Using Deep Learning to do Continuous Scoring in Practical Applications

Greg Makowski will talk about using Deep Learning to do real-time scoring in practical applications.  He will talk about the current state of the art and what will be done in the future.   The talk is based on both his academic and research background and his considerable experience in the front lines doing real-time analysis for banks and other enterprises.

The talk will cover a brief review of neural network basics and the following types of deep learning:

+ autocorrelational - unsupervised learning for extracting features compounding complexity with additional layers 
+ convolutional - shift invariant detection in speech, IoT or in images like self driving cars
+ real time & continuous systems
+ reinforcement learning or Q Learning, such as learning how to play Atari video games
+ continuous space word models, such as word2vec, skipgram training, NLP understanding and translation