Day 0 NeurIPS 2019 Trip @ Vancouver Convention Center, Canada
To read about my Day -2 NeurIPS 2019 Trip https://medium.com/@ramji.b/day-1-neurips-2019-trip-vancouver-canada-27e02830ffcc
The D day started today at 6.30 AM. I got ready and left the hotel towards Vancouver convention center and wanted to grab breakfast at JJ bean Coffee roasters, but the one on Burrard street was closed on Sunday. So went to Giovane cafe and had a Latte and Butter croissant.
Reached convention center around 8 am and stood in a queue for registration, In about 30 mins I finished the registration, got my name tag, along with a beautiful mug with a small booklet of EXPO and conference schedule.
I went to a morning session in Westside ballroom (A + B) and sat for a talk from Intel AI about “Domain Specific Architectures for Inference & Training”. The speaker was talking about hardware related topics which I was not interested.
Left the hall after a while and went to the second level where demonstration by Lee of BioMind’s team was being held. He did his best to explain the project he is working on. In which the team is trying to segment the CT scan of a cancer patient using the U-Net version model and predicting the drug dosage on every pixel of the tumour.
At 11 am, I listened to one of the best talks of today from Beem Kim, Google AI on “Interpretability - Now What?”. In her talk she reflected on where as we going in the field, and what are the things we need to be aware and be careful as we make progress. With that perspective she was discussing a Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV) paper. This was followed by talks on “Machine Learning for Perception at Cruise” by Cruise and “Federated Learning in Healthcare” by doc.ai.
There was a break for lunch. I went downstairs and had Falafel Plate from Anatolia express and moved towards the Eastern side of the Convention center.
In Eastern side of the convention center, I just roamed around for a bit and checked out the sponsors. One of the interesting companies there was Benevolent AI, who is working on revolutionizing drug discovery using AI technologies. I was discussing with Felix, Machine Learning Engineer from BenevolentAI about the team’s work on drug discovery using multi-model concepts. Thanks to BenevolentAI & Team for patiently answering my questions.
Then, attended a workshop on “AI & Machine Learning and Finance: Applications and Opportunities” in the Western side of the convention center and got few inputs from JP Morgan team and Hudson river trading team about the nuances of using the data, process of building models and many more.
Around 6 PM I started to walk back to my hotel. Over the dinner from the same Brooklyn pizza which I had on Day -4, I started to reflect back on the forceful first day talks I attended and conversations I developed.
