Winning the Market Game with AI

Beating the Market Remains an elusive Dream. Market volatility and economic uncertainty have so far trumped even the most avid investors and brokers in search of a “quick-money” scheme. Deep learning, a new class of AI (Artificial Intelligence) algorithms, takes traditional technical market analysis to the next level by unlocking an unprecedented level of intelligence from voluminous forms of structured and unstructured market data. Time-series based learning, an important branch of AI, is gaining momentum and interest in the financial community as a mean to sift through the space of probabilities and uncover viable alternative investment scenarios that seeks to maximize returns at different market entry points. Still, the most perplexing part of most machine learning doesn’t lie in the algorithms craftsmanship but rather in the art of data representation, the mean by which viable and meaningful features are extracted from continuous market streaming data and fed into the AI machine. In my talk, I present a case study showcasing some of our work with Forex trading focusing on how to engineer signal sampling and pre-processing algorithms to morph historical market data points into discrete events-based string of data suitable for many classical data science algorithms. Through the semantic value of these features market analysts can now translate the insights generated from AI into viable investment strategies that will help better redirect their time and energy from the guessing game towards a profitable instrument that allows them to win with Big Data.