DeepLINK: Deep learning inference using knockoffs with applications to genomics

We propose a deep learning–based knockoffs inference framework, DeepLINK, that guarantees the false discovery rate (FDR) control in high-dimensional settings. DeepLINK is applicable to a broad class of covariate distributions described by the possibly nonlinear latent factor models. It consists of two major parts: an autoencoder network for the knockoff variable construction and a multilayer perceptron network for feature selection with the FDR control. The empirical performance of DeepLINK is investigated through extensive simulation studies, where it is shown to achieve FDR control in feature selection with both high selection power and high prediction accuracy. We also apply DeepLINK to three real data applications to demonstrate its practical utility.

A Bayesian Precision Medicine Framework for Calibrating Individualized Therapeutic Indices in Cancer

The development and clinical implementation of evidence-based precision medicine strategies has become a realistic possibility, primarily due to the rapid accumulation of large-scale genomics and pharmacological data from diverse model systems: patients, cell-lines and drug perturbation studies. We introduce a novel Bayesian modeling framework called the individualized Rapeutic index (iRx) model to integrate high-throughput pharmacogenomic data across model systems. Our iRx model achieves three main goals: first, it exploits the conserved biology between patients and cell-lines to calibrate therapeutic response of drugs in patients; second, it finds optimal cell line avatars as proxies for patient(s); and finally, it identifies key genomic drivers explaining cell line-patient similarities. This is achieved through a semi-supervised learning approach, that conflates (unsupervised) sparse latent factor models with (supervised) penalized regression techniques. We propose a unified and tractable Bayesian model for estimation, and inference is conducted via efficient posterior sampling schemes. We illustrate and validate our approach using two existing clinical trial datasets in multiple myeloma and breast cancer studies. We show that our iRx model improves prediction accuracy compared to naive alternative approaches, and it consistently outperforms existing methods in literature in both in multiple simulation scenarios as well as real clinical examples.

Dual-Embedding based Deep Latent Factor Models for Recommendation

Among various recommendation methods, latent factor models are usually considered to be state-of-the-art techniques, which aim to learn user and item embeddings for predicting user-item preferences. When applying latent factor models to the recommendation with implicit feedback, the quality of embeddings always suffers from inadequate positive feedback and noisy negative feedback. Inspired by the idea of NSVD that represents users based on their interacted items, this article proposes a dual-embedding based deep latent factor method for recommendation with implicit feedback. In addition to learning a primitive embedding for a user (resp. item), we represent each user (resp. item) with an additional embedding from the perspective of the interacted items (resp. users) and propose attentive neural methods to discriminate the importance of interacted users/items for dual-embedding learning. We design two dual-embedding based deep latent factor models, DELF and DESEQ, for pure collaborative filtering and temporal collaborative filtering (i.e., sequential recommendation), respectively. The novel attempt of the proposed models is to capture each user-item interaction with four deep representations that are subtly fused for preference prediction. We conducted extensive experiments on four real-world datasets. The results verify the effectiveness of user/item dual embeddings and the superior performance of our methods on item recommendation.

A Survey on Adversarial Recommender Systems

Latent-factor models (LFM) based on collaborative filtering (CF), such as matrix factorization (MF) and deep CF methods, are widely used in modern recommender systems (RS) due to their excellent performance and recommendation accuracy. However, success has been accompanied with a major new arising challenge: Many applications of machine learning (ML) are adversarial in nature [146]. In recent years, it has been shown that these methods are vulnerable to adversarial examples, i.e., subtle but non-random perturbations designed to force recommendation models to produce erroneous outputs. The goal of this survey is two-fold: (i) to present recent advances on adversarial machine learning (AML) for the security of RS (i.e., attacking and defense recommendation models) and (ii) to show another successful application of AML in generative adversarial networks (GANs) for generative applications, thanks to their ability for learning (high-dimensional) data distributions. In this survey, we provide an exhaustive literature review of 76 articles published in major RS and ML journals and conferences. This review serves as a reference for the RS community working on the security of RS or on generative models using GANs to improve their quality.

Reliability in Latent Growth Curve Models

In this chapter I give an overview of reliability in the context of latent growth curve models. Although conceptually similar to latent factor models, reliability is more complex in a growth context. Two different conceptions of reliability were compared: growth curve reliability and growth rate reliability. The former evaluates the reliability of a measurement at a specific time point, the latter evaluates the reliability of the estimate of change over time. The differences, strengths, and limitations of these approaches are discussed, and demonstrated with an example on memory change in older adults. I show that using multiple measurements at each time point (with a second-order latent growth curve model) can improve growth curve reliability but will not necessarily improve growth rate reliability.

Estimating Endogenous Treatment Effects Using Latent Factor Models with and without Instrumental Variables

We provide evidence on the least biased ways to identify causal effects in situations where there are multiple outcomes that all depend on the same endogenous regressor and a reasonable but potentially contaminated instrumental variable that is available. Simulations provide suggestive evidence on the complementarity of instrumental variable (IV) and latent factor methods and how this complementarity depends on the number of outcome variables and the degree of contamination in the IV. We apply the causal inference methods to assess the impact of mental illness on work absenteeism and disability, using the National Comorbidity Survey Replication.