Xia Hu

Software Engineer @ Google DeepMind

About Me

I am currently a software engineer and researcher at Google DeepMind. I live in Los Angeles, California.

Prior to joining Google, I obtained my Ph.D. in Computing Science from Simon Fraser University in 2021, advised by Dr. Jian Pei. During my doctoral studies, I also closely worked and collaborated with Dr. Oliver Schulte. Before pursuing my Ph.D., I worked as a software engineer at Sogou and Baidu. I obtained my B.E. degree in Computer Science from the University of Science and Technology of China in 2013.

Research interests: I am currently working on multimodal LLM, in particular exploring their evolution to address application problems. Besides, I am interested in the foundemantal understanding of deep model structures from mathematical perspective - previously I focused on model complexity and interpretability.

Contact me: amber.hx01@gmail.com

Work Experiences

Software Engineer at Google DeepMind/Research (2022 - present)

Research Intern at Microsoft Research - Asia Lab (2019 - 2020)

Software Engineer at Sogou Inc (2014 - 2016)

Software Engineer at Baidu Inc (2012 - 2014)

Selected Publications

Wu, J., Hu, X., Wang, Y., Pang, B., & Soricut, R. (2024). Omni-SMoLA: Boosting Generalist Multimodal Models with Soft Mixture of Low-rank Experts. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 14205–14215.
```
@inproceedings{Wu_2024_CVPR,
  author = {Wu, Jialin and Hu, Xia and Wang, Yaqing and Pang, Bo and Soricut, Radu},
  title = {Omni-SMoLA: Boosting Generalist Multimodal Models with Soft Mixture of Low-rank Experts},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  month = jun,
  year = {2024},
  pages = {14205-14215},
  file = {mc04.pdf},
  arxiv = {arXiv:2312.00968}
}
```
Hu, X., Chu, L., Pei, J., Liu, W., & Bian, J. (2021). Model complexity of deep learning: A survey. Knowledge and Information Systems, 63, 2585–2619.
```
@inproceedings{hu2021model,
  title = {Model complexity of deep learning: A survey},
  author = {Hu, Xia and Chu, Lingyang and Pei, Jian and Liu, Weiqing and Bian, Jiang},
  journal = {Knowledge and Information Systems},
  volume = {63},
  pages = {2585--2619},
  year = {2021},
  publisher = {Springer},
  file = {mc01.pdf},
  arxiv = {arXiv:2103.05127}
}
```
Hu, X., Liu, W., Bian, J., & Pei, J. (2020). Measuring model complexity of neural networks with curve activation functions. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 1521–1531.
```
@inproceedings{hu2020measuring,
  title = {Measuring model complexity of neural networks with curve activation functions},
  author = {Hu, Xia and Liu, Weiqing and Bian, Jiang and Pei, Jian},
  booktitle = {Proceedings of the 26th ACM SIGKDD International Conference on knowledge discovery \& data mining},
  pages = {1521--1531},
  year = {2020},
  file = {mc02.pdf},
  arxiv = {arXiv:2006.08962},
  video = {https://www.youtube.com/watch?v=Zcl69pS_2lE}
}
```
Chu, L., Hu, X., Hu, J., Wang, L., & Pei, J. (2018). Exact and consistent interpretation for piecewise linear neural networks: A closed form solution. Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, 1244–1253.
```
@inproceedings{chu2018exact,
  title = {Exact and consistent interpretation for piecewise linear neural networks: A closed form solution},
  author = {Chu, Lingyang and Hu, Xia and Hu, Juhua and Wang, Lanjun and Pei, Jian},
  booktitle = {Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery \& Data Mining},
  pages = {1244--1253},
  year = {2018},
  file = {mc03.pdf},
  arxiv = {arXiv:1802.06259}
}
```

Tutorial

Deep Learning Model Complexity: Concepts and Approaches. In SIAM International Conference on Data Mining (SDM’21).
```
@tutorial{2021modelcomplexity,
  title = {Deep Learning Model Complexity: Concepts and Approaches},
  author = {},
  journal = {SIAM International Conference on Data Mining (SDM'21)},
  file = {SDM2021_tutorial_DLcomplexity.pdf},
  website = {sdm21_tutorial}
}
```