Picture for Thomas Powers

Thomas Powers

University of Delaware

Differentiable Greedy Networks

Oct 30, 2018
Thomas Powers, Rasool Fakoor, Siamak Shakeri, Abhinav Sethy, Amanjit Kainth, Abdel-rahman Mohamed, Ruhi Sarikaya

Figure 1 for Differentiable Greedy Networks

Figure 2 for Differentiable Greedy Networks

Figure 3 for Differentiable Greedy Networks

Figure 4 for Differentiable Greedy Networks

* Work in progress and under review

Access Paper or Ask Questions

Deep Recurrent NMF for Speech Separation by Unfolding Iterative Thresholding

Sep 21, 2017
Scott Wisdom, Thomas Powers, James Pitton, Les Atlas

Figure 1 for Deep Recurrent NMF for Speech Separation by Unfolding Iterative Thresholding

Figure 2 for Deep Recurrent NMF for Speech Separation by Unfolding Iterative Thresholding

* To be presented at WASPAA 2017

Access Paper or Ask Questions

Interpretable Recurrent Neural Networks Using Sequential Sparse Recovery

Nov 22, 2016
Scott Wisdom, Thomas Powers, James Pitton, Les Atlas

Figure 1 for Interpretable Recurrent Neural Networks Using Sequential Sparse Recovery

Figure 2 for Interpretable Recurrent Neural Networks Using Sequential Sparse Recovery

Figure 3 for Interpretable Recurrent Neural Networks Using Sequential Sparse Recovery

* Presented at NIPS 2016 Workshop on Interpretable Machine Learning in Complex Systems

Access Paper or Ask Questions

Full-Capacity Unitary Recurrent Neural Networks

Oct 31, 2016
Scott Wisdom, Thomas Powers, John R. Hershey, Jonathan Le Roux, Les Atlas

* 9 pages, to appear in NIPS

Access Paper or Ask Questions