Epistatic Net allows the sparse spectral regularization of deep neural networks for inferring fitness functions.

Despite recent advances in high-throughput combinatorial mutagenesis assays, the number of labeled sequences available to predict molecular functions has remained small for the vastness of the sequence space combined with the ruggedness of many fitness functions. While deep neural networks (DNNs) can capture high-order epistatic interactions among the mutational sites, they tend to overfit to the small number of labeled sequences available for training. Here, we developed Epistatic Net (EN), a method for spectral regularization of DNNs that exploits evidence that epistatic interactions in many fitness functions are sparse.

CRISPRL and: Interpretable large-scale inference of DNA repair landscape based on a spectral approach.

Summary: We propose a new spectral framework for reliable training, scalable inference and interpretable explanation of the DNA repair outcome following a Cas9 cutting. Our framework, dubbed CRISPRL and, relies on an unexploited observation about the nature of the repair process: the landscape of the DNA repair is highly sparse in the (Walsh-Hadamard) spectral domain. This observation enables our framework to address key shortcomings that limit the interpretability and scaling of current deep-learning-based DNA repair models.

Multiplexed coded illumination for Fourier Ptychography with an LED array microscope.

Fourier Ptychography is a new computational microscopy technique that achieves gigapixel images with both wide field of view and high resolution in both phase and amplitude. The hardware setup involves a simple replacement of the microscope's illumination unit with a programmable LED array, allowing one to flexibly pattern illumination angles without any moving parts. In previous work, a series of low-resolution images was taken by sequentially turning on each single LED in the array, and the data were then combined to recover a bandwidth much higher than the one allowed by the original imaging system.

Robust distributed multiview video compression for wireless camera networks.

We present a novel framework for robustly delivering video data from distributed wireless camera networks that are characterized by packet drops. The main focus in this work is on robustness which is imminently needed in a wireless setting. We propose two alternative models to capture interview correlation among cameras with overlapping views.

Robust video transmission with distributed source coded auxiliary channel.

We propose a novel solution to the problem of robust, low-latency video transmission over lossy channels. Predictive video codecs, such as MPEG and H.26x, are very susceptible to prediction mismatch between encoder and decoder or "drift" when there are packet losses.

PRISM: A video coding paradigm with motion estimation at the decoder.

We describe PRISM, a video coding paradigm based on the principles of lossy distributed compression (also called source coding with side information or Wyner-Ziv coding) from multiuser information theory. PRISM represents a major departure from conventional video coding architectures (e.g.