
ReMP: Rectified Metric Propagation for FewShot Learning
Fewshot learning features the capability of generalizing from a few exa...
read it

Improving Text Generation with StudentForcing Optimal Transport
Neural language models are often trained with maximum likelihood estimat...
read it

Repulsive Attention: Rethinking Multihead Attention as Bayesian Inference
The neural attention mechanism plays an important role in many natural l...
read it

Robust Conversational AI with Grounded Text Generation
This article presents a hybrid approach based on a Grounded Text Generat...
read it

Weakly supervised crossdomain alignment with optimal transport
Crossdomain alignment between image objects and text sequences is key t...
read it

StructureAware HumanAction Generation
Generating longrange skeletonbased human actions has been a challengin...
read it

SOLOIST: Fewshot TaskOriented Dialog with A Single Pretrained Autoregressive Model
This paper presents a new method SOLOIST, which uses transfer learning t...
read it

POINTER: Constrained Text Generation via Insertionbased Generative Pretraining
Largescale pretrained language models, such as BERT and GPT2, have ac...
read it

Oscar: ObjectSemantics Aligned Pretraining for VisionLanguage Tasks
Largescale pretraining methods of learning crossmodal representations...
read it

Optimus: Organizing Sentences via Pretrained Modeling of a Latent Space
When trained effectively, the Variational Autoencoder (VAE) can be both ...
read it

Feature Quantization Improves GAN Training
The instability in GAN training has been a longstanding problem despite...
read it

MultiView Learning for VisionandLanguage Navigation
Learning to navigate in a visual environment following natural language ...
read it

Survival Cluster Analysis
Conventional survival analysis approaches estimate risk scores or indivi...
read it

Fewshot Natural Language Generation for TaskOriented Dialog
As a crucial component in taskoriented dialog systems, the Natural Lang...
read it

Towards Learning a Generic Agent for VisionandLanguage Navigation via Pretraining
Learning to navigate in a visual environment following naturallanguage ...
read it

StraightThrough Estimator as Projected Wasserstein Gradient Flow
The StraightThrough (ST) estimator is a widely used technique for back...
read it

Robust Navigation with Language Pretraining and Stochastic Sampling
Core to the visionandlanguage navigation (VLN) challenge is building r...
read it

Implicit Deep Latent Variable Models for Text Generation
Deep latent variable models (LVM) such as variational autoencoder (VAE)...
read it

Twin Auxiliary Classifiers GAN
Conditional generative models enjoy remarkable progress over the past fe...
read it

DoubleTransfer at MEDIQA 2019: MultiSource Transfer Learning for Natural Language Understanding in the Medical Domain
This paper describes our competing system to enter the MEDIQA2019 compe...
read it

Towards Amortized RankingCritical Training for Collaborative Filtering
Collaborative filtering is widely used in modern recommender systems. Re...
read it

Cyclical Annealing Schedule: A Simple Approach to Mitigating KL Vanishing
Variational autoencoders (VAEs) with an autoregressive decoder have bee...
read it

Cyclical Stochastic Gradient MCMC for Bayesian Deep Learning
The posteriors over neural network weights are high dimensional and mult...
read it

Adversarial Learning of a Sampler Based on an Unnormalized Distribution
We investigate adversarial learning in the case when only an unnormalize...
read it

Generative Adversarial Network Training is a Continual Learning Problem
Generative Adversarial Networks (GANs) have proven to be a powerful fram...
read it

Policy Optimization as Wasserstein Gradient Flows
Policy optimization is a core component of reinforcement learning (RL), ...
read it

Baseline Needs More Love: On Simple WordEmbeddingBased Models and Associated Pooling Mechanisms
Many deep learning architectures have been proposed to model the composi...
read it

Joint Embedding of Words and Labels for Text Classification
Word embeddings are effective intermediate representations for capturing...
read it

Measuring the Intrinsic Dimension of Objective Landscapes
Many recently trained neural networks employ large numbers of parameters...
read it

Adversarial TimetoEvent Modeling
Modern health data science applications leverage abundant molecular and ...
read it

Learning Structural Weight Uncertainty for Sequential DecisionMaking
Learning probability distributions on the weights of neural networks (NN...
read it

Triangle Generative Adversarial Networks
A Triangle Generative Adversarial Network (ΔGAN) is developed for semi...
read it

Symmetric Variational Autoencoder and Connections to Adversarial Learning
A new form of the variational autoencoder (VAE) is proposed, based on th...
read it

ALICE: Towards Understanding Adversarial Learning for Joint Distribution Matching
We investigate the nonidentifiability issues associated with bidirectio...
read it

ContinuousTime Flows for Deep Generative Models
Normalizing flows have been developed recently as a method for drawing s...
read it

Scalable Bayesian Learning of Recurrent Neural Networks for Language Modeling
Recurrent neural networks (RNNs) have shown promising performance for la...
read it

Learning Generic Sentence Representations Using Convolutional Neural Networks
We propose a new encoderdecoder approach to learn distributed sentence ...
read it

Unsupervised Learning with Truncated Gaussian Graphical Models
Gaussian graphical models (GGMs) are widely used for statistical modelin...
read it

Stochastic Gradient MCMC with Stale Gradients
Stochastic gradient MCMC (SGMCMC) has played an important role in large...
read it

Variational Autoencoder for Deep Learning of Images, Labels and Captions
A novel variational autoencoder is developed to model images, as well as...
read it

Bridging the Gap between Stochastic Gradient MCMC and Stochastic Optimization
Stochastic gradient Markov chain Monte Carlo (SGMCMC) methods are Bayes...
read it

Preconditioned Stochastic Gradient Langevin Dynamics for Deep Neural Networks
Effective training of deep neural networks suffers from two main issues....
read it

HighOrder Stochastic Gradient Thermostats for Bayesian Learning of Deep Models
Learning in deep models using Bayesian methods has generated significant...
read it

A Deep Generative Deconvolutional Image Model
A deep generative model is developed for representation and analysis of ...
read it

Deep Temporal Sigmoid Belief Networks for Sequence Modeling
Deep dynamic generative models are developed to learn sequential depende...
read it
Chunyuan Li
is this you? claim profile
Research Assistant at Duke University since 2014, Research Intern at UBER 2017, Research Intern at ADOBE 2016, Guest Researcher at National Institute of Standards and Technology from 20132014, Research Intern at INRIA 2013.