Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Discover a revolutionary approach to machine learning that replaces complex arithmetic with simple Boolean logic in this 20-minute conference talk. Learn how the Tsetlin Machine, based on forgotten Soviet mathematical work from the 1960s, creates interpretable AI models using only AND, OR, and NOT gates without any multiplication operations. Explore the technical process of Booleanization that converts input data into binary questions for learning automata, and understand how these finite state machines work in parallel to create logical patterns for decision-making. Examine the natural sparsity of resulting models where over 99% of potential features are automatically excluded for complex tasks like image recognition, and see how further optimization through removing "weak includes" achieves remarkable efficiency improvements. Analyze performance benchmarks showing 10x faster inference times than Binarized Neural Networks, dramatically reduced memory footprints, and 20x energy efficiency improvements on embedded platforms. Review the latest microchip implementation that consumes just 8 nanojoules per frame for MNIST character recognition, representing potentially the lowest energy consumption ever published for this benchmark. Consider the implications for edge computing and IoT applications where power constraints are critical, and understand how this "white box" approach addresses the growing demand for explainable AI in an era of tightening regulations around artificial intelligence systems.
