Customer scoring models are the core of scalable direct marketing. Uplift models provide an estimate of the incremental benefit from a treatment that is used for operational decision-making. Training and monitoring of uplift models require experimental data. However, the collection of data under randomized treatment assignment is costly, since random targeting deviates from an established targeting policy. To increase the cost-efficiency of experimentation and facilitate frequent data collection and model training, we introduce supervised randomization. It is a novel approach that integrates existing scoring models into randomized trials to target relevant customers, while ensuring consistent estimates of treatment effects through correction for active sample selection. An empirical Monte Carlo study shows that data collection under supervised randomization is cost-efficient, while downstream uplift models perform competitively.

Yijun Bian 1, Qingquan Song 2, Mengnan Du 2, Jun Y ao 3, Huanhuan Chen 1, Xia Hu 2 1 University of Science and Technology of China 2 Department of Computer Science and Engineering, Texas A&M University 3 Data Science and Analytics Department, WeBank Abstract Neural architecture search (NAS) is gaining more and more attention in recent years due to its flexibility and the remarkable capability of reducing the burden of neural network design. To achieve better performance, however, the searching process usually costs massive computation, which might not be affordable to researchers and practitioners. While recent attempts have employed ensemble learning methods to mitigate the enormous computation, an essential characteristic of diversity in ensemble methods is missed out, causing more similar sub-architectures to be gathered and potential redundancy in the final ensemble architecture. To bridge this gap, we propose a pruning method for NAS ensembles, named as " Sub-Architecture Ensemble Pruning in Neural Architecture Search (SAEP) ." It targets to utilize diversity and achieve sub-ensemble architectures in a smaller size with comparable performance to the unpruned ensemble architectures. Three possible solutions are proposed to decide which sub-architectures should be pruned during the searching process. Experimental results demonstrate the effectiveness of the proposed method in largely reducing the size of ensemble architectures while maintaining the final performance. Moreover, distinct deeper architectures could be discovered if the searched sub-architectures are not diverse enough. Introduction Designing neural network architectures usually requires manually elaborated architecture engineering, extensive expertise as well as expensive costs. Neural architecture search (NAS), which aims at mitigating these challenges, is attracting more and more attention recently (Zoph et al. 2018; Elsken, Metzen, and Hutter 2019; Wistuba, Rawat, and Pedapati 2019).

ADDITIONAL SHARED DECODER ON SIAMESE MUL TI-VIEW ENCODERS FOR LEARNING ACOUSTIC WORD EMBEDDINGS Myunghun Jung, Hyungjun Lim, Jahyun Goo, Y oungmoon Jung, and Hoirin Kim School of Electrical Engineering, KAIST, Daejeon, Republic of Korea ABSTRACT Acoustic word embeddings -- fixed-dimensional vector representations of arbitrary-length words -- have attracted increasing interest in query-by-example spoken term detection. Recently, on the fact that the orthography of text labels partly reflects the phonetic similarity between the words' pronunciation, a multi-view approach has been introduced that jointly learns acoustic and text embeddings. It showed that it is possible to learn discriminative embeddings by designing the objective which takes text labels as well as word segments. In this paper, we propose a network architecture that expands the multi-view approach by combining the Siamese multi-view encoders with a shared decoder network to maximize the effect of the relationship between acoustic and text em-beddings in embedding space. Discriminatively trained with multi-view triplet loss and decoding loss, our proposed approach achieves better performance on acoustic word discrimination task with the WSJ dataset, resulting in 11.1% relative improvement in average precision. Index T erms -- acoustic word embedding, query-by- example spoken term detection, multi-view learning, Siamese network, encoder-decoder 1. INTRODUCTION Query-by-example spoken term detection (QbE-STD) is the task of retrieving a spoken query from a set of speech utterances. Amazon Echo, Google Home, Apple Siri), the QbE-STD has drawn interest as a technique that can be applied to wake-up or command word detection, search engine, etc.

Exposure bias refers to the train-test discrepancy that seemingly arises when an autoregressive generative model uses only ground-truth contexts at training time but generated ones at test time. We separate the contributions of the model and the learning framework to clarify the debate on consequences and review proposed counter-measures. In this light, we argue that generalization is the underlying property to address and propose unconditional generation as its fundamental benchmark. Finally, we combine latent variable modeling with a recent formulation of exploration in reinforcement learning to obtain a rigorous handling of true and generated contexts. Results on language modeling and variational sentence auto-encoding confirm the model's generalization capability.

Copies have been proposed as a viable alternative to endow machine learning models with properties and features that adapt them to changing needs. A fundamental step of the copying process is generating an unlabelled set of points to explore the decision behavior of the targeted classifier throughout the input space. In this article we propose two sampling strategies to produce such sets. We validate them in six well-known problems and compare them with two standard methods. We evaluate our proposals in terms of both their accuracy performance and their computational cost.

B M ORE V ISUALIZATIONS We compare the results of TriMap to LargeVis in Figure 7 and 8. We also provide more visualizations obtained using TriMap in Figure 9. C D ISCUSSION We briefly discuss the results of TriMap and draw a comparison to the other methods. TriMap generally provides better global accuracy compared to the competing methods. It also successfully maintains the continuity of the underlying manifold. This can be seen from the COIL-20 result where certain clusters are located farther away from the remaining clusters. However, the underlying structure for the main cluster resembles the one provided by the other methods. TriMap also preserves the continuous structure in the Fashion MNIST and the TV News datasets. TriMap is also efficient in uncovering the possible outliers in the data. For instance, PCA reveals a large number of outliers in the Tabula Muris and the 360 K Lyrics datasets.

Machine learning in context of physical systems merits a re-examination of the learning strategy. In addition to data, one can leverage a vast library of physical prior models (e.g. kinematics, fluid flow, etc) to perform more robust inference. The nascent sub-field of \emph{physics-based learning} (PBL) studies the blending of neural networks with physical priors. While previous PBL algorithms have been applied successfully to specific tasks, it is hard to generalize existing PBL methods to a wide range of physics-based problems. Such generalization would require an architecture that can adapt to variations in the correctness of the physics, or in the quality of training data. No such architecture exists. In this paper, we aim to generalize PBL, by making a first attempt to bring neural architecture search (NAS) to the realm of PBL. We introduce a new method known as physics-based neural architecture search (PhysicsNAS) that is a top-performer across a diverse range of quality in the physical model and the dataset.

Under Review.I NTERPRETATIONS ARE USEFUL: PENALIZING EXPLA - NATIONS TO ALIGN NEURAL NETWORKS WITH PRIOR KNOWLEDGE Laura Rieger DTU Compute DTU 2800 Kgs. A BSTRACT For an explanation of a deep learning model to be effective, it must provide both insight into a model and suggest a corresponding action in order to achieve some objective. Too often, the litany of proposed explainable deep learning methods stop at the first step, providing practitioners with insight into a model, but no way to act on it. In this paper, we propose contextual decomposition explanation penalization (CDEP), a method which enables practitioners to leverage existing explanation methods in order to increase the predictive accuracy of deep learning models. In particular, when shown that a model has incorrectly assigned importance to some features, CDEP enables practitioners to correct these errors by directly regularizing the provided explanations. Using explanations provided by contextual decomposition (CD) (Murdoch et al., 2018), we demonstrate the ability of our method to increase performance on an array of toy and real datasets. However, in order to achieve that accuracy, they sometimes latch onto spurious correlations, leading to undesirable behavior as a result of dataset bias (Winkler et al., 2019), racial and ethnic stereotypes (Garg et al., 2018), or simply overfitting. While recent work into explaining neural network predictions (Murdoch et al., 2019; Doshi-V elez & Kim, 2017) has demonstrated an ability to uncover the relationships learned by a model, it is still unclear how to actually alter the model in order to remove incorrect, or undesirable, relationships. We introduce c ontextual d ecomposition e xplanation p enalization (CDEP), a method which leverages existing explanation techniques for neural networks in order to prevent a model from learning unwanted relationships and ultimately improve predictive accuracy.

In this paper, well-calibrated model uncertainty is obtained by using temperature scaling together with Monte Carlo dropout as approximation to Bayesian inference. The proposed approach can easily be derived from frequentist temperature scaling and yields well-calibrated model uncertainty as well as softmax likelihood.

In recent times, machine learning (ML) and artificial intelligence (AI) based systems have evolved and scaled across different industries such as finance, retail, insurance, energy utilities, etc. Among other things, they have been used to predict patterns of customer behavior, to generate pricing models, and to predict the return on investments. But the successes in deploying machine learning models at scale in those industries have not translated into the healthcare setting. There are multiple reasons why integrating ML models into healthcare has not been widely successful, but from a technical perspective, general-purpose commercial machine learning platforms are not a good fit for healthcare due to complexities in handling data quality issues, mandates to demonstrate clinical relevance, and a lack of ability to monitor performance in a highly regulated environment with stringent security and privacy needs. In this paper, we describe Isthmus, a turnkey, cloud-based platform which addresses the challenges above and reduces time to market for operationalizing ML/AI in healthcare. Towards the end, we describe three case studies which shed light on Isthmus capabilities. These include (1) supporting an end-to-end lifecycle of a model which predicts trauma survivability at hospital trauma centers, (2) bringing in and harmonizing data from disparate sources to create a community data platform for inferring population as well as patient level insights for Social Determinants of Health (SDoH), and (3) ingesting live-streaming data from various IoT sensors to build models, which can leverage real-time and longitudinal information to make advanced time-sensitive predictions.