ResearchGuru Research
The ResearchGuru team is actively researching and building in reinforcement learning, vision-language models, large language models and robotics, with several papers under review.
What we work on
- Reinforcement learningRepresentation learning, exploration and the theory that ties them together, including our IEEE Access paper on bisimulation geometry.
- Vision-language modelsModels that connect images and language.
- Large language modelsLanguage models, their evaluation and their use in research tools.
- RoboticsLearning-based control and decision-making for robots.
Publications
- Bisimulation Geometry: Unifying Representation Learning and Exploration in Reinforcement LearningOne loss, the Bisimulation Infomax Loss, ties representation learning and curiosity-driven exploration to the same geometric object, the bisimulation metric, with no hand-tuned trade-off weight. The resulting method, Adaptive Bisimulation Mirror-Descent, cuts steps-to-threshold by up to 33% on MiniGrid and lifts returns by up to 17% on the Distracting Control Suite.Read the paper Download PDF
Several more papers are under review and will be listed here once they are published.
The team
Collaborate with us
Open to research collaboration. Send us a message here and we will reply.

