Welcome to our February 2023 month-to-month digest, the place you possibly can meet up with any AIhub tales you might have missed, get the low-down on latest occasions, and way more. This month, we glance again on the AAAI convention, congratulate award winners, and discover out extra about multimodal studying, and generative modelling.
We have been fortunate sufficient to have the ability to attend the thirty seventh AAAI Convention on Synthetic Intelligence (AAAI2023) in particular person. In addition to going to talks, poster periods, and different occasions, we gave a hands-on tutorial on “science communication for AI researchers”.
The invited talks are all the time an necessary a part of any convention, and this 12 months didn’t disappoint. The programme chairs had put collectively a diversified and fascinating choice of matters, starting from healthcare to robotics, from finance to physics. To this point, we’ve written summaries of the talks by Tuomas Sandholm and Manuela Veloso, and you’ll learn these write-ups right here:#AAAI2023 invited discuss: Tuomas Sandholm on organ exchanges#AAAI2023 invited discuss: Manuela Veloso on experience-based insights from AI in robotics and AI in finance
Yow will discover out what convention attendees, audio system and organisers thought concerning the week in our two Twitter round-ups:#AAAI2023 tweet round-up from the primary two days#AAAI2023 tweet round-up: the convention in abstract
In a particular award session, one of the best papers of the convention have been introduced. The AAAI-2023 excellent paper award went to Joar Skalse and Alessandro Abate for his or her work Misspecification in Inverse Reinforcement Studying. The AAAI-2023 excellent scholar paper award was given to Beautify the Newcomers: Visible Area Immediate for Continuous Check Time Adaptation, authored by Yulu Gan, Yan Bai, Yihang Lou, Xianzheng Ma, Renrui Zhang, Nian Shi, and Lin Luo. There have been additionally 12 distinguished paper award winners, the small print of which will be discovered right here.
In addition to these finest paper awards, a variety of prestigious AAAI awards have been introduced on the convention. These included the AAAI Award for Synthetic Intelligence for the Good thing about Humanity, which was gained by Tuomas Sandholm. Yow will discover out extra about this prize, and the others awarded, right here.
There shall be a lot extra content material to return as we proceed to cowl the convention, and listen to from individuals about their work. Yow will discover our convention protection right here, and this assortment shall be up to date as quickly as we add new content material.
ACM SIGAI Autonomous Brokers award
Congratulations to Edith Elkind on profitable the 2023 ACM/SIGAI Autonomous Brokers Analysis Award. This prestigious award is made for excellence in analysis within the space of autonomous brokers. It’s supposed to acknowledge researchers in autonomous brokers whose present work is a crucial affect on the sphere. The award choice committee recognised Edith for “her important analysis contributions in computational social alternative and algorithms for cooperative video games, and extraordinary service to the neighborhood”.
A survey on knowledge-enhanced multimodal studying
Multimodal studying is a subject of accelerating curiosity within the analysis neighborhood. On this blogpost, Maria Lymperaiou and Giorgos Stamou survey the state of the sphere of knowledge-enhanced multimodal studying, with a give attention to visiolinguistic (KVL) studying.
Riemannian score-based generative modelling
Valentin De Bortoli, Emile Mathieu, Michael John Hutchinson, James Thornton, Yee Whye Teh and Arnaud Doucet gained a NeurIPS2022 excellent paper award for his or her work Riemannian Rating-Primarily based Generative Modelling, by which they generalise score-based generative fashions from Euclidean house to Riemannian manifolds. On this blogpost, Valentin explains the work.
EurAI dissertation award
Nominations are invited for the 2022 Synthetic Intelligence Dissertation Award sponsored by EurAI (European Affiliation of Synthetic Intelligence). The award recognises superior analysis and writing by doctoral candidates in synthetic intelligence. It’s introduced yearly on the European Convention on Synthetic Intelligence, with the winner invited to current a chat on the convention. The deadline for nominations is 31 Could 2023, and yow will discover out extra data right here.
AfriSenti: sentiment datasets for tweets in African languages
This month noticed the discharge of AfriSenti: A Twitter Sentiment Evaluation Benchmark for African Languages. This paper introduces AfriSenti, a dataset of 110,000+ tweets in 14 African languages, annotated by native audio system. The dataset is out there right here. The paper describes the info assortment methodology, annotation course of, and associated challenges when curating every of the datasets.
AI and neurotechnology: studying from AI ethics
On this article, Sara Berger and Francesca Rossi speak about a few of the potential moral points regarding neurotechnology, and recommend that classes discovered from AI ethics might assist on this area. Their intention is to provoke a dialog and work in direction of addressing recognized neuroethics points. Francesca summarises the goals of the work on this video.
A blurry jpeg of the net
Ted Chiang gives his tackle ChatGPT (and huge language fashions usually) on this piece in The New Yorker. “Consider ChatGPT as a blurry jpeg of all of the textual content on the Internet. It retains a lot of the knowledge …, however, when you’re searching for a precise sequence of bits, you gained’t discover it; all you’ll ever get is an approximation.”
Jordan Harrod does 5-minute crafts with the assistance of GPT-3
Jordan Harrod recurrently posts movies about AI and know-how on her fashionable YouTube channel. This month, Jordan used GPT-3 to recommend some 5-minute crafts, which she tried out on digicam.
Stochastic Parrots Day
Friday 17 March marks the 2 12 months anniversary because the publication of On the Risks of Stochastic Parrots: Can Language Fashions Be Too Large?🦜. Be a part of the co-authors of the paper and varied company at a particular (digital) occasion to mirror on what has occurred within the final two years, what the big language mannequin panorama at present seems like, what trajectory are we following, and the way this compares to the trail we must be taking.
Our sources pageForthcoming and previous seminars for 2023AI world wide focus seriesUN SDGs focus seriesNew voices in AI collection
tags: month-to-month digest
, Managing Editor for AIhub.
Leave a Reply