Insight x Fastai AI SF Meetup -- March 18th

We are excited to share with you our March monthly AI Meetup supported by Insight and fast.ai. The meetup provides an opportunity for Insight and fast.ai fellows, as well as other like-minded AI communities and professionals to come together to share and present their work, take part in engaging workshops, collaborate on projects, and network.

As we strive to make the monthly Meetups diverse and appealing to everyone, we bring you another unique Meetup for this month. We start by two inspiring presentations by Insight and fastai fellows, followed by a hands-on Deep Learning Workshop on Hyperparameter Sweeps by Weights & Biases. The last hour is project collaborations and networking.

Meetup Details:

Date & Time: Wednesday, March 18th, 6p.m - 8:30p.m

Host: Insight Data Science, 500 3rd St, San Francisco, CA

• 6 - 6:30 p.m - Refreshments and pizza

• 6:30 p.m - Doors close!

• 6:30 - 6:40 p.m - Announcements

• 6:40 - 7:15 p.m - Presentations by Dr. Zigfried Hampel and Hao He (details below)
  

• 7:15 - 8 p.m - Hands-on Deep Learning Workshop by Weights & Biases (details below)

• 8 - 8:30 p.m - Project collaborations & Networking

Meetup mission, vision, and structure contact can be found HERE.

Meetup expects attendees to have some familiarity with Deep Learning.

(We recommend watching first couple Fast.ai Practical Deep Learning for Coders - Part 1 lessons prior to the event if you are not).

Presentations:

The first presentation is by Dr. Zigfried Hampel, an Insight AI Fellow and a Data Scientist at In-Q-Tel. Dr. Hampel will discuss his recent work with GANs (generative adversarial networks). He will address a fundamental question in generative models: What artificial fingerprints do GANs leave behind? Dr. Hampel will share how learning data with strict statistical properties can directly allow probing of such fingerprints; hence, providing insights into the capabilities and limitations of generative models.

The second presentation is by Hao He, an Insight & Fastai fellow and co-founder of Global Ocean Information System Inc. Hao will share his deep learning journey, from a fast.ai novice DL practitioner to winning his first Kaggle silver medal. He will share insights for exploring (EDA) and preprocessing Kaggle data, model architecture search, and ensemble modeling. He will also share his experience applying some of the Fastai tricks in the APTOS competition and his results.

Workshop:

This month’s workshop is a super hands-on where you get to build a vision emotion classifier with W&B hyperparameter sweeps. Attendees will break up into groups and take a machine learning model and iteratively improve the accuracy. Folks unfamiliar with vision models will get guidance through the the process and more advanced attendees will have a chance to experiment with more advanced models. The workshop is conducted by Weights & Biases founder and CEO, Lukas Biewald.

The last hour is for project collaborations, networking, and catching up with cohort fellows. Attendees are welcome to stay until 9p.m (hard stop!)

Space is limited so please RSVP in advance to attend.

(If on waitlist, you'll be contacted as soon as a spot opens up)