Deep-Thinking before Deep-LearningRichard Stebbing
|Thursday 26 November 2015|
|It has never been easier than now to apply ML algorithms to new problems and domains. Implementations of the most powerful algorithms can be accessed from many of the most popular high-level programming languages, and coupled with the availability of cloud computing, deployment couldn't be simpler. However, while the implementation barrier for selecting and pipelining different algorithms has been reduced, the consequences and assumptions of the underlying models often go unstated or are outright neglected.
The trajectory of this talk is to go (in depth) through some examples I've encountered most recently where the distinction between models and algorithms has been done well, and where it hasn't. With this theme, I'll also go through some of the practical technical and non-technical lessons that I've learned since my (short) transition between academia and industry.
Bio: My name is Dr Richard Stebbing and I'm a Senior Engineer and AI Lead at a San Francisco-based AI/ML startup. I finished my PhD in Engineering Science at Oxford University last year, and during this time, I also worked for seven months at Microsoft Research (Cambridge, UK) in the Machine Learning and Computer Vision Group. Before this, I completed my BE (Hons) in Electrical and Electronic Engineering at the University of Auckland.