AI Singapore X DSSG

For the April meetup, we are very happy and pleased to have speakers from AI Singapore to share their work in data science and related fields.

RSVP Time

Start: 05 April 2018, Thursday, 12:00pm

End: 17 April 2018, Tues, 6:00 PM

Please note of our attendance policy ( https://www.meetup.com/DataScience-SG-Singapore/events/228169542/ ). Attendees are REQUIRED to RSVP to the event to be allowed into the venue. People who are neither in the "going" nor "wait" lists may be turned away.

Please be considerate and update your RSVP if you are not able to make it.

Proposed Schedule & Synopsis

• 1845 - 1900: Networking

• 1900 - 2030:

Speaker: Kong Hai Xun

Title: Deep Learning without a power-hungry GPU
Speaking about deep learning, it almost always be confined in the hardware realm of GPU racks, CUDA computing etc. These are bulk hardware that are costy and power hungry. What if you could run the state-of-the-art neural network with the help of a a USB stick on a raspberry pi and achieves decent run-time speed? Find out how this can be done using Intel NCS compute stick.

Speaker Bio

Hai Xun is an AI Engineer with AI Singapore, a national program to boast AI capabilities in Singapore. He has previous work experience in self driving vehicles, mainly in the area of robot perception system. His main interest is in applied machine learning and computer vision. He also holds a master degree in engineering from NUS.

Speaker: Jeanne Choo
Title : Pipelines, packages and practices for better care and feeding of ML systems

The traditional Python Data Science Stack is incomplete. While areas like data cleaning, data visualisation and model development are addressed by Pandas, Scikit-learn and Matplotlib, other, equally important features like model reproducibility, interpretability, retraining and lifecycle management have received little attention. Lately, however, academic papers tackling these fields have been published, often with accompanying Python packages. In parallel, practitioners are solving these problems with Python solutions that co-opt ideas from Dev-ops and scientific computing. As a result, we can now build workflows that are accountable to end-users and that adapt well to new data. This talk traces the developments in this space, lists the pipelines, packages and practices that have resulted, and ties these ideas together with a final, end-to-end example.

Speaker Bio

I started programming at Code and Cake events. In the beginning I mostly ate cake. Over time the ratio shifted in favour of code. I got involved with R Ladies, spoke at Google Developer Dev Fest and ran workshops for people interested data analysis. In my previous life I was a scientist, studying everything from animal behaviour to blood cancer to the predator starfish in our oceans. I believe the key word in data science is science, and enjoy uncovering the ways science and computing inform each other.

For direction to venue: https://goo.gl/WFGj9Q

RSVP Time

Start: 05 April 2018, Thursday, 12:00pm

End: 17 April 2018, Tues, 6:00 PM