Published Jan 31, 2021

#040 - Adversarial Examples (Dr. Nicholas Carlini, Dr. Wieland Brendel, Florian Tramèr)

Explore the intriguing complexities of machine learning with experts Dr. Nicholas Carlini, Dr. Wieland Brendel, and Florian Tramèr as they delve into the challenges of adversarial examples, model robustness, and the intricate balance between performance and privacy in AI systems.

Episode Highlights

Topics covered

Episode Highlights

Related Episodes

#52 - Unadversarial Examples (Hadi Salman, MIT)
Answers 383 questions
#48 Machine Learning Security - Andy Smith
Answers 383 questions
Adversarial Examples and Data Modelling - Andrew Ilyas (MIT)
Answers 383 questions
Nicholas Carlini (Google DeepMind)
Answers 383 questions
Explainability, Reasoning, Priors and GPT-3
Answers 383 questions
047 Interpretable Machine Learning - Christoph Molnar
Answers 383 questions
Kernels!
Answers 383 questions
#032- Simon Kornblith / GoogleAI - SimCLR and Paper Haul!
Answers 383 questions
#65 Prof. PEDRO DOMINGOS [Unplugged]
Answers 383 questions
#86 - Prof. YANN LECUN and Dr. RANDALL BALESTRIERO - SSL, Data Augmentation, Reward isn't enough [NEURIPS2022]
Answers 383 questions
#57 - Prof. Melanie Mitchell - Why AI is harder than we think
Answers 383 questions
Prof. Melanie Mitchell 2.0 - AI Benchmarks are Broken!
Answers 383 questions
#114 - Secrets of Deep Reinforcement Learning (Minqi Jiang)
Answers 383 questions
#106 - Prof. KARL FRISTON 3.0 - Collective Intelligence [Special Edition]
Answers 383 questions
Dr. Paul Lessard - Categorical/Structured Deep Learning
Answers 383 questions

Dexa/Machine Learning Street Talk (MLST)

#040 - Adversarial Examples (Dr. Nicholas Carlini, Dr. Wieland Brendel, Florian Tramèr)

Topics covered

Popular Clips

Biologically Inspired Defenses

Adversarial Examples Discussion

Adversarial Examples Discussion

Biases in CNN Architectures

Machine Learning Security Insights

Mindset in Research

Learning and Privacy

Neural Network Memorization

Security and Neural Networks

Adversarial Examples Discussion

Adversarial Training Insights

Generalizing Adversarial Features

Human vs Machine

Adversarial Examples Insight

Episode Highlights

Research Practices

Learning Challenges

Adversarial Examples

Related Episodes

#52 - Unadversarial Examples (Hadi Salman, MIT)

#48 Machine Learning Security - Andy Smith

Adversarial Examples and Data Modelling - Andrew Ilyas (MIT)

Nicholas Carlini (Google DeepMind)

Explainability, Reasoning, Priors and GPT-3

047 Interpretable Machine Learning - Christoph Molnar

Kernels!

#032- Simon Kornblith / GoogleAI - SimCLR and Paper Haul!

#65 Prof. PEDRO DOMINGOS [Unplugged]

#86 - Prof. YANN LECUN and Dr. RANDALL BALESTRIERO - SSL, Data Augmentation, Reward isn't enough [NEURIPS2022]

#57 - Prof. Melanie Mitchell - Why AI is harder than we think

Prof. Melanie Mitchell 2.0 - AI Benchmarks are Broken!

#114 - Secrets of Deep Reinforcement Learning (Minqi Jiang)

#106 - Prof. KARL FRISTON 3.0 - Collective Intelligence [Special Edition]

Dr. Paul Lessard - Categorical/Structured Deep Learning

#040 - Adversarial Examples (Dr. Nicholas Carlini, Dr. Wieland Brendel, Florian Tramèr)

Topics covered

Popular Clips

Episode Highlights

Research PracticesThe discussion shifts to the challenges of evaluating machine learning models and the incentives that drive research publications. Nicholas Carlini, Florian Tramèr, and Wieland Brendel explore how these factors contribute to errors and hinder progress in the field.

Research Practices

Learning ChallengesThe discussion on memorization versus generalization in machine learning highlights the delicate balance between model performance and privacy. Experts explore how shortcut learning and model robustness challenges impact the effectiveness and reliability of AI systems.

Learning Challenges

Adversarial ExamplesFlorian Tramèr, Wieland Brendel, and Nicholas Carlini explore the perplexing nature of adversarial examples in machine learning. They discuss the challenges of feature generalization, security risks, and the inherent complexity of achieving robust model performance.

Adversarial Examples

Related Episodes