While current benchmark reinforcement learning (RL) tasks have been useful to drive progress in the field, they are in many ways poor substitutes for learning with real-world data. By testing increasingly complex RL algorithms on low-complexity simulation environments, we often end up with brittle RL policies that generalize poorly beyond the very specific domain. To combat this, we propose three new families of benchmark RL domains that contain some of the complexity of the natural world, while still supporting fast and extensive data acquisition. The proposed domains also permit a characterization of generalization through fair train/test separation, and easy comparison and replication of results. Through this work, we challenge the RL research community to develop more robust algorithms that meet high standards of evaluation.

Deep neural networks have shown superior performance in many regimes to remember familiar patterns with large amounts of data. However, the standard supervised deep learning paradigm is still limited when facing the need to learn new concepts efficiently from scarce data. In this paper, we present a memory-augmented neural network which is motivated by the process of human concept learning. The training procedure, imitating the concept formation course of human, learns how to distinguish samples from different classes and aggregate samples of the same kind. In order to better utilize the advantages originated from the human behavior, we propose a sequential process, during which the network should decide how to remember each sample at every step. In this sequential process, a stable and interactive memory serves as an important module. We validate our model in some typical one-shot learning tasks and also an exploratory outlier detection problem. In all the experiments, our model gets highly competitive to reach or outperform those strong baselines.

In real-world applications of natural language generation, there are often constraints on the target sentences in addition to fluency and naturalness requirements. Existing language generation techniques are usually based on recurrent neural networks (RNNs). However, it is non-trivial to impose constraints on RNNs while maintaining generation quality, since RNNs generate sentences sequentially (or with beam search) from the first word to the last. In this paper, we propose CGMH, a novel approach using Metropolis-Hastings sampling for constrained sentence generation. CGMH allows complicated constraints such as the occurrence of multiple keywords in the target sentences, which cannot be handled in traditional RNN-based approaches. Moreover, CGMH works in the inference stage, and does not require parallel corpora for training. We evaluate our method on a variety of tasks, including keywords-to-sentence generation, unsupervised sentence paraphrasing, and unsupervised sentence error correction. CGMH achieves high performance compared with previous supervised methods for sentence generation. Our code is released at https://github.com/NingMiao/CGMH

Sufficient physical activity and restful sleep play a major role in the prevention and cure of many chronic conditions. Being able to proactively screen and monitor such chronic conditions would be a big step forward for overall health. The rapid increase in the popularity of wearable devices provides a significant new source, making it possible to track the user's lifestyle real-time. In this paper, we propose a novel unsupervised representation learning technique called activity2vec that learns and "summarizes" the discrete-valued activity time-series. It learns the representations with three components: (i) the co-occurrence and magnitude of the activity levels in a time-segment, (ii) neighboring context of the time-segment, and (iii) promoting subject-invariance with adversarial training. We evaluate our method on four disorder prediction tasks using linear classifiers. Empirical evaluation demonstrates that our proposed method scales and performs better than many strong baselines. The adversarial regime helps improve the generalizability of our representations by promoting subject invariant features. We also show that using the representations at the level of a day works the best since human activity is structured in terms of daily routines

Code generation maps a program description to executable source code in a programming language. Existing approaches mainly rely on a recurrent neural network (RNN) as the decoder. However, we find that a program contains significantly more tokens than a natural language sentence, and thus it may be inappropriate for RNN to capture such a long sequence. In this paper, we propose a grammar-based structural convolutional neural network (CNN) for code generation. Our model generates a program by predicting the grammar rules of the programming language; we design several CNN modules, including the tree-based convolution and pre-order convolution, whose information is further aggregated by dedicated attentive pooling layers. Experimental results on the HearthStone benchmark dataset show that our CNN code generator significantly outperforms the previous state-of-the-art method by 5 percentage points; additional experiments on several semantic parsing tasks demonstrate the robustness of our model. We also conduct in-depth ablation test to better understand each component of our model.

This paper surveys the recent attempts, both from the machine learning and operations research communities, at leveraging machine learning to solve combinatorial optimization problems. Given the hard nature of these problems, state-of-the-art methodologies involve algorithmic decisions that either require too much computing time or are not mathematically well defined. Thus, machine learning looks like a promising candidate to effectively deal with those decisions. We advocate for pushing further the integration of machine learning and combinatorial optimization and detail methodology to do so. A main point of the paper is seeing generic optimization problems as data points and inquiring what is the relevant distribution of problems to use for learning on a given task.

In Business Intelligence, accurate predictive modeling is the key for providing adaptive decisions. We studied predictive modeling problems in this research which was motivated by real-world cases that Microsoft data scientists encountered while dealing with e-commerce transaction fraud control decisions using transaction streaming data in an uncertain probabilistic decision environment. The values of most online transactions related features can return instantly, while the true fraud labels only return after a stochastic delay. Using partially mature data directly for predictive modeling in an uncertain probabilistic decision environment would lead to significant inaccuracy on risk decision-making. To improve accurate estimation of the probabilistic prediction environment, which leads to more accurate predictive modeling, two frameworks, Current Environment Inference (CEI) and Future Environment Inference (FEI), are proposed. These frameworks generated decision environment related features using long-term fully mature and short-term partially mature data, and the values of those features were estimated using varies of learning methods, including linear regression, random forest, gradient boosted tree, artificial neural network, and recurrent neural network. Performance tests were conducted using some e-commerce transaction data from Microsoft. Testing results suggested that proposed frameworks significantly improved the accuracy of decision environment estimation.

Recent successes and advances in Deep Neural Networks (DNN) in machine vision and Natural Language Processing (NLP) have motivated their use in traditional signal processing and communications systems. In this paper, we present results of such applications to the problem of automatic modulation recognition. Variations in wireless communication channels are represented by statistical channel models and their parameterization will increase with the advent of 5G. In this paper, we report effect of simple two path channel model on our naive deep neural network based implementation. We also report impact of adversarial perturbation to the input signal.

Deep learning involves a difficult non-convex optimization problem, which is often solved by stochastic gradient (SG) methods. While SG is usually effective, it may not be robust in some situations. Recently, Newton methods have been investigated as an alternative optimization technique, but nearly all existing studies consider only fully-connected feedforward neural networks. They do not investigate other types of networks such as Convolutional Neural Networks (CNN), which are more commonly used in deep-learning applications. One reason is that Newton methods for CNN involve complicated operations, and so far no works have conducted a thorough investigation. In this work, we give details of all building blocks including function, gradient, and Jacobian evaluation, and Gauss-Newton matrix-vector products. These basic components are very important because with them further developments of Newton methods for CNN become possible. We show that an efficient MATLAB implementation can be done in just several hundred lines of code and demonstrate that the Newton method gives competitive test accuracy.

The performance of an organic photovoltaic device is intricately connected to its active layer morphology. This connection between the active layer and device performance is very expensive to evaluate, either experimentally or computationally. Hence, designing morphologies to achieve higher performances is non-trivial and often intractable. To solve this, we first introduce a deep convolutional neural network (CNN) architecture that can serve as a fast and robust surrogate for the complex structure-property map. Several tests were performed to gain trust in this trained model. Then, we utilize this fast framework to perform robust microstructural design to enhance device performance.