AstriCon 2017

Within the last year we've seen a proliferation of services spring up (IBM Watson, Google Speech API, Amazon Lex) that help make machine learning accessible to a far wider audience. At the same time, open-source frameworks have also been evolving to fill these roles at a lightening pace -- and with fantastic results. This talk will dive into this ecosystem of ASR, TTS, and NLU frameworks, including DeepSpeech and WaveNet, and highlight their practicality in the real world. Knowledge of machine learning isn't required to understand the concepts and practical applications. Demonstrations of these frameworks in action can help illuminate the possibilities and challenges of utilizing the DIY approach. Many companies would benefit from having a tailored solution to even just one of the above -- perhaps you want your own custom TTS voice, or need sentiment analysis on stored voicemails or customer service calls. Pick and choose and make your machine learning dreams come true.