Publications

For an exhaustive updated list of scientific publications, please check my Google Scholar Page. My extended resume is here
Thesis
- Akakzia, Ahmed. (2022). Teaching Predicate-based Autotelic Agents: Learning Goal Representations with a Social Caregiver for Intrinsically Motivated Agents. (Link to manuscript).
Link to video and slides will be provided after defense.
Articles
2022
- A Akakzia, O Sigaud (2022). Learning Object-Centered Autotelic Behaviors with Graph Neural Networks.
Accepted at ICLR’s ALOE Workshop 2022 and CoLLAs 2022.
Available code. - A Akakzia, O Serris, O Sigaud, C Colas (2022). Help Me Explore: Minimal Social Interventions for Graph-Based Autotelic Agents.
Accepted at IMOL 2022.
Available code.
2021
- A Akakzia, C Colas, PY Oudeyer, M Chetouani, O Sigaud (2021). Grounding language to autonomously-acquired skills via goal generation.
Accepted at ICLR 2021.
Available talk.
Available code. - Olivier Sigaud, Ahmed Akakzia, Hugo Caselles-Dupré, Cédric Colas, Pierre-Yves Oudeyer, Mohamed Chetouani (2021). Towards Teachable Autotelic Agents.
Under review at IEEE Transactions on Cognitive and Developmental Systems. - F Jarboui, A Akakzia (2021). Delayed Geometric Discounts: An alternative criterion for Reinforcement Learning
Available code.
2020
- C Colas, A Akakzia, PY Oudeyer, M Chetouani, O Sigaud (2020). Language-conditioned goal generation: a new approach to language grounding for RL.
Accepted at ICML’s LaReL Workshop 2020.
Available code.
2019
- A Akakzia, M Chetouani, O Sigaud (2019). GARGAML: Addressing Multitask and Multi-goal Transfer with Goal-conditioned Meta Reinforcement Learning and Replay.
Available slides