View this email in your browser


Hello everyone! If you happen to be in the states, then this period is a holiday traditionally associated with giving thanks. Even if you're not, especially with the period we've been through this year, now seems as good of time as any to pause and meditate on some of the good that has continued on through this time. 

For us at ContinualAI, this is certainly synonymous with our members. Within the community that we have amassed through grassroots efforts, with a shared desire to make AI a little bit better, maybe, through putting our heads together and solving hundreds and thousands of problems in AI -of all sizes, the tool of AI will emerge to help solve problems outside of the field. 

We are grateful for each and every one of our over 1000 members! Our community is a key example of what a globalized group working towards a shared vision may be able to accomplish. Open source projects, sharing the latest research, workshops, answering questions, and valuable discussions have not only been exciting, but have helped catalyze this field of research. The future only knows what will emerge, but there is a high probability that the whole is more than the sum of its parts.  

We thank every one of you who have been along for the ride, for however long you have been with us. We hope that ContinualAI has been a silver lining to a crazy year, pushing this branch of AI forward, continually.  


We're happy to share what has been working on, towards accelerating research surrounding continual learning AI: a necessary step in the direction of strong AI. Passionate about our mission? Join us on slack if you haven't already and feel free to donate if you are passionate about this goal.  

A Few Recent Announcements

  • We are excited to announce the next ContinualAI Online Meetup (This Friday 5.30 PM CET)! This meetup will be about "Benchmarks and Evaluation for Continual Learning". See the speakers and topics for the meetup above, and prepare your questions! We will make sure to make a strong panel discussion at the end of the meetup! To join, the Eventbrite link is here and MS Teams link is here!
  • Miss our last meetup about "Generalization and Robustness in Continual Learning"? We had a wonderful group of speakers and exciting discussion! Find the recording of the event here, or view the discussion here.
  • Still haven't checked out the updated ContinualAI Wiki? Bring yourself up to speed by watching the meetup event talking all about it, which you can watch here , or if you want to get involved reach out on slack or check out our discussion here. A big thank you to Andrea Cossu for heading up this project. 
  • We've maintained a great reading group line up. Can't make this week's reading group? No worries! See the past papers here, and you can also watch the recordings of all the events that we have had.
  • The ContinualAI Lab collaborative team is always looking for contributors to the many open-source projects we have under development. Contact us on Slack if you want to learn more about them and join us! We are always looking for motivated people willing to give back to this awesome community!
Not on our mailing list? Join now!

ContinualAI Sponsored Programs

  • Please reach out if you would be interested in us sponsoring your program!
ContinualAI has been an open community from the beginning. From the start, we have strived to make it a more diverse, equitable, and inclusive organization, which will help our mission of making research in continual learning & AI more fair, open and collaborative. This is what we have always hoped ContinualAI would embody!

Towards this goal, we are excited to announce the creation of an Inclusion & Diversity committee within ContinualAI to improve the quality of our community with a particular attention to fulfilling this mission. If you have an idea how to further reach this goal, please feel free to contact us, we would love to hear your ideas. 

Top paper picks: 

A paper we think you should read, if you have not yet, as chosen by the community:

Multi-Task Incremental Learning for Object Detection

Xialei LiuHao YangAvinash RavichandranRahul BhotikaStefano Soatto 

Multi-task learns multiple tasks, while sharing knowledge and computation among them. However, it suffers from catastrophic forgetting of previous knowledge when learned incrementally without access to the old data. Most existing object detectors are domain-specific and static, while some are learned incrementally but only within a single domain. Training an object detector incrementally across various domains has rarely been explored. In this work, we propose three incremental learning scenarios across various domains and categories for object detection. To mitigate catastrophic forgetting, attentive feature distillation is proposed to leverages both bottom-up and top-down attentions to extract important information for distillation. We then systematically analyze the proposed distillation method in different scenarios. We find out that, contrary to common understanding, domain gaps have smaller negative impact on incremental detection, while category differences are problematic. For the difficult cases, where the domain gaps and especially category differences are large, we explore three different exemplar sampling methods and show the proposed adaptive sampling method is effective to select diverse and informative samples from entire datasets, to further prevent forgetting. Experimental results show that we achieve the significant improvement in three different scenarios across seven object detection benchmark datasets.

Other Useful Links: 


Copyright © 2020 ContinualAI, All rights reserved.

Our mailing address is:

Want to change how you receive these emails? ;(
I suppose that you can update your preferences or unsubscribe from this list.

This email was sent to <<Email Address>>
why did I get this?    unsubscribe from this list    update subscription preferences
ContinualAI · Via Caduti della Via Fani, 7 · Bologna, Bo 40121 · Italy

Email Marketing Powered by Mailchimp