ML-India: Gurgaon chapter | Second Meetup

ML India is delighted to inform that the second meet-up in its Gurgaon chapter was held last Saturday (25th June 2016).

They had a guest speaker - Rishabh Mehrotra, PhD student from University College, London. Rishabh has a hands-on experience of working with some of the leading researchers on ‘Proactive Search and Digital Assistants’. His talk on the subject was very impressive and enlightening. He was able to tune with the audience and focused on the big picture ideas and key takeaways. Everyone found his talk to be very useful.

Some topics of discussion were:

How do search companies use user data for personalization?
What is task based search?
Discussion on personalization vs biases - Exploration vs Exploitation.
Different features that are utilized by search engines like Google.
Little digression toward deep learning, its pros and cons and its hype.
Contextual modelling in conversation bots - memory based models.
Search tasks: focused, multi-tasking, super-tasking users.
Non-parametric Bayesian models and Chinese restaurant.

After the talk, some small group Q/A sessions ensued in which people talked about their personal doubts and experiences in machine learning. There was a debate going around on whether Python is appropriate for production related big data systems. One of the participants proposed that Scala does a way better job for production-related things while Python should be used for rapid prototyping and experimentation.

Following this, they had a small discussion on how to improve these meet-ups and carry the initiative forward. Some important suggestions that came up were:

Members can contribute to the meetups by suggesting guest speakers and inviting more people for such talks.
Meet-ups can be made more immersive by steering the discussions into a more technical direction by focusing on specific topics and discussing research papers.
Once the structure is more organised and established, splitting the meet-up into two groups, one for beginners and another for more technical people, can be considered.

Overall, the meet-up was super successful with a significant turnout of roughly 35 participants. The crowd was very diverse and enthusiastic. They had 5-6 undergraduate students, 2-3 graduates, two in their PhDs and one professor. One participant was from a firm working to materialize new hardware optimized for machine learning algorithms. Another participant was from an HR consultancy firm looking forward to using machine learning to select better candidates, something that Aspiring Minds works on.

We really look forward to having such engaging and interesting meetups, and having researchers and practitioners involved in ML to discuss their work and insights.