Alert button
Picture for Kishore Konda

Kishore Konda

Alert button

Building effective deep neural network architectures one feature at a time

Oct 19, 2017
Martin Mundt, Tobias Weis, Kishore Konda, Visvanathan Ramesh

Figure 1 for Building effective deep neural network architectures one feature at a time
Figure 2 for Building effective deep neural network architectures one feature at a time
Figure 3 for Building effective deep neural network architectures one feature at a time
Figure 4 for Building effective deep neural network architectures one feature at a time
Viaarxiv icon

Dropout as data augmentation

Jan 08, 2016
Xavier Bouthillier, Kishore Konda, Pascal Vincent, Roland Memisevic

Figure 1 for Dropout as data augmentation
Figure 2 for Dropout as data augmentation
Figure 3 for Dropout as data augmentation
Figure 4 for Dropout as data augmentation
Viaarxiv icon

How far can we go without convolution: Improving fully-connected networks

Nov 09, 2015
Zhouhan Lin, Roland Memisevic, Kishore Konda

Figure 1 for How far can we go without convolution: Improving fully-connected networks
Figure 2 for How far can we go without convolution: Improving fully-connected networks
Figure 3 for How far can we go without convolution: Improving fully-connected networks
Figure 4 for How far can we go without convolution: Improving fully-connected networks
Viaarxiv icon

Zero-bias autoencoders and the benefits of co-adapting features

Apr 08, 2015
Kishore Konda, Roland Memisevic, David Krueger

Figure 1 for Zero-bias autoencoders and the benefits of co-adapting features
Figure 2 for Zero-bias autoencoders and the benefits of co-adapting features
Figure 3 for Zero-bias autoencoders and the benefits of co-adapting features
Figure 4 for Zero-bias autoencoders and the benefits of co-adapting features
Viaarxiv icon

EmoNets: Multimodal deep learning approaches for emotion recognition in video

Mar 30, 2015
Samira Ebrahimi Kahou, Xavier Bouthillier, Pascal Lamblin, Caglar Gulcehre, Vincent Michalski, Kishore Konda, Sébastien Jean, Pierre Froumenty, Yann Dauphin, Nicolas Boulanger-Lewandowski, Raul Chandias Ferrari, Mehdi Mirza, David Warde-Farley, Aaron Courville, Pascal Vincent, Roland Memisevic, Christopher Pal, Yoshua Bengio

Figure 1 for EmoNets: Multimodal deep learning approaches for emotion recognition in video
Figure 2 for EmoNets: Multimodal deep learning approaches for emotion recognition in video
Figure 3 for EmoNets: Multimodal deep learning approaches for emotion recognition in video
Figure 4 for EmoNets: Multimodal deep learning approaches for emotion recognition in video
Viaarxiv icon

Modeling sequential data using higher-order relational features and predictive training

Feb 10, 2014
Vincent Michalski, Roland Memisevic, Kishore Konda

Figure 1 for Modeling sequential data using higher-order relational features and predictive training
Figure 2 for Modeling sequential data using higher-order relational features and predictive training
Figure 3 for Modeling sequential data using higher-order relational features and predictive training
Figure 4 for Modeling sequential data using higher-order relational features and predictive training
Viaarxiv icon

Unsupervised learning of depth and motion

Dec 16, 2013
Kishore Konda, Roland Memisevic

Figure 1 for Unsupervised learning of depth and motion
Figure 2 for Unsupervised learning of depth and motion
Figure 3 for Unsupervised learning of depth and motion
Figure 4 for Unsupervised learning of depth and motion
Viaarxiv icon