Session Theme

Live Webinar - Scalable Active Learning for Image Classification and Detection

Computer Vision is being leveraged everywhere today; be it Healthcare, Infrastructure, Transportation, Security, or Manufacturing. Use of Computer Vision is going to increase exponentially. This trajectory will include examining millions, or even hundreds of millions, of images, depending on how robust or how complex one wants their model to be.

While collecting large amounts of unlabelled data is already a major challenge, only a small subset of it can be labelled by humans due to the effort needed for high-quality annotation. Human effort also adds errors and inconsistency in labelling. Labelling is one of the greatest stumbling blocks to the wide adoption of Computer Vision in Enterprises.

Active Learning seems to be one of the answers to this conundrum. How can we obtain high-quality annotation with minimal supervision? What role does active learning play in this whole mechanism? How can we make it more scalable? What are the challenges in this approach?

Let us explore the role of active learning in annotation…

Distinguished Speakers

Sirisha Peyyeti
Head – Consulting
Srinivas Atreya
Chief Data Scientist
Ramyasree Devarakonda
Sr. Consultant
Dheeraj Kasavajjala
Software Engineer
Kanaka Raju
Data Scientist
Akhil Sambar
Data Scientist

Session Highlights

It may have looked a little different this time, but the 4th edition of #herAIstory was an altogether unique experience for all the 60-odd attendees. The transition to a virtual webinar saw a Global audience from varied backgrounds and experiences. We may have been physically apart, but we still shared an exciting evening of learning, socializing, and fun!

The first segment was a deep dive into understanding the new business imperative of today’s ML world.

Our Consulting expert, Ramyasree Devarakonda, initiated the conversation by examining the way data labeling and augmentation have become the true bottleneck in AI/ ML projects. This step is indeed eating up most of the time, budget, and effort.

She highlighted some of the practical problems, including the ambiguity in the labeling process, the importance of hierarchy (particularly the element of context setting, and how different it will be for every project), the lack of collaboration between different teams, and so on.

Paying heed to the challenges faced by enterprises, Dheeraj steps in during the next segment, to unveil an enterprise-grade technology solution that not only uses state-of-the-art active learning techniques to deliver labeling that is faster, easier, and more accurate, but also addresses the labeling complexities of today. This solution is packed with all the features, framework support, models, and workflows to truly blaze a way out of the current-day ML problems. Curious to know more about Zaastra? Click here.

Up next, in the third Segment, it was Raju’s turn to walk us through the world of active learning – the why, what, and how dimensions of it. He sprinkled some tough geeky ML on us, including the concept of Empirical Risk Minimization (ERM), Vapnik-Chervonenkis (VC) Theory, and Disagreement based Active Learning.

These concepts were developed further by Akhil, with a detailed explanation of the most commonly used active learning strategies and the challenges faced by teams in adopting them.

Following a brief Q&A session, Srinivas Atreya revisited some of the conventional wisdom in deep learning. He dug deeper into why disagreement-based learning does not work in the deep neural network world. He highlighted some of the interesting developments in this area that will evolve rapidly over the next few years.

We are thankful for all the wonderful and constructive feedback received on the event. We will keep them in mind for the next one. However, we are thrilled that the most common refrain was

To know the level direction inat which technology is cruising, #herAIstory is a must attend event.

It is an excellent program for people who needs it, especially for working professionals and AI enthusiasts. #herAIstory will help us to explore more on AI or Data science.

We hope you enjoyed going through the session recap! We have plenty more exciting webinars coming your way in the next few weeks, so keep an eye out on our social media channels for more information!

To get an email invite for every session of herAIstory, register here.

A BIG THANK YOU to all the speakers and guests. We are overwhelmed by the number of people that came, the selfless manner in which knowledge was shared, and how engaged everyone was.

We are humbled by the thought that we can make a difference to this world, however small that might be. Such support motivates us and makes our resolve stronger to invest more time in this event.

Do watch out for another edition in October…

#herAIstory – let us change the Industry, one woman at a time