Dylan Cope

Dylan Cope

Mr

Personal profile

Research interests

PhD student working in studying emergent communication in multi-agent reinforcement learning, with broad interesting spanning from AI to evolutionary biology. Outside my current PhD focus, I have worked on the following projects:

  • Characterising "explanation" in multiagent communication. What does it mean for one agent to "explain" something to another?
  • Studying robustness to distributional shift in neural networks. We experimented with the learned latent representations within networks to investigate how novel situations are encoded.
  • Proposing an algorithm for applying reinforcement learning to select nodes to expand in a tree search process over a complex continuous environment.
  • Investigating the evolution of multicellularity within a real-time physically simulated environment. In particular, looking at the role of gene regulation for cell specialisation within a group of cells acting as a whole.

Keywords

  • QA75 Electronic computers. Computer science

Fingerprint

Dive into the research topics where Dylan Cope is active. These topic labels come from the works of this person. Together they form a unique fingerprint.
  • 1 Similar Profiles

Collaborations and top research areas from the last five years

Recent external collaboration on country/territory level. Dive into details by clicking on the dots or
  • Improving Activation Steering in Language Models with Mean-Centring

    Jorgensen, O., Cope, D., Schoots, N. & Shanahan, M., 26 Feb 2024, Responsible Language Models Workshop (ReLM) at AAAI-24.

    Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

    File
    88 Downloads (Pure)
  • Learning Translations: Emergent Communication Pretraining for Cooperative Language Acquisition

    Cope, D. & McBurney, P., 2 Aug 2024, Proceedings of the 33rd International Joint Conference on Artificial Intelligence (IJCAI-24).

    Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

    File
    63 Downloads (Pure)
  • Learning Translations: Emergent Communication Pretraining for Cooperative Language Acquisition

    Cope, D. & McBurney, P., 26 Feb 2024, Ad Hoc Teamwork Workshop at AAAI-24.

    Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

    File
    51 Downloads (Pure)
  • Learning to Plan with Tree Search via Deep RL

    Cope, D., Svegliato, J. & Russell, S., 2023, (Accepted/In press) Bridging the Gap Between AI Planning and Reinforcement Learning Workshop (PRL @ IJCAI 2023) at the International Joint Conference on Artificial Intelligence.

    Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

    Open Access
    File
    139 Downloads (Pure)
  • Low-Entropy Latent Variables Hurt Out-of-Distribution Performance

    Schoots, N. & Cope, D., 2023, (Accepted/In press) International Conference on Learning Representations 2023: Workshop on Domain Generalization.

    Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

    Open Access
    File
    63 Downloads (Pure)