SK Reddy

Abstract: There are so many open sourced and other AI frameworks available. Tensorflow, Caffe, Torch, PyTorch, Caffe2, MXNet, Spark MLlib, CNTK, H2O, DMTK, Theano, Scikit, etc. Each of these frameworks have certain strengths. Some of tem work with certain languages and some do not. In addition, some of these frameworks have been used by popular authors of tech papers and have open sourced the code. Some of these are free and some are not. Some frameworks are better to use for image processing use cases and some are not. Organizations and individuals are confused on which framework is best for what type of use case and which framework is better. I would like to discuss and compare various frameworks. Also I would like to pick a few use cases that will be implemented as a hands-on exercise using a few frameworks.

With global commerce on the rise, collaboration between individuals speaking different languages is essential. The need for real time and accurate Language translation using deep networks is picking up fast. Historically, language translation was done using rules and examples. This is obviously not scalable and does not cater to the evolution of languages. Later word-based and phrase-based translations were invented that still could not address the need of the hour. The modern day translations are done using deep networks. I would like to discuss the topic of language translation and deep learning using Tensorflow. The attendees will get a chance to work hands-on to create the building blocks of Language Translator using Tensorflow.

Use cases in Patient records, Contract Analyses, Patent Evaluations, etc. can be done using NLP. The above could be just a 40-60min speech. Or I could do a workshop to discuss a couple of use-cases and do hands-on exercises.