Comparing with other biometrics, gait has advantages of being unobtrusive and difficult to conceal. Inertial sensors such as accelerometer and gyroscope are often used to capture gait dynamics. Nowadays, these inertial sensors have commonly been integrated in smartphones and widely used by average person, which makes it very convenient and inexpensive to collect gait data. In this paper, we study gait recognition using smartphones in the wild. Unlike traditional methods that often require the person to walk along a specified road and/or at a normal walking speed, the proposed method collects inertial gait data under a condition of unconstraint without knowing when, where, and how the user walks. To obtain a high performance of person identification and authentication, deep-learning techniques are presented to learn and model the gait biometrics from the walking data. Specifically, a hybrid deep neural network is proposed for robust gait feature representation, where features in the space domain and in the time domain are successively abstracted by a convolutional neural network and a recurrent neural network. In the experiments, two datasets collected by smartphones on a total of 118 subjects are used for evaluations. Experiments show that the proposed method achieves over 93.5% and 93.7% accuracy in person identification and authentication, respectively.

Fake news are nowadays an issue of pressing concern, given their recent rise as a potential threat to high-quality journalism and well-informed public discourse. The Fake News Challenge (FNC-1) was organized in 2017 to encourage the development of machine learning-based classification systems for stance detection (i.e., for identifying whether a particular news article agrees, disagrees, discusses, or is unrelated to a particular news headline), thus helping in the detection and analysis of possible instances of fake news. This article presents a new approach to tackle this stance detection problem, based on the combination of string similarity features with a deep neural architecture that leverages ideas previously advanced in the context of learning efficient text representations, document classification, and natural language inference. Specifically, we use bi-directional Recurrent Neural Networks, together with max-pooling over the temporal/sequential dimension and neural attention, for representing (i) the headline, (ii) the first two sentences of the news article, and (iii) the entire news article. These representations are then combined/compared, complemented with similarity features inspired on other FNC-1 approaches, and passed to a final layer that predicts the stance of the article towards the headline. We also explore the use of external sources of information, specifically large datasets of sentence pairs originally proposed for training and evaluating natural language inference methods, in order to pre-train specific components of the neural network architecture (e.g., the RNNs used for encoding sentences). The obtained results attest to the effectiveness of the proposed ideas and show that our model, particularly when considering pre-training and the combination of neural representations together with similarity features, slightly outperforms the previous state-of-the-art.

Stochastic regularisation is an important weapon in the arsenal of a deep learning practitioner. However, despite recent theoretical advances, our understanding of how noise influences signal propagation in deep neural networks remains limited. By extending recent work based on mean field theory, we develop a new framework for signal propagation in stochastic regularised neural networks. Our noisy signal propagation theory can incorporate several common noise distributions, including additive and multiplicative Gaussian noise as well as dropout. We use this framework to investigate initialisation strategies for noisy ReLU networks. We show that no critical initialisation strategy exists using additive noise, with signal propagation exploding regardless of the selected noise distribution. For multiplicative noise (e.g. dropout), we identify alternative critical initialisation strategies that depend on the second moment of the noise distribution. Simulations and experiments on real-world data confirm that our proposed initialisation is able to stably propagate signals in deep networks, while using an initialisation disregarding noise fails to do so. Furthermore, we analyse correlation dynamics between inputs. Stronger noise regularisation is shown to reduce the depth to which discriminatory information about the inputs to a noisy ReLU network is able to propagate, even when initialised at criticality. We support our theoretical predictions for these trainable depths with simulations, as well as with experiments on MNIST and CIFAR-10

A recent line of work has uncovered a new form of data poisoning: so-called backdoor attacks. These attacks are particularly dangerous because they do not affect a network's behavior on typical, benign data. Rather, the network only deviates from its expected output when triggered by a perturbation planted by an adversary. In this paper, we identify a new property of all known backdoor attacks, which we call spectral signatures. This property allows us to utilize tools from robust statistics to thwart the attacks. We demonstrate the efficacy of these signatures in detecting and removing poisoned examples on real image sets and state of the art neural network architectures. We believe that understanding spectral signatures is a crucial first step towards designing ML systems secure against such backdoor attacks.

Starting from the Fermat's principle of least action, which governs classical and quantum mechanics and from the theory of exterior differential forms, which governs the geometry of curved manifolds, we show how to derive the equations governing neural networks in an intrinsic, coordinate invariant way, where the differential dW of the parameters W appears as the cotangent pullback of the differential of the loss function L: dW = -eta f*(dL) where f denotes the action of the network, and eta the learning rate. To be covariant, these equations imply a layer metric which is instrumental in pretraining and explains the role of conjugation when using complex numbers. The differential formalism also clarifies the relation of the gradient descent optimizer with Aristotelian and Newtonian mechanics and why large learning steps break the logic of the linearization procedure. We hope that this formal presentation of the differential geometry of neural networks will encourage some physicists to dive into deep learning, and reciprocally, that the specialists of deep learning will better appreciate the close interconnection of their subject with the foundations of classical and quantum field theory.

The developments of deep neural networks (DNN) in recent years have ushered a brand new era of artificial intelligence. DNNs are proved to be excellent in solving very complex problems, e.g., visual recognition and text understanding, to the extent of competing with or even surpassing people. Despite inspiring and encouraging success of DNNs, thorough theoretical analyses still lack to unravel the mystery of their magics. The design of DNN structure is dominated by empirical results in terms of network depth, number of neurons and activations. A few of remarkable works published recently in an attempt to interpret DNNs have established the first glimpses of their internal mechanisms. Nevertheless, research on exploring how DNNs operate is still at the initial stage with plenty of room for refinement. In this paper, we extend precedent research on neural networks with piecewise linear activations (PLNN) concerning linear regions bounds. We present (i) the exact maximal number of linear regions for single layer PLNNs; (ii) a upper bound for multi-layer PLNNs; and (iii) a tighter upper bound for the maximal number of liner regions on rectifier networks. The derived bounds also indirectly explain why deep models are more powerful than shallow counterparts, and how non-linearity of activation functions impacts on expressiveness of networks.

Question answering (QA) has achieved promising progress recently. However, answering a question in real-world scenarios like the medical domain is still challenging, due to the requirement of external knowledge and the insufficient of high-quality training data. In the light of these challenges, we study the task of generating medical QA pairs in this paper. With the insight that each medical question can be considered as a sample from the latent distribution conditioned on the corresponding answer, we propose an automated medical QA pair generation framework, consisting of an unsupervised key phrase detector that explores unstructured material for validity, and a generator that involves multi-pass decoder to integrate with structural knowledge for diversity. Series of experiments have been conducted on a real-world dataset collected from the National Medical Licensing Examination of China. Both automatic evaluation and human annotation demonstrate the effectiveness of the proposed method. Further investigation shows that, by incorporating the generated QA pairs for training, significant improvement in terms of accuracy can be achieved for the examination QA system.

This paper proposes a new architecture - Attentive Tensor Product Learning (ATPL) - to represent grammatical structures in deep learning models. ATPL is a new architecture to bridge this gap by exploiting Tensor Product Representations (TPR), a structured neural-symbolic model developed in cognitive science, aiming to integrate deep learning with explicit language structures and rules. The key ideas of ATPL are: 1) unsupervised learning of role-unbinding vectors of words via TPR-based deep neural network; 2) employing attention modules to compute TPR; and 3) integration of TPR with typical deep learning architectures including Long Short-Term Memory (LSTM) and Feedforward Neural Network (FFNN). The novelty of our approach lies in its ability to extract the grammatical structure of a sentence by using role-unbinding vectors, which are obtained in an unsupervised manner. This ATPL approach is applied to 1) image captioning, 2) part of speech (POS) tagging, and 3) constituency parsing of a sentence. Experimental results demonstrate the effectiveness of the proposed approach.

Recently, considerable research effort has been devoted to developing deep architectures for topic models to learn topic structures. Although several deep models have been proposed to learn better topic proportions of documents, how to leverage the benefits of deep structures for learning word distributions of topics has not yet been rigorously studied. Here we propose a new multi-layer generative process on word distributions of topics, where each layer consists of a set of topics and each topic is drawn from a mixture of the topics of the layer above. As the topics in all layers can be directly interpreted by words, the proposed model is able to discover interpretable topic hierarchies. As a self-contained module, our model can be flexibly adapted to different kinds of topic models to improve their modelling accuracy and interpretability. Extensive experiments on text corpora demonstrate the advantages of the proposed model.

Generative moment matching networks are introduced as quasi-random number generators for multivariate distributions. So far, quasi-random number generators for non-uniform multivariate distributions require a careful design, often need to exploit specific properties of the distribution or quasi-random number sequence under consideration, and are limited to few models. Utilizing generative neural networks, in particular, generative moment matching networks, allows one to construct quasi-random number generators for a much larger variety of multivariate distributions without such restrictions. Once trained, the presented generators only require independent quasi-random numbers as input and are thus fast in generating non-uniform multivariate quasi-random number sequences from the target distribution. Various numerical examples are considered to demonstrate the approach, including applications inspired by risk management practice.