Jacob Yeung

I recently graduated with a B.S. in Electrical Engineering and Computer Science (EECS) from UC Berkeley. My research is advised by Dr. Kristofer Bouchard and Dr. Ji Hyun Bak at the Lawrence Berkeley National Laboratory. I am also working with Medhini Narasimhan and Prof. Trevor Darrell at Berkeley AI Research (BAIR).

My primary interests lie in understanding how vision can be used to guide motor behavior and using inspiration from biology towards improving computational models. To this end, my recent research explores subspace identification in neural and behavioral data, and self-supervised learning for images.

Email  /  Github  /  CV  /  LinkedIn

We introduce external Dynamics Components Analysis (eDCA), a linear dimensionality reduction method that finds subspaces of past neural data that have maximal mutual information with subspaces of future behavior data.

We generate pseudo-labels and identify subsets of images for annotation using self-supervised learning.

We introduce a dataset with known hierarchical and compositional structure and show deep unsupervised models poorly learn the latent generative structures.

We present a novel convolutional filter pruning technique based on the Tucker decomposition that utilizes spatial information from convolutional outputs.

We present a novel visualization technique for finding relationships among histone modifications for gene expression and find the signals learned by DeepChrome are only as informative as the average histone modification counts.

Cloned from here!
Last updated: Jan 7, 2023