Training deep neural networks via direct loss minimization
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
Training deep neural networks via direct loss minimization
Y Song, AG Schwing, RS Zemel… - Proceedings of the 33rd …, 2016 - dl.acm.org
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
[PDF][PDF] Training Deep Neural Networks via Direct Loss Minimization
Y Song, T EDU, AG Schwing, T EDU, RS Zemel… - cs.toronto.edu
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
[PDF][PDF] Training Deep Neural Networks via Direct Loss Minimization
Y Song, T EDU, AG Schwing, T EDU, RS Zemel… - utstat.toronto.edu
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
[PDF][PDF] Training Deep Neural Networks via Direct Loss Minimization
Y Song, T EDU, AG Schwing, T EDU, RS Zemel… - utstat.utoronto.ca
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
Training Deep Neural Networks via Direct Loss Minimization
Y Song, AG Schwing, RS Zemel, R Urtasun - arXiv e-prints, 2015 - ui.adsabs.harvard.edu
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
[PDF][PDF] Training Deep Neural Networks via Direct Loss Minimization
Y Song, T EDU, AG Schwing, T EDU, RS Zemel… - utstat.toronto.edu
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
[PDF][PDF] Training Deep Neural Networks via Direct Loss Minimization
Y Song, T EDU, AG Schwing, T EDU, RS Zemel… - jmlr.csail.mit.edu
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
Training Deep Neural Networks via Direct Loss Minimization
Y Song, AG Schwing, RS Zemel, R Urtasun - arXiv preprint arXiv …, 2015 - arxiv.org
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …
[PDF][PDF] Training Deep Neural Networks via Direct Loss Minimization
Y Song, T EDU, AG Schwing, T EDU, RS Zemel… - alexander-schwing.de
Supervised training of deep neural nets typically relies on minimizing cross-entropy.
However, in many domains, we are interested in performing well on metrics specific to the …
However, in many domains, we are interested in performing well on metrics specific to the …