In this video:
- Romeo Kienzler, Chief Data Scientist, IBM
Romeo leads off with an overall discussion of the Internet-of-Things (IoT), focusing on IBM’s particular approach to that subject. He then demonstrates a game he has created (called “BoomBoomShakeShake” or “Harlem Shake”) where players compete to see who can generate and send to the cloud the most accelerometric energy.
Following the game demo, Romeo discusses machine learning, which he says comes in two varieties: batch and online streaming. In a discussion of neural elements, he encourages the audience to suggest algorithms for modeling software after the behavior of biological nerve systems.
Romeo focuses specifically on the area of “deep learning,” where machine learning contains multiple layers of information. Pursuing the subject further, he introduces the concept of “Long short-term memory (LSTM),” an architecture that specializes in remembering values for either long or short durations of time.
Romeo concludes by relating his BoomBoomShakeShake game to the theoretical information he has provided.
This presentation packs into its 51 minutes a tremendous amount of information, surely to be of great interest to developers involved in the area of machine learning and neural networks.