• Publications
  • Influence
Real-Time Adaptive Image Compression
TLDR
We present a machine learning-based approach to lossy image compression which outperforms all existing codecs, while running in real-time. Expand
Scalable Bayesian Optimization Using Deep Neural Networks
TLDR
We show that performing adaptive basis function regression with a neural network as the parametric form performs competitively with state-of-the-art GP-based approaches, but scales linearly with the number of data rather than cubically. Expand
Spectral Representations for Convolutional Neural Networks
TLDR
We propose spectral pooling, which performs dimensionality reduction by projecting onto the frequency basis set and then truncating the representation in the frequency domain. Expand
Metric Learning with Adaptive Density Discrimination
TLDR
In this work, we propose a novel approach explicitly designed to address a number of subtle yet important issues which have stymied earlier DML algorithms. Expand
Avoiding pathologies in very deep networks
TLDR
We propose to approach the design of deep architectures by examining the problem of assigning priors to nested compositions of functions. Expand
Learning Ordered Representations with Nested Dropout
TLDR
In this paper, we present results on ordered representations of data in which different dimensions have different degrees of importance. Expand
High-Dimensional Probability Estimation with Deep Density Models
TLDR
We exploit insights from deep learning to construct a bijective map to a representation space under which the transformation of the distribution of the data is approximately factorized and has identical and known marginal densities. Expand
Learned Video Compression
TLDR
We present a new algorithm for video coding, learned end-to-end for the low-latency mode. Expand
ELF-VC: Efficient Learned Flexible-Rate Video Coding
TLDR
We propose a new ML video codec, ELF-VC (Efficient, Learned and Flexible-Rate Video Coding) for the low-latency mode, which aims to improve three key weaknesses of ML-based video compression: bitrate flexibility, compression efficiency and speed. Expand
Sculpting representations for deep learning
TLDR
In machine learning, the choice of space in which to represent our data is of vital importance to their effective and efficient analysis. Expand
...
1
2
...