AIQC - Framework for Reproducible and Rapid Deep Learning

This month's talk is in collaboration with Boston Data Science Meetup (https://meetup.com/Boston-Data-Science-Meetup/)!

• We will have breakout rooms for small group discussions and networking during this event.
• The Zoom link to this virtual event will be uploaded closer to the event date.
• This talk will be recorded.

AGENDA

7:00pm: Introductions
7:15pm: Breakout Rooms
7:30pm: Talk by Layne Sadler
8:00pm: Q&A
8:15pm: Breakout Rooms
8:30pm: End of Event

TALK SUMMARY

AIQC simplifies data preparation and parameter tuning for batches of deep learning models without an expensive cloud backend. It empowers researchers by reducing the programming and data science know-how required to integrate machine learning into their research. It makes machine learning less of a black box by reproducibly recording experiments in a file-based database that requires no configuration.

The talk will:
(1) Describe the challenges I ran into when implementing best practices that academics encouraged me to adopt after reviewing my paper. Show the gaps this revealed in the mainstream machine learning workflow.
(2) Demo the AIQC Python package on Kepler satellite data and brain tumor MRI scans.

Learn more:

• Read: https://aiqc.readthedocs.io/
• Watch: https://youtube.com/watch?v=cN7d8c-3Vxc&list=PLzDUt2WiohNj7MUrYL3YxoPbXjt5iDEPz

ABOUT THE SPEAKER

An autodidact at heart, Layne Sadler actually started off his journey on the business side of technology. Over time, an innate technical curiosity drew him to learn how to build applications and algorithms alike. This ultimately led to not only founding an API-driven Content Management startup, but also spearheading scientific product development at a unicorn biotech company in the genomics industry. Most recently, and more directly related to PyData, Layne led the 2021 JupyterLab Survey which gathered use cases, pain points, and competitive intel from over one thousand data scientists and researchers in order to help guide Jupyter's development efforts.