• Publications
  • Influence
RETAIN: An Interpretable Predictive Model for Healthcare using Reverse Time Attention Mechanism
The REverse Time AttentIoN model (RETAIN) is developed for application to Electronic Health Records (EHR) data and achieves high accuracy while remaining clinically interpretable and is based on a two-level neural attention model that detects influential past visits and significant clinical variables within those visits. Expand
Explainable Prediction of Medical Codes from Clinical Text
An attentional convolutional network that predicts medical codes from clinical text using a convolutionAL neural network and uses an attention mechanism to select the most relevant segments for each of the thousands of possible codes is presented. Expand
The TPR*-Tree: An Optimized Spatio-Temporal Access Method for Predictive Queries
This paper proposes a new index structure called the TPR*- tree, which takes into account the unique features of dynamic objects through a set of improved construction algorithms and provides cost models that determine the optimal performance achievable by any data-partition spatio-temporal access method. Expand
Social influence analysis in large-scale networks
Topical Affinity Propagation (TAP) is designed with efficient distributed learning algorithms that is implemented and tested under the Map-Reduce framework and can take results of any topic modeling and the existing network structure to perform topic-level influence propagation. Expand
GraphScope: parameter-free mining of large time-evolving graphs
The efficiency and effectiveness of the GraphScope is demonstrated, which is designed to operate on large graphs, in a streaming fashion, on real datasets from several diverse domains, and produces meaningful time-evolving patterns that agree with human intuition. Expand
Multi-layer Representation Learning for Medical Concepts
This work proposes Med2Vec, which not only learns the representations for both medical codes and visits from large EHR datasets with over million visits, but also allows us to interpret the learned representations confirmed positively by clinical experts. Expand
GRAM: Graph-based Attention Model for Healthcare Representation Learning
Compared to the basic RNN, GRAM achieved 10% higher accuracy for predicting diseases rarely observed in the training data and 3% improved area under the ROC curve for predicting heart failure using an order of magnitude less training data. Expand
Temporal recommendation on graphs via long- and short-term preference fusion
This work proposes Session-based Temporal Graph (STG) which simultaneously models users' long-term and short-term preferences over time and proposes a novel recommendation algorithm Injected Preference Fusion (IPF) and extends the personalized Random Walk for temporal recommendation. Expand
Scalable Tensor Decompositions for Multi-aspect Data Mining
  • T. Kolda, Jimeng Sun
  • Mathematics, Computer Science
  • Eighth IEEE International Conference on Data…
  • 15 December 2008
Memory-Efficient Tucker (MET) is proposed, which achieves over 1000X space reduction without sacrificing speed; it also allows us to work with much larger tensors that were too big to handle before. Expand
MetaFac: community discovery via relational hypergraph factorization
The proposed MetaFac (MetaGraph Factorization), a framework that extracts community structures from various social contexts and interactions, outperform baseline methods by an order of magnitude and is able to extract meaningful communities based on the social media contexts. Expand