Bernard Traversat, Vice President of Software Development, Java SE Platform , Oracle Java and the JVM are the leading big data runtime within the enterprise. The gigantic demands of new machine learning, AI, and big data workloads are pushing the limit of scale and performance beyond what current leading ML/AI platforms can scale to. Java and the JVM have historically been the runtime of choice for developers looking at massive scale, robustness, and performance. This session discusses several new innovations in the Java language and JVM to address new machine learning and AI workloads at scale (JIT compilation, data layout, data access, GC latency, off-heap management, native OS interactions, crypto support and startup) and presents upcoming improvements in the platform. It also covers the leading Java and JVM-based ML and AI libraries.
We’ve seen fully automated bot beats us in Go, one-on-one Poker and Dota II, now what’s going to happen for trading financial markets? Listen to A.I. Capital Management sharing their research, a Deep Learning trading agent that over-performs us in trading FX markets. Marshall has been trading FX markets for 5 years. As a Master in Finance graduate from Brandeis International Business school, he combines his insight in financial markets with a passion for machine learning and expertise in programming, striving to build the first game-changing A.I. trading system to disrupt the markets. He has a strong passion for quantitative trading and machine learning and started AI Capital Management in September 2016. His inspiration came from Google DeepMind’s AlphaGo project, which is a Deep Learning agent that beats human Go world champions. Go, as a game with complexity at a number more than the atom in the universe, is arguably as hard as or even harder than trading financial markets, which they believe is the next game to be solved with Artificial Intelligence. This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at https://www.ted.com/tedx