Mohammad Taha Bahadori

Deep Computational Phenotyping

We apply deep learning to the problem of discovery and detection of characteristic patterns of physiology in clinical time series data. We propose two novel modifications to standard neural net training that address challenges and exploit properties that are peculiar, if not exclusive, to medical data. First, we examine a general framework for using prior knowledge to regularize parameters in the topmost layers. This framework can leverage priors of any form, ranging from formal ontologies (e.g., ICD9 codes) to data-derived similarity. Second, we describe a scalable procedure for training a collection of neural networks of different sizes but with partially shared architectures. Both of these innovations are well-suited to medical applications, where available data are not yet Internet scale and have many sparse outputs (e.g., rare diagnoses) but which have exploitable structure (e.g., temporal order and relationships between labels). However, both techniques are sufficiently general to be applied to other problems and domains. We demonstrate the empirical efficacy of both techniques on two real-world hospital data sets and show that the resulting neural nets learn interpretable and clinically relevant features.

Multi-layer Representation Learning for Medical Concepts

Proper representations of medical concepts such as diagnosis, medication, procedure codes and visits from Electronic Health Records (EHR) has broad applications in healthcare analytics. Patient EHR data consists of a sequence of visits over time, where each visit includes multiple medical concepts, e.g., diagnosis, procedure, and medication codes. This hierarchical structure provides two types of relational information, namely sequential order of visits and co-occurrence of the codes within a visit. In this work, we propose 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. In the experiments, Med2Vec shows significant improvement in prediction accuracy in clinical applications compared to baselines such as Skip-gram, GloVe, and stacked autoencoder, while providing clinically meaningful interpretation.

GRAM: Graph-based Attention Model for Healthcare Representation Learning

Deep learning methods exhibit promising performance for predictive modeling in healthcare, but two important challenges remain: - Data insufficiency: Often in healthcare predictive modeling, the sample size is insufficient for deep learning methods to achieve satisfactory results. Interpretation: The representations learned by deep learning methods should align with medical knowledge. To address these challenges, we propose GRaph-based Attention Model (GRAM) that supplements electronic health records (EHR) with hierarchical information inherent to medical ontologies. Based on the data volume and the ontology structure, GRAM represents a medical concept as a combination of its ancestors in the ontology via an attention mechanism. We compared predictive performance (i.e. accuracy, data needs, interpretability) of GRAM to various methods including the recurrent neural network (RNN) in two sequential diagnoses prediction tasks and one heart failure prediction task. 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. Additionally, unlike other methods, the medical concept representations learned by GRAM are well aligned with the medical ontology. Finally, GRAM exhibits intuitive attention behaviors by adaptively generalizing to higher level concepts when facing data insufficiency at the lower level concepts.

Learning with Minimum Supervision: A General Framework for Transductive Transfer Learning

Transductive transfer learning is one special type of transfer learning problem, in which abundant labeled examples are available in the source domain and only \textit{unlabeled} examples are available in the target domain. It easily finds applications in spam filtering, microblogging mining and so on. In this paper, we propose a general framework to solve the problem by mapping the input features in both the source domain and target domain into a shared latent space and simultaneously minimizing the feature reconstruction loss and prediction loss. We develop one specific example of the framework, namely latent large-margin transductive transfer learning (LATTL) algorithm, and analyze its theoretic bound of classification loss via Rademacher complexity. We also provide a unified view of several popular transfer learning algorithms under our framework. Experiment results on one synthetic dataset and three application datasets demonstrate the advantages of the proposed algorithm over the other state-of-the-art ones.

FBLG: a simple and effective approach for temporal dependence discovery from time series data

Discovering temporal dependence structure from multivariate time series has established its importance in many applications. We observe that when we look in reversed order of time, the temporal dependence structure of the time series is usually preserved after switching the roles of cause and effect. Inspired by this observation, we create a new time series by reversing the time stamps of original time series and combine both time series to improve the performance of temporal dependence recovery. We also provide theoretical justification for the proposed algorithm for several existing time series models. We test our approach on both synthetic and real world datasets. The experimental results confirm that this surprisingly simple approach is indeed effective under various circumstances.

Fast structure learning in generalized stochastic processes with latent factors

Understanding and quantifying the impact of unobserved processes is one of the major challenges of analyzing multivariate time series data. In this paper, we analyze a flexible stochastic process model, the generalized linear auto-regressive process (GLARP) and identify the conditions under which the impact of hidden variables appears as an additive term to the evolution matrix estimated with the maximum likelihood. In particular, we examine three examples, including two popular models for count data, i.e, Poisson and Conwey-Maxwell Poisson vector auto-regressive processes, and one powerful model for extreme value data, i.e., Gumbel vector auto-regressive processes. We demonstrate that the impact of hidden factors can be separated out via convex optimization in these three models. We also propose a fast greedy algorithm based on the selection of composite atoms in each iteration and provide a performance guarantee for it. Experiments on two synthetic datasets, one social network dataset and one climatology dataset demonstrate the the superior performance of our proposed models.

A general framework for scalable transductive transfer learning

Transductive transfer learning is one special type of transfer learning problem, in which abundant labeled examples are available in the source domain and only unlabeled examples are available in the target domain. It easily finds applications in spam filtering, microblogging mining, and so on. In this paper, we propose a general framework to solve the problem by mapping the input features in both the source domain and the target domain into a shared latent space and simultaneously minimizing the feature reconstruction loss and prediction loss. We develop one specific example of the framework, namely latent large-margin transductive transfer learning algorithm, and analyze its theoretic bound of classification loss via Rademacher complexity. We also provide a unified view of several popular transfer learning algorithms under our framework. Experiment results on one synthetic dataset and three application datasets demonstrate the advantages of the proposed algorithm over the other state-of-the-art ones.

FLASH: Fast Bayesian Optimization for Data Analytic Pipelines

Modern data science relies on data analytic pipelines to organize interdependent computational steps. Such analytic pipelines often involve different algorithms across multiple steps, each with its own hyperparameters. To achieve the best performance, it is often critical to select optimal algorithms and to set appropriate hyperparameters, which requires large computational efforts. Bayesian optimization provides a principled way for searching optimal hyperparameters for a single algorithm. However, many challenges remain in solving pipeline optimization problems with high-dimensional and highly conditional search space. In this work, we propose Fast LineAr SearcH (FLASH), an efficient method for tuning analytic pipelines. FLASH is a two-layer Bayesian optimization framework, which firstly uses a parametric model to select promising algorithms, then computes a nonparametric model to fine-tune hyperparameters of the promising algorithms. FLASH also includes an effective caching algorithm which can further accelerate the search process. Extensive experiments on a number of benchmark datasets have demonstrated that FLASH significantly outperforms previous state-of-the-art methods in both search speed and accuracy. Using 50% of the time budget, FLASH achieves up to 20% improvement on test error rate compared to the baselines. FLASH also yields state-of-the-art performance on a real-world application for healthcare predictive modeling.

Time Series Feature Learning with Applications to Health Care

Exponential growth in mobile health devices and electronic health records has resulted in a surge of large-scale time series data, which demands effective and fast machine learning models for analysis and discovery. In this chapter, we discuss a novel framework based on deep learning which automatically performs feature learning from heterogeneous time series data. It is well-suited for healthcare applications, where available data have many sparse outputs (e.g., rare diagnoses) and exploitable structures (e.g., temporal order and relationships between labels). Furthermore, we introduce a simple yet effective knowledge-distillation approach to learn an interpretable model while achieving the prediction performance of deep models. We conduct experiments on several real-world datasets and show the empirical efficacy of our framework and the interpretability of the mimic models.