Skip to content

Details

We can learn from the data. If we take a step ahead, we can learn from the models as well. This is called meta-learning.

But exactly why does meta-learning work? How to make them work even better? How to obtain an optimal meta-learner from an army of basic models?

Building a meta-learner is not a silver bullet to win that Kaggle competition. We also need to think about putting it in to production while making it embarrassingly parallel. We need to incorporate our domain expertise (if there is any) in it. We also might need to use a pre-trained model in the meta-learner to build our next generation cat detector Meta-meow.

In this talk we will be exploring all of these quirks about meta-learning using first principles and some poorly drawn sketches. We also will explore how ML-Ens can effectively answer most of the above questions.

Shah Mahdi Hasan is a Senior Machine Learning Scientist, PhD in Statistical Signal Processing from UoN. Specialities are convex optimization, timeseries forecasting and deep learning. Jazz nerd and applied machine learning fiend.

##-------------------------------------------------------------------##

VENUE:

NEW LOCATION!

The university's brand new Q building in Honeysuckle!

https://goo.gl/maps/35LwcLhKV1nvoFAh7

IMPORTANT this is NOT the NewSpace building where we met last year.

We will meet in the Level 2 Seminar Space. You'll need to enter through the main doors on Worth Place and will be directed to the elevators to join the meetup.

PARKING: The closest parking is Wright Lane Carpark.
https://goo.gl/maps/rj6Yg8EkpjYzUmbi7

After 5pm it is free and is a 400m - 5 minute walk.

##-------------------------------------------------------------------##

TIME: 5.30pm - 7.30pm

FOOD: 🍕 + 🥤

  • Wood fired pizza, sponsored by Newy Tech People
  • Please advise dietary requirements in RSVP
  • Soft drinks

SCHEDULE:

  • 5:30pm: Networking + Food + Drinks
    (Don't stress if you are held up in traffic)
  • 6:15pm: Presentation
  • 7:15pm: Packing up + Bonus questions
  • 7:30pm: Close

Related topics

Events in Newcastle
Artificial Intelligence
Machine Learning
Data Mining
Data Science
Predictive Analytics

You may also like