manocha
* denotes equal contribution.
WACV, 2022
arXiv / code / slides / bibtex

The multi-representation, multi-scale, mutual-relation 3D object detector with transformers.

InfoFocus: 3D Object Detection for Autonomous Driving with Dynamic Information Modeling
Jun Wang*, Shiyi Lan*, Mingfei Gao, Larry S. Davis
* denotes equal contribution.
ECCV, 2020
arXiv / slides / bibtex

3D Object Detection with the effective dynamic attention module.

Misc Projects
RTL Design of R10K Out-of-Order 2-way Superscalar Processor with Simultaneous Multithreading
Ruobai Feng, Wang Cao, Jun Wang, Yujun Yan, Jiapeng Zhao
EECS 470 Computer Architecture, 2016

The two-way superscalar SMT processor design based on MIPS R10K out- of-order execution architecture.

Design and Layout of a 16-bit RISC Pipelined Processor
Farzad Asgarian, Harsha Chawla, Isaac Jarman, Cody Piekarz, Jun Wang
EECS 427 VLSI Design I, 2015

The baseline processor design with a customized kogge-stone adder based on a 16-bit RISC architecture using IBM’s 130nm CMOS process.

Academic Services
  • Conference Reviewer: CVPR, ICCV, ECCV, ICML, NeurIPS, ICLR, AAAI, BMVC, WACV, ACM MM
  • Journal Reviewer: Neurocomputing, IEEE Transactions on Circuits and Systems for Video Technology
  • Student Volunteer: INTERSPEECH'22
  • [2023] Outstanding Overseas Student Scholarship - Government of China
  • [2022] International Conference Student Support Award - University of Maryland
  • [2022] CVPR Travel Grant
  • [2022] Jacob K. Goldhaber Travel Grant - University of Maryland
  • [2019] Teaching Assistant Training and Development (TATD) Fellow - University of Maryland
  • [2015] Outstanding Graduates - Beijing Institute of Technology
  • [2014] Honorable Mention - Mathematical Contest in Modeling