Shagun Sodhani

📢 2 weeks until #CoLLAs2025!

Here’s what you need to make the most of it

🗓 Program: lifelong-ml.cc/Conferences/...

✅ Papers: lifelong-ml.cc/Conferences/...

📍 Venue: lifelong-ml.cc/Conferences/...

🔗 Info & registration: lifelong-ml.cc/Conferences/...

#AI #ML #ContinualLearning #LifelongLearning

We've been using fairseq2 internally for over a year now. It's stable, developer-friendly, and integrates seamlessly with frameworks like vLLM and accelerate. Highly recommend giving it a try!

Papers We Love A repository of computer science papers and a community of people who love reading them

4. Try to get at least one other person as a co-organizer so that you are not the single pint of failure. A person who regularly attends the sessions maybe willing to step up as a co-organizer.

You may find paperswelove.org useful.

Best of luck with your reading club!

3. Give people incentives to present. e.g. publicize the sessions/speakers. Make it easier for people to present by maintaining a list of “suggested papers”. Make sure you have read the paper so that there is at least one person in the audience who has read the paper. 3/

2. These sessions are lot more fun if the audience has read the paper. You can make it easier for the audience by organizing the sessions around themes so that papers in subsequent sessions are “related”. This also helps the speakers because they can build on previous sessions. 2/

I ran a reading group at my grad school. Couple of thoughts :

1. Consistency is key - you want to stick to a schedule and not cancel sessions. Sometimes that means you are presenting back to back weeks - because the speaker had to cancel at last minute or maybe there was no signed up speaker 1/

Excited to announce that Dr. Martha White will be an invited speaker at #CoLLAs2025! 🎤

An Associate Professor at the U of Alberta, Canada CIFAR AI Chair, & member of RLAI & Amii, her research focuses on adaptive systems that #continually learn and improve.

Stay tuned for more details on her talk!

DTensor - Status, Design and Looking Forward This post outlines the current status, design principles, and future work items for PyTorch DTensor. Overall Design Objective and Principles DTensor is the PyTorch native tensor sharding primitive. F...

A very detailed note from #PyTorch devs about DTensor (Distributed Tensors) - how they are designed, what is supported / not supported and whats planned for the future: dev-discuss.pytorch.org/t/dtensor-st...

Any chance the recording would be available?

Harnessing small projectors and multiple views for efficient vision... Recent progress in self-supervised (SSL) visual representation learning has led to the development of several different proposed frameworks that rely on augmentations of images but use different...

I will be at @neuripsconf.bsky.social next week! If you're around, let's connect! Would love to discuss #agents, #reasoning, continual learning, careers in industrial research labs & more! We are also presenting a paper on efficient SSL for vision openreview.net/forum?id=Y5D...

Created a starter pack for people working in continual / lifelong ML. This is obviously incomplete - please comment names (including your own) and I will add them

go.bsky.app/RB4ekFs

Could you add me as well!

We've had three successful editions since 2021, with our fourth edition announced just recently!

Follow @collasconf.bsky.social for updates and join the conversation! 🚀

A couple of organizers are also attending NeurIPS 2024 so if you are around, come say hi 👋 3/3

As the @icml.bsky.social , @neuripsconf.bsky.social , and @iclr-conf.bsky.social communities grow, there's a need for focused conferences that foster strong communities around specific ML sub-topics and drive progress. That's where CoLLAs comes in! 🎯 2/

Collas (@collasconf.bsky.social) is on 🦋!

CoLLAs is the International Conference on Lifelong Learning Agents. It was founded in 2021 to advance research in lifelong learning, meta-learning, multi-task learning, transfer learning & OOD generalization.

Learn more here: lifelong-ml.cc 1/