Ques 1:

Which of the following is currently the most difficult computer vision (CV) task?

- Take a picture and turn it into a 3D description

- Produce accurate image classifiers that use learned features

- Given an image, detect an object and its geometric location

- Given a description of a 3D scene, produce a picture of it

Ques 2:

In robotics, what challenge does reinforcement learning address?

- A model of the dynamic world cannot be updated for new tasks

- It can be difficult to write down a dynamic model of the world

- It can be difficult to know the expected outcome of possible actions in a dynamic world

- It can be difficult to write down a stochastic model of the world