Lingjuan Lyu

Timeline

1. 2011–2017

   Robots, chips and a Melbourne PhD

   Her research begins with humanoid dancing robots and robust image watermarking, and she publishes her first paper on IP protection in 2011. Secure chip design research in her master's program deepens that focus, and she graduates from the University of Chinese Academy of Sciences before her PhD. She completes a PhD at the University of Melbourne, where her work on privacy and robustness in machine learning begins in 2015, and IBM names her a PhD Fellow in 2017.

2. By 2021

   Fellowships and Ant Group

   Before Sony AI, she is a Research Fellow at the Australian National University at Level B3, the level of a lecturer or assistant professor. She also serves as a Research Fellow at the National University of Singapore and as an expert researcher at Ant Group, where she leads federated learning. Her paper wins the best paper award at FL-IJCAI in 2020.

3. 2021

   Building Sony AI's privacy team

   In 2021, she is a senior research scientist and team leader at Sony AI, where she builds the Privacy-Preserving Machine Learning team. The team works toward trustworthy sensing platforms for imaging and sensing use cases, powered by federated learning, privacy-preserving computer vision and on-device AI.

4. 2022

   Privacy for Free at ICML

   With Tian Dong and Bo Zhao, she co-authors "Privacy for Free: How does Dataset Condensation Help Privacy?", which receives an Outstanding Paper Award at ICML in 2022. The paper shows that dataset condensation, designed to make training more efficient, also provides a route to private data generation.

5. 2023

   IJCAI Early Career Spotlight

   In 2023, she receives an IJCAI Early Career Spotlight, and her team widens its focus to low-cost, high-impact vision foundation models and visual generative AI. Federated learning enters the vision foundation model development pipeline to reduce risk and enhance privacy.

6. 2024

   TR35 and AI's 10 to Watch

   MIT Technology Review names her to its Innovators Under 35 list for China in 2024, for developing Sony's first vision-centric federated learning platform. IEEE Intelligent Systems also counts her among its 2024 recipients of AI's 10 to Watch, an award for 10 rising stars in AI.

7. 2025

   Argus at CVPR

   Her team's Argus, a vision foundation model of about 100 million parameters that supports 17 vision tasks, appears at CVPR in 2025. The paper "Stretching Each Dollar: Diffusion Training from Scratch on a Micro-Budget" also appears there, and its diffusion transformer trains from scratch for $1,890.

8. As of October 2026

   Head of Sony AI's privacy team

   She is the Head of the Privacy-Preserving Machine Learning team at Sony AI, where she leads scientists and engineers on privacy and security initiatives across the company. Her research interests include federated learning, responsible foundation model development, data privacy, model robustness, IP protection and on-device AI.