Friends from the Canadian CS & Computer/Software Engineering community, consider submitting to the CS-Can Student Symposium (cscan-infocan.ca/spac/symposium/) happening McGill University between June 2-4 2019. Submission Deadline: April 15 2019. Please share with your colleagues! Thanks

Diversity Panel at cscan-infocan.ca/spac/symposium/ talking about 'Diversity and Representation of Underrepresented Groups in CS Workshop' (Moderated by Ariella Smofsky) Khimya Charles C Onu

Julie Nutini, winner of CS-CAN Award gives a talk at SPAC symposium (cscan-infocan.ca/spac/) on 'story telling in academia'. Drop by at mcgill education building 624 if you are around!

AI4Good Lab is a beautifully crafted, extensive, and ambitious program! Personally, it was more challenging than a regular TA. I felt more responsible and hence more cautious and anxious. But it was ALL worth it. Consider applying to be a TA in the coming years.

Really happy to share our work on using reinforcement learning for accelerating MRI acquisition, which we presented at @MICCAI2020 today (joint work with amazing collaborators Sumana Basu Adriana Romero-Soriano @ CVPR Roberto Calandra Michal Drozdzal). 1/7

If you missed our talk this morning AAAI, or want to chat about RL and healthcare, or just want to say Hi, please visit our poster tonight! Poster ID: 274 When: 7-9PM Full Paper: arxiv.org/abs/2301.00512 Short Video: youtube.com/watch?v=YwhvZN… See you soon!

🚨New Paper!🚨 We posit and show that human-like language lies in a narrow, flat stable entropy zone, and violations of this entropy band result in degenerate behavior. joint work w/ Timothy J. O’Donnell, Doina Precup, Jason Weston, & Jackie Cheung. 1/n arxiv.org/abs/2302.06784

Please reach out if you have any questions, comments, suggestions, or just want to chat about LLMs, RL+NLP, and NLG, in general. The paper is available at: arxiv.org/abs/2302.06784 Entropy-Aware Decoding implementation: github.com/kushalarora/tr… 8/n

Excited to share our #ACL2023NLP (#NLProc) paper on "Information Synchronization Across Multilingual Semi-Structured Tables"! 📚🔍 We explore the problem of Information Synchronization across multilingual tables and construct a large-scale dataset InfoSync. #NLPResearch -1/n

I prefer to operate in “GPU-Poor” mode. I don’t agree with the take from the semianalysis piece. Creative breakthroughs often occur under constraints—new systems, models, and methods that can better take advantage of even larger-scale compute semianalysis.com/p/google-gemin…