Kevin Chen

I am at Apple. I graduated with my PhD from the Stanford Vision and Learning Lab, where I was advised by Silvio Savarese. During my PhD, I worked on problems related to computer vision, robotics, and natural language processing.

I was an intern at Robotics @ Google where I worked on semantic visual navigation and Argo AI where I worked on 3D detection and segmentation.

Email  /  Google Scholar  /  Github

We learn language-conditioned visuomotor skills on real robots from entirely offline, pre-collected datasets and crowdsourced language annotation.

Modular approach to vision-and-language navigation utilizing transformers for generating navigation plans which are subsequently executed by a low-level controller.

Navigation to an object in an unexplored environment with low-level control by training with scaled up soft actor critic.

Using splines to register lidar observations with hand-drawn maps.

Introducing updates to the Gibson simulator for interactive navigation.

Visual navigation with topological maps using graph neural networks for localization and pre-defined behaviors for navigation.

Learning to generate 3D shapes (chairs, tables) from natural language descriptions.

Predicting a behavioral plan (turn left, turn right, etc.) from a natural language instruction and topological map.

An extension of the LSTM architecture for action recognition which learns independent hidden state transitions for individual spatial locations.

Introducing an end-to-end 3D reconstruction model that unifies single- and multi-view reconstruction.

Spatial layout estimation using convolutional neural networks.

Factored near-eye light field head-mounted displays.

Source