Towards Frequency-Based Explanation for Robust CNN | AISC

Channel:
Canada LLMs Explained - Aggregate Intellect - AI.SCIENCE
Subscribers:
22,300
Published on ● Video Link: https://www.youtube.com/watch?v=dHo9QUm85Ms


Duration: 41:31
342 views
17

Speaker(s): Zifan Wang

Find the recording, slides, and more info at https://ai.science/e/towards-frequency-based-explanation-for-robust-cnn--hBiCzxIOpi3iOm0woCyY

Motivation / Abstract
Computer Vision implementations based on Convolutional Neural Networks (CNN) have been reported to be unstable and vulnerable to adversarial attacks that are not visible to humans. The authors are presenting novel work to better identify algorithms that are vulnerable to these attacks and present explainable options to help design better models.

What was discussed?
1) Is your method helping models ignore the noise in a similar way that human cognition filters out irrelevant information?
2) Why is adversarial training expensive? How does your method improve on adversarial training?
3) Do you need access to the model to generate your explanations?



------
#AISC hosts 3-5 live sessions like this on various AI research, engineering, and product topics every week! Visit https://ai.science for more details



Other Videos By LLMs Explained - Aggregate Intellect - AI.SCIENCE


2020-07-29Quantum Technologies: State of Play | AISC
2020-07-29NLP on Noisy User-generated text - NER for StackOverflow | AISC
2020-07-28Overview of Machine Learning in Marketing | AISC
2020-07-24Cognitive Model Priors for Predicting Human Decisions | AISC
2020-07-22Machine Learning to Assess Trends and Alignment of Funded Research Output | AISC
2020-07-22COVID and Racial Inequity, and Implications for AI
2020-07-21TGN: Temporal Graph Networks for Deep Learning on Dynamic Graphs [Paper Explained by the Author]
2020-07-20Founders in Fundraising, and AI Applications
2020-07-16Uncertainty-Aware Action Advising for Deep Reinforcement Learning Agents | AISC
2020-07-16Machine Learning for Forecasting Global Atmospheric Models | AISC
2020-07-15Towards Frequency-Based Explanation for Robust CNN | AISC
2020-07-14Lagrangian Neural Networks | AISC
2020-07-14TeslaSenti: Near real-time sentiment analysis of Tesla tweets | Workshop Capstone
2020-07-14See.Know.Bias - Using AI to Develop Media Literacy and Keep News Neutral | workshop capstone
2020-07-10Video Action Transformer Network | AISC
2020-07-09DLIME - Let's dig into the code (model explainability stream)
2020-07-09Overview of Machine Learning in Behavioral Economics | AISC
2020-07-08Compact Neural Representation Using Attentive Network Pruning | AISC
2020-07-08Navigating the Idea Maze: Continuous discovery frameworks for (AI?) products | AISC
2020-07-02Nvidia's RAPIDS.ai: Massively Accelerated Modern Data-Science | AISC
2020-07-02Building AI Products; The Journey | Overview