# CMU Seminar Series : Rohan Paul, 2021
###### tags: `robotics` `ai` `cmutalks`

[LINK TO SOURCE](https://youtu.be/JTruH1NvWmQ)

Pretty interesting talk about developing generalized models of world understanding through abstract representations of language and vision for robotic task learning. Some of the topics introduced:

- Cognitive model for Robots
- Formulation of a world model
- Grounding Graphs : A tool to correlate verbal instructions with robot world model
- Adaptive Distributed Correspondence graphs
- Temporal Gronding Graphs - Remembering facts about the world
- A imitation learning based tool for predicting Tool interaction
- Common Sense Learning
- Technical challenges forward