Explore the limitations of machine learning systems through a unique steampunk perspective in this 25-minute conference talk from BSidesSF 2018. Delve into the world of neural networks and their fundamental constraints as Thomas Phillips presents an innovative approach to understanding machine learning. Learn how to conceptualize a physical implementation of a machine learning system using tubes, valves, and gears, moving beyond traditional mathematical explanations. Discover the essential elements of neural nets and gain insights into why these systems generate false positives. Gain a fresh understanding of terms like deep learning and neural nets, often touted as magical solutions to security problems, and uncover their inherent limitations in a captivating and accessible manner.
Overview
Syllabus
BSidesSF 2018 - Machine Learning: Too Smart for its Own Good (Thomas Phillips)
Taught by
Security BSides San Francisco
Related Courses
-
Getting Started with Machine Learning at the Edge on Arm
-
Practical Threat Hunting With Machine Learning
-
Machine Duping - Pwning Deep Learning Systems
-
Using Machine Learning to Detect Sensitive Documents on SharePoint
-
Secure Because Math - A Deep-Dive on Machine Learning-Based Monitoring
-
ATT&CKing the Status Quo - Improving Threat Intel and Cyber Defense with MITRE ATT&CK
