Deep Learning in Mammography; Building Clever Chatbots

This is a past event

222 people went

Location image of event venue

Details

Thanks to Vivek Anand for sponsoring!

For the first time our Meetup will be live-streamed! Head over to http://www.pressnplay.co/streaming/bmlg to join from anywhere. Talk 1: Deep Learning in Mammography
Speaker: Ryan Henderson
Abstract: Radiology is at an inflection point, with computer algorithms approaching and sometimes exceeding the capabilities of trained radiologists. In this talk, we'll look at the particular case of mammographic screening and how we are approaching this problem at Merantix. We'll briefly touch on the societal and business potential of the technology and quickly review relevant deep learning concepts. Then we'll take an in-depth look at the FRCNN model that we trained to detect masses and some of the interesting engineering concerns and results.
Bio: Ryan Henderson holds a PhD in Chemistry from Cornell University, and is currently an engineer at Merantix. He has previously worked at Intel and BigchainDB; a Berlin-based blockchain startup. Talk 2: Building Clever Chatbots with Rasa Core
Speaker: Akela Drissner
Abstract: State machines have been widely used for chatbots, but those only really work for the happy path. I’ll talk about how we use machine learning for dialogue handling to make your bots smarter. While reinforcement learning may seem like an attractive option for this (learning from experience), it’s tricky to define reward functions for this type of problem. Instead, we use a combination of supervised learning and interactive learning.
Bio: Akela has a MSc in AI from the University of Edinburgh. Now she’s a machine learning engineer at Rasa working on conversational AI.