Efficient Continual Learning with Modular Networks and Task-Driven Priors

We organize online meetups and events for the continual learning community!

ContinualAI

ContinualAI Reading Group: "Efficient Continual Learning with Modular Networks and Task-Driven Priors"

This event is free and open to everyone. It will be hosted on Google Meet: <a href="https://join.slack.com/t/continualai/shared_invite/enQtNjQxNDYwMzkxNzk0LTBhYjg2MjM0YTM2OWRkNDYzOGE0ZTIzNDQ0ZGMzNDE3ZGUxNTZmNmM1YzJiYzgwMTkyZDQxYTlkMTI3NzZkNjU" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">join us on slack to get the link</a> (the #reading-group channel). Otherwise you can grab you ticket (link inside) with a voluntary donation (even just 1$ will do)!Please consider that:<ol><li>this is a community effort: the speakers &amp; organizers of the event do not receive any compensation.</li><li>the full amount collected through donations will be used exclusively to face CAI costs for event hosting and community services.</li></ol>--------- Event Description -----------This Friday 15-1-2021, 5.30pm CET, for the ContinualAI Reading Group, Tom Veniat (Sorbonne Universite) will present the paper:Title: <a href="https://openreview.net/forum?id=EKV158tSfwv" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">Efficient Continual Learning with Modular Networks and Task-Driven Priors</a>Abstract: Existing literature in Continual Learning (CL) has focused on overcoming catastrophic forgetting, the inability of the learner to recall how to perform tasks observed in the past. There are however other desirable properties of a CL system, such as the ability to transfer knowledge from previous tasks and to scale memory and compute sub-linearly with the number of tasks. Since most current benchmarks focus only on forgetting using short streams of tasks, we first propose a new suite of benchmarks to probe CL algorithms across these new axes. Finally, we introduce a new modular architecture, whose modules represent atomic skills that can be composed to perform a certain task. Learning a task reduces to figuring out which past modules to re-use, and which new modules to instantiate to solve the current task. Our learning algorithm leverages a task-driven prior over the exponential search space of all possible ways to combine modules, enabling efficient learning on long streams of tasks. Our experiments show that this modular architecture and learning algorithm perform competitively on widely used CL benchmarks while yielding superior performance on the more challenging benchmarks we introduce in this work.The event will be moderated by: <a href="https://www.vincenzolomonaco.com/" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">Vincenzo Lomonaco</a>.--------- ContinualAI Reading Group -----------The ContinualAI Reading Group is hosted every Friday in collaboration with MILA and it is based on roughly 60 minutes discussion about a particular Continual Learning paper. Occasionally, a speaker is invited to explain his work. Join our reading group on slack! It will be the occasion to talk with other researchers of this amazing field!--------------- About ContinualAI ---------------ContinualAI is an official non-profit research organization and the largest open community on Continual Learning for AI. We aim at connecting people and working better together on this fascinating topic we consider fundamental for the future of AI.• Official website: <a href="https://www.continualai.org" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">https://www.continualai.org</a>• Join us now at <a href="https://www.continualai.org/join_us" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">https://www.continualai.org/join_us</a>Please consider supporting us with a small donation at: <a href="https://www.continualai.org/supporters" target="_blank" data-msys-clicktrack="0" rel="nofollow noopener noreferrer">https://www.continualai.org/supporters</a>It&#x27;s thanks to people like you that we are making ContinualAI a reality!

Our {communityGuidelinesLink} describe the sort of content we prohibit on Eventbrite. If you suspect an event may be in violation, you can report it to us so we can investigate.

To report other categories of prohibited or illegal content, submit {link}

Efficient Continual Learning with Modular Networks and Task-Driven Priors

More events from ContinualAI

Follow organizers to get events picked for you

Still looking for the right event?

Explore all online events to browse and filter by date, category, and more.

Efficient Continual Learning with Modular Networks and Task-Driven Priors

More events from ContinualAI

Follow organizers to get events picked for you

You might also like...

Browse more events with different dates, prices, and formats to find your next great experience.

Still looking for the right event?

Explore all online events to browse and filter by date, category, and more.