Publications

Notable publications

Deep Learning

Bioinformatics

Towards the The Latent Transcriptome
[arXiv:1810.03442]
Learning to predict if protein model is correct
Towards Gene Expression Convolutions using Gene Interaction Graphs
[ICML-WCB2017 Workshop on Computational Biology] [arXiv:1806.06975]

Bridging the Gap Between Deep Learning and Neuroscience

Generalization of Equilibrium Propagation to Vector Field Dynamics
[arXiv:1808.04873]
Equivalence of Equilibrium Propagation and Recurrent Backpropagation
[arXiv:1711.08416]
Equilibrium Propagation: Bridging the Gap between Energy-Based Models and Backpropagation
[arXiv:1602.05179]
From STDP towards Biologically Plausible Deep Learning
Towards deep learning with spiking neurons in energy based models with contrastive Hebbian plasticity
[arXiv:1612.03214]
STDP-Compatible Approximation of Backpropagation in an Energy-Based Model

Computer Vision

Relativistic f-divergence
[arXiv:1901.02474]
Distribution Matching Losses Can Hallucinate Features in Medical Image Translation
[MICCAI2018] [arXiv:1805.08841]
Count-ception: Counting by Fully Convolutional Redundant Counting
[ICCV2017 Workshop on BioImage Computing] [arXiv:1703.08710]

Deep Learning for Speech and Signals

Char2Wav: End-to-End Speech Synthesis

Deep Learning for Video

Temporal Structure In Videos
[ICCV 2015] [arXiv:1502.08029] [Aigaion]
RNNs for Emotion Recognition
[ICMI 2015] [Aigaion]
Recurrent Convolutional Networks for Video
[ICLR 2016] [arXiv:1511.06432]
Describing Multimedia Content
[IEEETM2015] [arXiv:1507.01053]
Recurrent "Grammar Cells"
[NIPS 2014]

Efficient Implementations of Deep Nets

Large Output Spaces
[NIPS 2015] [arXiv:1412.7091]

Generic Methods to Improve Training and Generalization

A Many-Objective Evolutionary Algorithm with Two Interacting Processes: Cascade Clustering and Reference Point Incremental Learning
RandomOut: Using a convolutional gradient norm to rescue convolutional filters
[ICLRW2016 Workshop] [arXiv:1602.05931]
Recombinator Networks
[CVPR 2016] [arXiv:1511.07356]
Stochastic k-Neighborhood
[ICML2013]
Toronto Paper Matching System
[ICML2013]
Framework for Paper Matching
[UAI2011]

Graph Neural Networks

Graph Attention Networks
[ICLR2018] [arXiv:1710.10903] [Aigaion]

Graphical Model

Modeling User Exposure in Recommendation
[WWW 2016] [arXiv:1510.07025]
Dynamic Poisson Factorization
[RecSys 2015] [arXiv:1509.04640]
Content-based recommendations
[NIPS 2014]
Bootstrap Collaborative Filtering
[KDD 2014]
Active Learning for Matching Problems
[ICML2012] [arXiv:1206.4647]

hardware

Bit-Slicing FPGA Accelerator for Quantized Neural Networks
[ISCAS2019]

Learning Theory and Optimization for Deep Learning

A Many-Objective Evolutionary Algorithm with Two Interacting Processes: Cascade Clustering and Reference Point Incremental Learning
Deep Learning

Medical Applications

InfoMask: Masked Variational Latent Representation to Localize Chest Disease
A Survey of Mobile Computing for the Visually Impaired
[arXiv:1811.10120]
Distribution Matching Losses Can Hallucinate Features in Medical Image Translation
[MICCAI2018] [arXiv:1805.08841]
Count-ception: Counting by Fully Convolutional Redundant Counting
[ICCV2017 Workshop on BioImage Computing] [arXiv:1703.08710]

Multi-Agent Systems

Learning by Demonstration in Repeated Stochastic Games
[AAMAS '11] [Aigaion]

Neural Language Modelling and Natural Language Processing

Ordered Neurons: Integrating Tree Structures into Recurrent Neural Networks
Event Coreference Resolution
[*SEM2018] [arXiv:1805.10985] [Aigaion]

Publications With Demo

Artistic Style Transfer
[ICLR2017] [arXiv:1610.07629] [Aigaion]
Char2Wav: End-to-End Speech Synthesis
GuessWhat?!
[CVPR2017] [arXiv:1611.08481]
Temporal Structure In Videos
[ICCV 2015] [arXiv:1502.08029] [Aigaion]

Recurrent Neural Networks

Recurrent Batch Normalization
[arXiv:1603.09025]
Regularizing RNNs
[ICLR 2016] [arXiv:1511.08400]
Latent random variables
[arXiv:1506.02216] [Aigaion]

Reinforcement Learning

Learning Causal State Representations of Partially Observable Environments
Learning Powerful Policies by Using Consistent Dynamics Mode
Variational State Encoding as Intrinsic Motivation in Reinforcement Learning
Active Domain Randomization
Hyperbolic Discounting and Learning over Multiple Horizons
Separating value functions across time-scales
Prioritizing Starting States for Reinforcement Learning
Count-Based Exploration with the Successor Representation

Reinforcement learning & Planning

Learning by Demonstration in Repeated Stochastic Games
[AAMAS '11] [Aigaion]
Hierarchical POMDP
[UAI2008] [arXiv:1206.3291]
Planning in Observable Environments
[NIPS2006]

Robotics

INFER: INtermediate representations for FuturE pRediction
Deep Generative Modeling of LiDAR Data

Unsupervised, Generative Deep Learning

Relativistic f-divergence
[arXiv:1901.02474]
The relativistic discriminator: a key element missing from standard GAN
[ICLR2019] [arXiv:1807.00734]
Artistic Style Transfer
[ICLR2017] [arXiv:1610.07629] [Aigaion]
Deep Exponential Families
[AISTATS 2015] [arXiv:1411.2581]
Generative Models
[ICLR2014] [arXiv:1311.6184] [Aigaion]