Interesting applications of Semi-supervised learning and Computer Vision.

-- This is a FREE event--

Bangalore AI/ML Meetup invites you to this special meetup.
Interesting applications of Semi-supervised learning and Computer Vision will be discussed in this event.

Detailed Agenda:

10:15 AM - 10:30 AM: Registration & Welcome / Introduction

10:30 AM –11:00 AM: Talk 1 by Prashant Kumar Rai, Deep Learning Researcher at Clear Image AI, (https://www.linkedin.com/in/prashant83/)
Talk 1 Details:
Introduction to and applications of 3D Computer Vision Techniques
i.> Introduction to 3D Computer Vision
ii.> Reconstruction
iii.> 3D object detection
iv.> Segmentation
v.> Success Stories and Industrial Applications
vi.> Libraries relevant to 3D Computer Vision
vii.> Relevant Resources

11:00 AM –11:45 AM: Talk 2 by Nikhil Prasad Maroli, Founder At CoWork Central. (https://www.linkedin.com/in/mnikhilprasad/)

Talk 2 Details:
"RoadMetrics is an exciting startup that's using an innovative approach to monitoring road conditions across major cities in India. Through the help of data obtained from a simple smartphone fixed on a car's dashboard, they are able to extract image data, and vibration data from the accelerometer and gyroscope.

During this talk Nikhil would be speaking about their computer vision model that utilizes object detection and fuses that data from the accelerometer and gyroscope. Expect to have an exciting time during his talk and broaden your knowledge of Machine Learning and Artificial Intelligence and its applications."

11:45 AM – 12:00 Noon: Tea Break

12:00 Noon: 12:45 PM: Talk 3 by Samiran Roy, Data Scientist at Envestnet | Yodlee,(https://www.linkedin.com/in/samiranroy/)

Talk 3 Details:
Learnings from productionizing a semi-supervised deep learning system model at the petabyte scale.

Supervised deep learning methods require large amounts of labelled data to achieve good perform ance and generalization. However, manually constructing such a data set with is a labor-intensive and time-consuming task. The rate at which we acquire data is often greater than the rate at which we can label them. One solution is Semi-supervised learning(SSL), A family of methods that also make use of unlabelled data for training – typically a small amount of labelled data with a large amount of unlabelled data. There is an abundance published work in the field, but these methods have simplifying assumptions that fail to transfer to practical industry use cases. There is a lack of practical guidelines for deploying effective SSL solutions. This talk is an attempt to bridge that gap where Samiran shares his learnings from successful SSL models deployed in production.

12:45 PM: 1:00 PM: Networking

Heartiest thanks to CoWork Central for hosting / sponsoring this meetup.