[R][P] AdvFreeze - Adversarial training for robustness without sacrificing clean accuracy

• AdvFreeze

  2 1 2.6 Jupyter Notebook

  Novel training method for Adversarial Robustness in image classifiers

Hi all! Just some background, this is a project that I (along with my partner) worked on for an undergraduate class on Deep Neural Nets. The goal of the project was to design and implement a robust TinyImageNet classifier that would perform well on a hidden test set, despite unknown test-time perturbation. As a part of this project, we were required to research and come up with at least 1 novel idea, implement it, and report our findings. Our idea, which we dubbed "AdvFreeze", is described in here, and we were pretty happy with our findings: https://github.com/roshie548/AdvFreeze/blob/main/report.pdf

• InfluxDB

  www.influxdata.com featured

  Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.

  

Suggest a related project