Scaling Instructable Agents Across Many Simulated Worlds (SIMA)

Blog:

• https://deepmind.google/discover/blog/sima-generalist-ai-agent-for-3d-virtual-environments/

How is this different from GENIE, DreamerV4?

This is not a generative model, its just an agent. However, there’s no RL involved. Very similar to the LLM paradigm. Always off-policy.

It gets ground truth from how to play a game, the quality of data seems quite important here?c

Citation

(Team et al., 2024)

Team, S., Raad, M. A., Ahuja, A., Barros, C., Besse, F., Bolt, A., Bolton, A., Brownfield, B., Buttimore, G., Cant, M., Chakera, S., Chan, S. C. Y., Clune, J., Collister, A., Copeman, V., Cullum, A., Dasgupta, I., de Cesare, D., Trapani, J. D., … Young, N. (2024). Scaling Instructable Agents Across Many Simulated Worlds (SIMA). https://arxiv.org/pdf/2404.10179