Updates
News, Events & Community
ARCathon News
With a dual career spanning clinical psychology and software development, Jack Cole and his team have brought a unique perspective to tackling the ARC Challenge.
Leveraging cognitive insights and advanced machine learning techniques, the team’s approach stands out for its novel use of test-time fine-tuning and synthetic data enhancement.
The wait is over and the ARCathon is underway again. Make up to three submissions per week and join the ARC circle mailing list and monthly calls.
All information can be found in the announcement letter linked here.
The ARCathon 2023 is a wrap! We want to thank all 265 participating teams and are excited about all the innovative submissions. Take a look at our final standings and podium.
But the ARCathon journey does not end here. Stay tuned!
Win a trip to Lab42 in Davos, Switzerland, for a multi-day workshop alongside François Chollet! Travel expenses are covered and accommodations are provided by Lab42. To enter, submit an explanation of your ARC approach along with your code. The top three teams meeting the judging criteria will be selected at the conclusion of ARCathon 2023. Click the title link for more details.
Two-time Essay Competition winner and ARCathon participant Simon Ouellette discusses the intricacies of deep learning in the context of the ARC challenge. He highlights its difficulties in classifying prime numbers and underscores the limitations of both purely deep learning and DSL-based methods. He then proposes a modular architecture as a potential bridge between these two methodologies.
We are excited to welcome François Chollet for an exclusive Q&A session on October 18 at 6 p.m. CEST (GMT +2)! To kick things off, our moderators, Michael Hodel and Simon Ouellette, will discuss François's insights on the challenges of ARC. Afterward, we'll open the floor for audience questions.
ARCathon participants Jack Cole and Mohamed Osman discuss their strategy to tackle ARC with large language models and resources from Google's TPU Research Cloud. They emphasized the benefits of diverse datasets and meta-learning, discovering that test-time fine-tuning significantly boosted their model's performance.
Since September 10, teams from over 60 countries from 6 continents have registered. We are just waiting for you, Antarctica...
We have reached a total of over 200 teams in the ARCathon 2023! The Lab42 crew wishes all teams the best of luck in breaking the ARC world record of 30.5%!
In 2019, François Chollet published his paper "On the Measure of Intelligence" along where he introduced the Abstraction and Reasoning Corpus. Almost four years later, after a competition on Kaggle and with Lab42's second ARCathon in progress, the time is ripe to take a look at where we stand today, including an update from seven devoted ARC specialists.
The ARCathon evaluation is designed to be flexible and is deliberately based on Docker. Nevertheless, some have found the step of integrating their code into a Docker image a hurdle. For this reason, our ARCathon participant Taka has kindly published a repo with a ready-to-use Docker image. Fork this repo to submit your algorithms in the ARCathon with minimal hassle!
Simon Ouellette, the previous winner of the Principles of Intelligence essay competition, has done it again. In our latest essay competition, which ended in May, Simon detailed in 26 well-written pages how he intends to solve the ARC problem! Read his full essay to immerse yourself in the thought process of an ARC enthusiast, or simply read his executive summary for inspiration.
ARC Publications
April 10, 2024
March 18, 2024
Seungpil Lee, Woochang Sim, Donghyeon Shin, Sanha Hwang, Wongyu Seo, Jiwon Park, Seokki Lee, Sejin Kim, Sundong Kim
March 13, 2024
Gustaw Opiełka, Hannes Rosenbusch, Veerle Vijverberg, Claire E. Stevenson
February 14, 2024
Yudong Xu, Wenhao Li, Pashootan Vaezipoor, Scott Sanner, Elias B. Khalil
February 7, 2024
Natasha Butt, Blazej Manczak, Auke Wiggers, Corrado Rainone, David Zhang, Michaël Defferrard, Taco Cohen
February 5, 2024
Mikel Bober-Irizar, Soumya Banerjee
January 15, 2024
Chao Lei, Nir Lipovetzky, Krista A. Ehinger
December 13, 2023
Mukul Singh, Jose Cambronero, Sumit Gulwani, Vu Le, Gust Verbruggen
November 14, 2023
Melanie Mitchell, Alessandro B. Palmarini, Arseny Moskvichev
November 14, 2023
Luca H. Thoms, Karel A. Veldkamp, Hannes Rosenbusch, Claire E. Stevenson
November 1, 2023
October 31, 2023
James Ainooson, Deepayan Sanyal, Joel P. Michelson, Yuan Yang, Maithilee Kunda
October 12, 2023
Simon Ouellette, Rolf Pfister, Hansueli Jud
October 8, 2023
September 11, 2023
Ruocheng Wang, Eric Zelikman, Gabriel Poesia, Yewen Pu, Nick Haber, Noah D. Goodman
July 10, 2023
Suvir Mirchandani, Fei Xia, Pete Florence, Brian Ichter, Danny Driess, Montserrat Gonzalez Arenas, Kanishka Rao, Dorsa Sadigh, Andy Zeng
June 14, 2023
Jaehyun Park, Jaegyun Im, Sanha Hwang, Mintaek Lim, Sabina Ualibekova, Sejin Kim, …
June 6, 2023
Tan John Chong Min
June 5, 2023
Mattia Atzeni, Mrinmaya Sachan, Andreas Loukas
May 29, 2023
Di Huang, Ziyuan Nan, Xing Hu, Pengwei Jin, Shaohui Peng, Yuanbo Wen, Rui Zhang, Zidong Du, Qi Guo, Yewen Pu, Yunji Chen
May 26, 2023
Yudong Xu, Wenhao Li, Pashootan Vaezipoor, Scott Sanner, Elias B. Khalil
May 11, 2023
Arseny Moskvichev, Victor Vikram Odouard, Melanie Mitchell
May 5, 2023
Rim Assouel, Pau Rodriguez, Perouz Taslakian, David Vazquez, Yoshua Bengio
March 7, 2023
Giacomo Camposampiero, Loic Houmard, Benjamin Estermann, Joël Mathys, Roger Wattenhofer
February 18, 2023
James Ainooson, Deepayan Sanyal, Joel P. Michelson, Yuan Yang, Maithilee Kunda
January 8, 2023
Jonas Witt, Stef Rasing, Sebastijan Dumančić, Tias Guns, Claus-Christian Carbon
October 21, 2022
Subin Kim, Prin Phunyaphibarn, Donghyun Ahn, Sundong Kim
October 18, 2022
Yudong Xu, Elias B. Khalil, Scott Sanner
July 4, 2022
Daniel Zeng, Tailin Wu, Jure Leskovec
June 28, 2022
Victor Vikram Odouard, Melanie Mitchell
March 25, 2022
Rim Assouel, Pau Rodriguez, Perouz Taslakian, David Vazquez, Yoshua Bengio
December 1, 2021
October 22, 2021
Simon Alford, Anshula Gandhi, Akshay Rangamani, Andrzej Banburski, Tony Wang, …
June 2021
June 15, 2021
Samuel Acquaviva, Yewen Pu, Marta Kryven, Theodoros Sechopoulos, Catherine Wong, …
April 8, 2021
Yonggang Qi, Kai Zhang, Aneeshan Sain, Yi-Zhe Song
March 10, 2021
Aysja Johnson, Wai Keen Vong, Brenden M. Lake, Todd M. Gureckis
November 23, 2020
Andrzej Banburski, Simon Alford, Anshula Gandhi, Sylee Dandekar, Peter Chin, Tomaso Poggio
September, 2020
Raphael Fischer, Matthias Jakobs, Sascha Mücke, Katharina Morik
November 12, 2020
Victor Kolev, Bogdan Georgiev, Svetlin Penkov
November 5, 2019
François Chollet
Lab42 Updates
With a dual career spanning clinical psychology and software development, Jack Cole and his team have brought a unique perspective to tackling the ARC Challenge.
Leveraging cognitive insights and advanced machine learning techniques, the team’s approach stands out for its novel use of test-time fine-tuning and synthetic data enhancement.
The ARCathon 2023 is a wrap! We want to thank all 265 participating teams and are excited about all the innovative submissions. Take a look at our final standings and podium.
But the ARCathon journey does not end here. Stay tuned!
Win a trip to Lab42 in Davos, Switzerland, for a multi-day workshop alongside François Chollet! Travel expenses are covered and accommodations are provided by Lab42. To enter, submit an explanation of your ARC approach along with your code. The top three teams meeting the judging criteria will be selected at the conclusion of ARCathon 2023. Click the title link for more details.
Two-time Essay Competition winner and ARCathon participant Simon Ouellette discusses the intricacies of deep learning in the context of the ARC challenge. He highlights its difficulties in classifying prime numbers and underscores the limitations of both purely deep learning and DSL-based methods. He then proposes a modular architecture as a potential bridge between these two methodologies.
We are excited to welcome François Chollet for an exclusive Q&A session on October 18 at 6 p.m. CEST (GMT +2)! To kick things off, our moderators, Michael Hodel and Simon Ouellette, will discuss François's insights on the challenges of ARC. Afterward, we'll open the floor for audience questions.
ARCathon participants Jack Cole and Mohamed Osman discuss their strategy to tackle ARC with large language models and resources from Google's TPU Research Cloud. They emphasized the benefits of diverse datasets and meta-learning, discovering that test-time fine-tuning significantly boosted their model's performance.
Since September 10, teams from over 60 countries from 6 continents have registered. We are just waiting for you, Antarctica...
We have reached a total of over 200 teams in the ARCathon 2023! The Lab42 crew wishes all teams the best of luck in breaking the ARC world record of 30.5%!
In 2019, François Chollet published his paper "On the Measure of Intelligence" along where he introduced the Abstraction and Reasoning Corpus. Almost four years later, after a competition on Kaggle and with Lab42's second ARCathon in progress, the time is ripe to take a look at where we stand today, including an update from seven devoted ARC specialists.
Simon Ouellette, the previous winner of the Principles of Intelligence essay competition, has done it again. In our latest essay competition, which ended in May, Simon detailed in 26 well-written pages how he intends to solve the ARC problem! Read his full essay to immerse yourself in the thought process of an ARC enthusiast, or simply read his executive summary for inspiration.
For Humankind
Villa Fontana Obere Strasse 22B 7270 Davos, Switzerland
Powered by