Unsupervised Learning of Video Representations using LSTMs
- Nitish Srivastava, Elman Mansimov, R. Salakhutdinov
- Computer ScienceInternational Conference on Machine Learning
- 16 February 2015
This work uses Long Short Term Memory networks to learn representations of video sequences and evaluates the representations by finetuning them for a supervised learning problem - human action recognition on the UCF-101 and HMDB-51 datasets.
Deterministic Non-Autoregressive Neural Sequence Modeling by Iterative Refinement
- Jason Lee, Elman Mansimov, Kyunghyun Cho
- Computer ScienceConference on Empirical Methods in Natural…
- 19 February 2018
The proposed conditional non-autoregressive neural sequence model is evaluated on machine translation and image caption generation, and it is observed that it significantly speeds up decoding while maintaining the generation quality comparable to the autoregressive counterpart.
Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation
- Yuhuai Wu, Elman Mansimov, R. Grosse, Shu Liao, Jimmy Ba
- Computer ScienceNIPS
- 1 August 2017
This work proposes to apply trust region optimization to deep reinforcement learning using a recently proposed Kronecker-factored approximation to the curvature with trust region, which is the first scalable trust region natural gradient method for actor-critic methods.
Multi-Task Pre-Training for Plug-and-Play Task-Oriented Dialogue System
This study presents PPTOD, a unified plug-and-play model for task-oriented dialogue, and introduces a new dialogue multi-task pre-training strategy that allows the model to learn the primary TOD task completion skills from heterogeneous dialog corpora.
Generating Images from Captions with Attention
- Elman Mansimov, Emilio Parisotto, Jimmy Ba, R. Salakhutdinov
- Computer ScienceInternational Conference on Learning…
- 9 November 2015
It is demonstrated that the proposed model produces higher quality samples than other approaches and generates images with novel scene compositions corresponding to previously unseen captions in the dataset.
Molecular Geometry Prediction using a Deep Generative Graph Neural Network
- Elman Mansimov, O. Mahmood, Seokho Kang, Kyunghyun Cho
- Computer Science, ChemistryScientific Reports
- 31 March 2019
A conditional deep generative graph neural network that learns an energy function by directly learning to generate molecular conformations that are energetically favorable and more likely to be observed experimentally in data-driven manner is proposed.
A Generalized Framework of Sequence Generation with Application to Undirected Sequence Models
- Elman Mansimov, Alex Wang, Kyunghyun Cho
- Computer ScienceArXiv
- 29 May 2019
The proposed framework models the process of generation rather than the resulting sequence, and under this framework, various neural sequence models are derived as special cases, such as autoregressive, semi-autorgressive, and refinement-based non-autoregressive models, which enables to adapt decoding algorithms originally developed for directed sequence models to undirected sequence models.
Masked graph modeling for molecule generation
- O. Mahmood, Elman Mansimov, Richard Bonneau, Kyunghyun Cho
- Computer ScienceNature Communications
- 28 October 2020
A graph- based molecular generative model that outperforms previously proposed graph-based generative models of molecules and performs comparably to several SMILES-based models is proposed.
Initialization Strategies of Spatio-Temporal Convolutional Neural Networks
- Elman Mansimov, Nitish Srivastava, R. Salakhutdinov
- Computer ScienceArXiv
- 24 March 2015
We propose a new way of incorporating temporal information present in videos into Spatial Convolutional Neural Networks (ConvNets) trained on images, that avoids training Spatio-Temporal ConvNets…
Capturing document context inside sentence-level neural machine translation models with self-training
- Elman Mansimov, Gábor Melis, Lei Yu
- Computer ScienceCODI
- 11 March 2020
This work proposes an approach that doesn’t require training a specialized model on parallel document-level corpora and is applied to a trained sentence-level NMT model at decoding time and demonstrates that it has higher BLEU score and higher human preference than the baseline.