This paper introduces the QMDP-net, a neural network architecture for planning under partial observability. The QMDP-net combines the strengths of model-free learning and model-based planning. It is a recurrent policy network, but it represents a policy for a parameterized set of tasks by connecting a model with a planning algorithm that solves the model, thus embedding the solution structure of planning in a network learning architecture. The QMDP-net is fully differentiable and allows for end-to-end training. We train a QMDP-net on different tasks so that it can generalize to new ones in the parameterized task set and "transfer" to other similar tasks beyond the set. In preliminary experiments, QMDP-net showed strong performance on several robotic tasks in simulation. Interestingly, while QMDP-net encodes the QMDP algorithm, it sometimes outperforms the QMDP algorithm in the experiments, as a result of end-to-end learning.

Saliency methods aim to explain the predictions of deep neural networks. These methods lack reliability when the explanation is sensitive to factors that do not contribute to the model prediction. We use a simple and common pre-processing step --adding a constant shift to the input data-- to show that a transformation with no effect on the model can cause numerous methods to incorrectly attribute. In order to guarantee reliability, we posit that methods should fulfill input invariance, the requirement that a saliency method mirror the sensitivity of the model with respect to transformations of the input. We show, through several examples, that saliency methods that do not satisfy input invariance result in misleading attribution. Research to address this tension is urgently needed; reliable explanations build trust with users, help identify points of model failure and remove barriers to entry for the deployment of deep neural networks in domains like health care, security and transportation. In deep neural networks, data representation is delegated to the model and subsequently we cannot generally say in an informative way what led to a model prediction.

Abstract--This paper proposes a practical approach for automatic sleep stage classification based on a multilevel feature learning framework and Recurrent Neural Network (RNN) classifier using heart rate and wrist actigraphy derived from a wearable device. The feature learning framework is designed to extract low-and mid-level features. Low-level features capture temporal and frequency domain properties and mid-level features learn compositions and structural information of signals. Since sleep staging is a sequential problem with long-term dependencies, we take advantage of RNNs with Bidirectional Long Short-T erm Memory (BLSTM) architectures for sequence data learning. T o simulate the actual situation of daily sleep, experiments are conducted with a resting group in which sleep is recorded in resting state, and a comprehensive group in which both resting sleep and non-resting sleep are included. We evaluate the algorithm based on an eightfold cross validation to classify five sleep stages (W, N1, N2, N3, and REM). V arious comparison experiments demonstrate the effectiveness of feature learning and BLSTM. We further explore the influence of depth and width of RNNs on performance. Our method is specially proposed for wearable devices and is expected to be applicable for long-term sleep monitoring at home. Without using too much prior domain knowledge, our method has the potential to generalize sleep disorder detection. Index Terms--Heart rate, Long Short-T erm Memory, Recurrent neural networks, Sleep stage classification, Wearable device. Xin Zhang, Weixuan Kou, Y ubo Fan and Y an Xu are with the State Key Laboratory of Software Development Environment and the Key Laboratory of Biomechanics and Mechanobiology of Ministry of Education and Research Institute of Beihang University in Shenzhen and Beijing Advanced Innovation Centre for Biomedical Engineering, Beihang University, Beijing 100191, China (email: xinzhang0376@gmail.com; Eric I-Chao Chang, and Y an Xu are with Microsoft Research, Beijing 100080, China (email:echang@microsoft.com; xuyan04@gmail.com).

We consider the problem of learning a one-hidden-layer neural network: we assume the input $x\in \mathbb{R}^d$ is from Gaussian distribution and the label $y = a^\top \sigma(Bx) + \xi$, where $a$ is a nonnegative vector in $\mathbb{R}^m$ with $m\le d$, $B\in \mathbb{R}^{m\times d}$ is a full-rank weight matrix, and $\xi$ is a noise vector. We first give an analytic formula for the population risk of the standard squared loss and demonstrate that it implicitly attempts to decompose a sequence of low-rank tensors simultaneously. Inspired by the formula, we design a non-convex objective function $G(\cdot)$ whose landscape is guaranteed to have the following properties: 1. All local minima of $G$ are also global minima. 2. All global minima of $G$ correspond to the ground truth parameters. 3. The value and gradient of $G$ can be estimated using samples. With these properties, stochastic gradient descent on $G$ provably converges to the global minimum and learn the ground-truth parameters. We also prove finite sample complexity result and validate the results by simulations.

Deep neural networks (DNNs) are one of the most prominent technologies of our time, as they achieve state-of-the-art performance in many machine learning tasks, including but not limited to image classification, text mining, and speech processing. However, recent research on DNNs has indicated ever-increasing concern on the robustness to adversarial examples, especially for security-critical tasks such as traffic sign identification for autonomous driving. Studies have unveiled the vulnerability of a well-trained DNN by demonstrating the ability of generating barely noticeable (to both human and machines) adversarial images that lead to misclassification. Furthermore, researchers have shown that these adversarial images are highly transferable by simply training and attacking a substitute model built upon the target model, known as a black-box attack to DNNs. Similar to the setting of training substitute models, in this paper we propose an effective black-box attack that also only has access to the input (images) and the output (confidence scores) of a targeted DNN. However, different from leveraging attack transferability from substitute models, we propose zeroth order optimization (ZOO) based attacks to directly estimate the gradients of the targeted DNN for generating adversarial examples. We use zeroth order stochastic coordinate descent along with dimension reduction, hierarchical attack and importance sampling techniques to efficiently attack black-box models. By exploiting zeroth order optimization, improved attacks to the targeted DNN can be accomplished, sparing the need for training substitute models and avoiding the loss in attack transferability. Experimental results on MNIST, CIFAR10 and ImageNet show that the proposed ZOO attack is as effective as the state-of-the-art white-box attack and significantly outperforms existing black-box attacks via substitute models.

Collecting large training datasets, annotated with high-quality labels, is costly and time-consuming. This paper proposes a novel framework for training deep convolutional neural networks from noisy labeled datasets that can be obtained cheaply. The problem is formulated using an undirected graphical model that represents the relationship between noisy and clean labels, trained in a semi-supervised setting. In our formulation, the inference over latent clean labels is tractable and is regularized during training using auxiliary sources of information. The proposed model is applied to the image labeling problem and is shown to be effective in labeling unseen images as well as reducing label noise in training on CIFAR-10 and MS COCO datasets.

Ultrasound diagnosis is routinely used in obstetrics and gynecology for fetal biometry, and owing to its time-consuming process, there has been a great demand for automatic estimation. However, the automated analysis of ultrasound images is complicated because they are patient-specific, operator-dependent, and machine-specific. Among various types of fetal biometry, the accurate estimation of abdominal circumference (AC) is especially difficult to perform automatically because the abdomen has low contrast against surroundings, non-uniform contrast, and irregular shape compared to other parameters.We propose a method for the automatic estimation of the fetal AC from 2D ultrasound data through a specially designed convolutional neural network (CNN), which takes account of doctors' decision process, anatomical structure, and the characteristics of the ultrasound image. The proposed method uses CNN to classify ultrasound images (stomach bubble, amniotic fluid, and umbilical vein) and Hough transformation for measuring AC. We test the proposed method using clinical ultrasound data acquired from 56 pregnant women. Experimental results show that, with relatively small training samples, the proposed CNN provides sufficient classification results for AC estimation through the Hough transformation. The proposed method automatically estimates AC from ultrasound images. The method is quantitatively evaluated, and shows stable performance in most cases and even for ultrasound images deteriorated by shadowing artifacts. As a result of experiments for our acceptance check, the accuracies are 0.809 and 0.771 with the expert 1 and expert 2, respectively, while the accuracy between the two experts is 0.905. However, for cases of oversized fetus, when the amniotic fluid is not observed or the abdominal area is distorted, it could not correctly estimate AC.

Deep reinforcement learning (DRL) methods such as the Deep Q-Network (DQN) have achieved state-of-the-art results in a variety of challenging, high-dimensional domains. This success is mainly attributed to the power of deep neural networks to learn rich domain representations for approximating the value function or policy. Batch reinforcement learning methods with linear representations, on the other hand, are more stable and require less hyper parameter tuning. Yet, substantial feature engineering is necessary to achieve good results. In this work we propose a hybrid approach -- the Least Squares Deep Q-Network (LS-DQN), which combines rich feature representations learned by a DRL algorithm with the stability of a linear least squares method. We do this by periodically re-training the last hidden layer of a DRL network with a batch least squares update. Key to our approach is a Bayesian regularization term for the least squares update, which prevents over-fitting to the more recent data. We tested LS-DQN on five Atari games and demonstrate significant improvement over vanilla DQN and Double-DQN. We also investigated the reasons for the superior performance of our method. Interestingly, we found that the performance improvement can be attributed to the large batch size used by the LS method when optimizing the last layer.

If the game Cards Against Humanity and those refrigerator poetry magnets had a digital baby bestowed with machine learning, it would look something like Botnik. This Seattle-based startup is actually the comedic offspring of Jamie Brew, previously a head writer for ClickHole, a satirical website connected to The Onion, and Bob Mankoff, cartoon and humor editor of Esquire and former cartoon editor of The New Yorker. "Bob and I started Botnik after a series of long phone calls converging on the idea that comedy writing isn't a problem that an algorithm can solve," Brew said. "We didn't really care for fully automatic creativity (such as Google DeepMind's attempt to win The New Yorker Caption Contest) and were far more interested in human-machine collaboration." Botnik builds a "predictive keyboard" of words taken from various sources -- beauty ads, nature shows, famous poets, dialogue from "Seinfeld" episodes and even combinations of sources, including the unlikely triumvirate that is Beowulf/Maya Angelou/forklift manual.

Which authors of this paper are endorsers? Disable MathJax (What is MathJax?)