基本信息
views: 311
![](https://originalfileserver.aminer.cn/sys/aminer/icon/show-trajectory.png)
Bio
My main research focus is to design and execute rigorous experiments to understand the solutions found by deep neural networks and most critically, their bottlenecks, so that we can intelligently design machine learning systems. I am also interested in developing methods to measure and induce abstract representations in neural networks.
Most recently, my work has focused on understanding properties of data and how these properties lead to desirable and useful representations. I have worked on a variety of topics, including self-supervised learning, the lottery ticket hypothesis, the mechanisms underlying common regularizers, and the properties predictive of generalization, as well as methods to compare representations across networks, the role of single units in computation, and on strategies to induce and measure abstraction in neural network representations.
Most recently, my work has focused on understanding properties of data and how these properties lead to desirable and useful representations. I have worked on a variety of topics, including self-supervised learning, the lottery ticket hypothesis, the mechanisms underlying common regularizers, and the properties predictive of generalization, as well as methods to compare representations across networks, the role of single units in computation, and on strategies to induce and measure abstraction in neural network representations.
Research Interests
Papers共 66 篇Author StatisticsCo-AuthorSimilar Experts
By YearBy Citation主题筛选期刊级别筛选合作者筛选合作机构筛选
时间
引用量
主题
期刊级别
合作者
合作机构
arxiv(2024)
Cited0Views0Bibtex
0
0
NeurIPS (2023)
CoRR (2023)
CoRR (2023)
Load More
Author Statistics
Co-Author
Co-Institution
D-Core
- 合作者
- 学生
- 导师
Data Disclaimer
The page data are from open Internet sources, cooperative publishers and automatic analysis results through AI technology. We do not make any commitments and guarantees for the validity, accuracy, correctness, reliability, completeness and timeliness of the page data. If you have any questions, please contact us by email: report@aminer.cn