AutoML-Conf 2022 Competitions - DAC4AutoML
Raghu Rajan
The AutoML Freiburg-Hannover group is organising the following two competitions:
1) Configuring hyperparameters of RL agents dynamically at https://codalab.lisn.upsaclay.fr/competitions/3727
The environments are variants of Pendulum, Acrobot, MountainCarContinuous, LunarLander and CartPole with factors, for example pole length and gravity for CartPole, varying across train and test sets.
2) Dynamically adapting the learning rate of the SGD optimizer at https://codalab.lisn.upsaclay.fr/competitions/3672
Dataset instances are based on the extended version of the SGDBenchmark included in DACBench.
Deadline: June 18, 2022.
Extended Deadline: July 9, 2022
The two competitions are separate tracks of the DAC4AutoML Competition at AutoML-Conf 2022.
Awards
We will provide you with certificates of participation (physical as well as digital), including your placement and medals for the top 3 teams of each track. Further prizes for the top 3 teams are copies of the AutoML book signed by the editors. These teams will also receive monetary prizes sponsored by ChaLearn:
250$ for first place
150$ for second
100$ for third
Furthermore, winners that present novel approaches will be invited to co-author a publication presenting the competition results with the organisers.
Best regards,
DAC4AutoML Organising Team,