“If we make them smaller, we can make them faster” has been the approach to building faster computers over the last 70-years starting with the first transistors and continuing as we built increasingly more complex integrated circuits. This was known as “Moore’s Law”. Recently, our technology is no longer achieving higher densities as we scale to smaller features sizes first with NAND Flash and now with SRAM and DRAM. New techniques known as Beyond Moore are required to continue to build faster and more complex computers. David Bondurant reviews the emergence of 3D memory and 3D packaging as today’s approach to building the fastest Supercomputers and AI Processors. Virtual: https://events.vtools.ieee.org/m/455111

Humans make decisions and solve problems using heuristics (“thinking fast”) or abstract approaches such as modeling (“thinking slow”). Artificial intelligence approaches can similarly use either heuristics that are related to correlation and categorization, or use models that are related to causation. Predictive Engineering, which melds engineering modeling with probabilistic thinking, aligns closely with causation and an aspect of artificial intelligence called Causal Learning. Issues with some artificial intelligence approaches will be explored, with real (and sometimes controversial and provocative) examples, and promising approaches encompassing causation /predictive engineering will be discussed. [] Speaker(s): Eric Maass Virtual: https://events.vtools.ieee.org/m/445270