PeopleofSuperintelligence

Researcher

Jingkuan Song

Also known as 宋井宽

Jingkuan Song is a computer scientist whose research on multimedia retrieval connects learning to hash, video captioning and multimodal large language models. After a PhD at the University of Queensland in 2014, he held research posts at the University of Trento and Columbia University, then served as a professor at the University of Electronic Science and Technology of China from 2017 to 2024. He coauthored A Survey on Learning to Hash in 2016 and the IJCAI 2017 paper on hLSTMat for video captioning, and he won the ACM China SIGMM Rising Star Award in 2019 and the SIGMM Rising Star Award in 2021. In 2024, he coauthored MMEvol, a framework for evolving multimodal instruction data. He is a professor in the School of Computer Science and Technology at Tongji University and is also affiliated with the Shanghai Innovation Institute.

Timeline

  1. 2014

    A PhD at Queensland

    He earns a PhD in 2014 from the University of Queensland in Australia, advised by Heng Tao Shen.

  2. 2014–2017

    Trento and Columbia posts

    He is a research fellow at the University of Trento from 2014 to 2016, advised by Nicu Sebe. From 2016 to 2017, he is a postdoctoral research scientist at Columbia University.

  3. 2016

    A survey on learning to hash

    In June 2016, he coauthors A Survey on Learning to Hash with Jingdong Wang, Ting Zhang, Nicu Sebe and Heng Tao Shen on arXiv. The survey sorts learning-to-hash methods for nearest neighbor search into pairwise, multiwise and implicit similarity preserving, plus quantization, and reports that quantization performs best on search accuracy, search time and space cost. It is slated to appear in IEEE TPAMI.

  4. 2016–2017

    Best paper awards

    He wins the Best Paper Award at ICPR in 2016 and the Best Student Paper Award at the Australian Database Conference in 2017, and receives a Best Paper Honorable Mention at SIGIR in 2017.

  5. 2017–2024

    Professor at UESTC

    He is a professor at the University of Electronic Science and Technology of China from 2017 to 2024. In 2017, he is first author of Hierarchical LSTM with Adjusted Temporal Attention for Video Captioning at IJCAI, with Lianli Gao, Zhao Guo, Wu Liu, Dongxiang Zhang and Heng Tao Shen. The hLSTMat model decides whether each word relies on visual information or language context, and it outperforms earlier methods on the MSVD and MSR-VTT datasets.

  6. 2019–2021

    SIGMM Rising Star awards

    He wins the ACM China SIGMM Rising Star Award in 2019 and the SIGMM Rising Star Award in 2021, the latter for contributions to multimedia compact representation and analysis. His team takes first place at the CVPR 2021 AI Security Challenge in July 2021.

  7. September 2024

    MMEvol for multimodal models

    In September 2024, he coauthors MMEvol with Run Luo, Haonan Zhang, Yongbin Li and others, a framework that evolves multimodal instruction data starting from the SEED-163K instruction set. Against baselines trained on the initial seed data, it improves average accuracy by 3.1 percentage points across 13 vision-language tasks.

  8. As of October 2026

    Professor at Tongji University

    He is a professor in the School of Computer Science and Technology at Tongji University and is also affiliated with the Shanghai Innovation Institute. His research covers large-scale multimedia retrieval, large language models and deep learning, and he is an associate editor of IEEE TMM and ACM TOMM.

Career

  1. NowTongji UniversityProfessor
  2. NowShanghai Innovation Institute
  3. 2017–2024University of Electronic Science and Technology of ChinaProfessor
  4. 2016–2017Columbia UniversityPostdoctoral research scientist
  5. 2014–2016University of TrentoResearch fellow

Education

  1. 2014University of QueenslandPhD

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Jingkuan Song. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Tongji University. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  3. The company share of AI researchers’ top papers was 31% in 2020–2024 and 8% in 1970–1974Includes Tongji University. We estimate where the people shaping AI worked when their most-cited papers appeared: the company share was 31% in 2020–2024 and 8% in 1970–1974, and Google led all organizations in 2020–2024.
  4. 5 of the 25 people with the most-cited recent papers also lead on career citationsIncludes Shanghai Innovation Institute. We compare citations to 2023–2025 papers with career citations and h-index, after correcting reviewed publication-date errors: who leads, when they started and where they work.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Jingkuan Song yet. Send us a free-licensed photo.

  1. Jingkuan Song cfm.uestc.edu.cn. Checked 21 October 2026.
  2. Homepage of Jingkuan Song, Tongji University jingkuansong.github.io. Checked 21 October 2026.
  3. A Survey on Learning to Hash arxiv.org. Checked 21 October 2026.
  4. Hierarchical LSTM with Adjusted Temporal Attention for Video Captioning ijcai.org. Checked 21 October 2026.
  5. ACM SIGMM Rising Star Award 2021 sigmm.org. Checked 21 October 2026.
  6. MMEvol: Empowering Multimodal Large Language Models with Evol-Instruct arxiv.org. Checked 21 October 2026.
  7. A Survey on Efficient Vision-Language-Action Models arxiv.org. Checked 21 October 2026.