Anthill Inside 2018

DL research has progressed at tremendous speed in recent years. From the demo of Natural conversations at google I/O to BMW X3+ driving itself on road, AI is no more only a topic of research or interest. AI is already everywhere. This talk presents advancements in Deep Learning both in research and field from practitioner’s perspective. This talk has two parts. First part demonstrates the advancements in research. The second part deals with improvements in hardware, software and frameworks which is helping DL grow exponentially.

Outline

DL research has progressed at tremendous speed in recent years. From the demo of Natural conversations at google I/O to BMW X3+ driving itself on road, AI is no more only a topic of research or interest. AI is already everywhere. This talk presents advancements in Deep Learning both in research and field from practitioner’s perspective. This talk has two parts. First part demonstrates the advancements in research. The second part deals with improvements in hardware, software and frameworks which is helping DL grow exponentially.

CNNs and LSTMs have been the traditional DL powerhouses. Part one would extend these powerful concepts and walk through the advances in DL models from the lens of researchers. Part two would walk through how teams across the world are experimenting and breaking barriers in terms of reducing training time, improving accuracy or launching AI enabled products in record time.

Part 1:
Wavenet, Inception v4, AutoML, Spiking CNNs, Shortcut CNNs, Energy based optimization algorithms.
Part 2:
2.1) Practical aspects of DL:
a) Precision
b) Distributed Training
c) Transfer Learning
d) Behavioural Cloing
2.2) Identification of fruits/veggies in real world market in less than a week.
2.3) Fraud Detection in minutes from days in real world bank.

Speaker bio

Akhilesh works as Strategic Cloud Engineer with Google in Australia. As part of his daily job, he works on enabling global enterprises adopt ML/AI on Google Cloud technologies. He is interested in scalable machine learning models. He is particulary interested in text processing using deep learning techniques.