Neuromorphic hardware architectures represent a growing family of potential post-Moore’s Law Era platforms. Largely due to event-driving processing inspired by the human brain, these computer platforms can offer significant energy benefits compared to traditional von Neumann processors. Unfortunately there still remains considerable difficulty in successfully programming, configuring and deploying neuromorphic systems. We present the Fugu framework as an answer to this need. Rather than necessitating a developer attain intricate knowledge of how to program and exploit spiking neural dynamics to utilize the potential benefits of neuromorphic computing, Fugu is designed to provide a higher level abstraction as a hardware-independent mechanism for linking a variety of scalable spiking neural algorithms from a variety of sources. Individual kernels linked together provide sophisticated processing through compositionality. Fugu is intended to be suitable for a wide-range of neuromorphic applications, including machine learning, scientific computing, and more brain-inspired neural algorithms. Ultimately, we hope the community adopts this and other open standardization attempts allowing for free exchange and easy implementations of the ever-growing list of spiking neural algorithms.
http://arxiv.org/abs/1905.12130
Sparse rewards are one of the most important challenges in reinforcement learning. In the single-agent setting, these challenges have been addressed by introducing intrinsic rewards that motivate agents to explore unseen regions of their state spaces. Applying these techniques naively to the multi-agent setting results in individual agents exploring independently, without any coordination among themselves. We argue that learning in cooperative multi-agent settings can be accelerated and improved if agents coordinate with respect to what they have explored. In this paper we propose an approach for learning how to dynamically select between different types of intrinsic rewards which consider not just what an individual agent has explored, but all agents, such that the agents can coordinate their exploration and maximize extrinsic returns. Concretely, we formulate the approach as a hierarchical policy where a high-level controller selects among sets of policies trained on different types of intrinsic rewards and the low-level controllers learn the action policies of all agents under these specific rewards. We demonstrate the effectiveness of the proposed approach in a multi-agent learning domain with sparse rewards.
http://arxiv.org/abs/1905.12127
Massively multi-label prediction/classification problems arise in environments like health-care or biology where very precise predictions are useful. One challenge with massively multi-label problems is that there is often a long-tailed frequency distribution for the labels, which results in few positive examples for the rare labels. We propose a solution to this problem by modifying the output layer of a neural network to create a Bayesian network of sigmoids which takes advantage of ontology relationships between the labels to help share information between the rare and the more common labels. We apply this method to the two massively multi-label tasks of disease prediction (ICD-9 codes) and protein function prediction (Gene Ontology terms) and obtain significant improvements in per-label AUROC and average precision for less common labels.
http://arxiv.org/abs/1905.12126
In recent years, pretrained word embeddings have proved useful for multimodal neural machine translation (NMT) models to address the shortage of available datasets. However, the integration of pretrained word embeddings has not yet been explored extensively. Further, pretrained word embeddings in high dimensional spaces have been reported to suffer from the hubness problem. Although some debiasing techniques have been proposed to address this problem for other natural language processing tasks, they have seldom been studied for multimodal NMT models. In this study, we examine various kinds of word embeddings and introduce two debiasing techniques for three multimodal NMT models and two language pairs – English-German translation and English-French translation. With our optimal settings, the overall performance of multimodal models was improved by up to +1.93 BLEU and +2.02 METEOR for English-German translation and +1.73 BLEU and +0.95 METEOR for English-French translation.
https://arxiv.org/abs/1905.10464
The reliable segmentation of retinal vasculature can provide the means to diagnose and monitor the progression of a variety of diseases affecting the blood vessel network, including diabetes and hypertension. We leverage the power of convolutional neural networks to devise a reliable and fully automated method that can accurately detect, segment, and analyze retinal vessels. In particular, we propose a novel, fully convolutional deep neural network with an encoder-decoder architecture that employs dilated spatial pyramid pooling with multiple dilation rates to recover the lost content in the encoder and add multiscale contextual information to the decoder. We also propose a simple yet effective way of quantifying and tracking the widths of retinal vessels through direct use of the segmentation predictions. Unlike previous deep-learning-based approaches to retinal vessel segmentation that mainly rely on patch-wise analysis, our proposed method leverages a whole-image approach during training and inference, resulting in more efficient training and faster inference through the access of global content in the image. We have tested our method on three publicly available datasets, and our state-of-the-art results on both the DRIVE and CHASE-DB1 datasets attest to the effectiveness of our approach.
http://arxiv.org/abs/1905.12120
The vast majority of processors in the world are actually microcontroller units (MCUs), which find widespread use performing simple control tasks in applications ranging from automobiles to medical devices and office equipment. The Internet of Things (IoT) promises to inject machine learning into many of these every-day objects via tiny, cheap MCUs. However, these resource-impoverished hardware platforms severely limit the complexity of machine learning models that can be deployed. For example, although convolutional neural networks (CNNs) achieve state-of-the-art results on many visual recognition tasks, CNN inference on MCUs is challenging due to severe finite memory limitations. To circumvent the memory challenge associated with CNNs, various alternatives have been proposed that do fit within the memory budget of an MCU, albeit at the cost of prediction accuracy. This paper challenges the idea that CNNs are not suitable for deployment on MCUs. We demonstrate that it is possible to automatically design CNNs which generalize well, while also being small enough to fit onto memory-limited MCUs. Our Sparse Architecture Search method combines neural architecture search with pruning in a single, unified approach, which learns superior models on four popular IoT datasets. The CNNs we find are more accurate and up to $4.35\times$ smaller than previous approaches, while meeting the strict MCU working memory constraint.
http://arxiv.org/abs/1905.12107
In many real world situations, collective decisions are made using voting. Moreover, scenarios such as committee or board elections require voting rules that return multiple winners. In multi-winner approval voting (AV), an agent may vote for as many candidates as they wish. Winners are chosen by tallying up the votes and choosing the top-$k$ candidates receiving the most votes. An agent may manipulate the vote to achieve a better outcome by voting in a way that does not reflect their true preferences. In complex and uncertain situations, agents may use heuristics to strategize, instead of incurring the additional effort required to compute the manipulation which most favors them.In this paper, we examine voting behavior in multi-winner approval voting scenarios with complete information. We show that people generally manipulate their vote to obtain a better outcome, but often do not identify the optimal manipulation. Instead, voters tend to prioritize the candidates with the highest utilities. Using simulations, we demonstrate the effectiveness of these heuristics in situations where agents only have access to partial information.
http://arxiv.org/abs/1905.12104
To learn useful dynamics on long time scales, neurons must use plasticity rules that account for long-term, circuit-wide effects of synaptic changes. In other words, neural circuits must solve a credit assignment problem to appropriately assign responsibility for global network behavior to individual circuit components. Furthermore, biological constraints demand that plasticity rules are spatially and temporally local; that is, synaptic changes can depend only on variables accessible to the pre- and postsynaptic neurons. While artificial intelligence offers a computational solution for credit assignment, namely backpropagation through time (BPTT), this solution is wildly biologically implausible. It requires both nonlocal computations and unlimited memory capacity, as any synaptic change is a complicated function of the entire history of network activity. Similar nonlocality issues plague other approaches such as FORCE (Sussillo et al. 2009). Overall, we are still missing a model for learning in recurrent circuits that both works computationally and uses only local updates. Leveraging recent advances in machine learning on approximating gradients for BPTT, we derive biologically plausible plasticity rules that enable recurrent networks to accurately learn long-term dependencies in sequential data. The solution takes the form of neurons with segregated voltage compartments, with several synaptic sub-populations that have different functional properties. The network operates in distinct phases during which each synaptic sub-population is updated by its own local plasticity rule. Our results provide new insights into the potential roles of segregated dendritic compartments, branch-specific inhibition, and global circuit phases in learning.
http://arxiv.org/abs/1905.12100
Embeddings are a fundamental component of many modern machine learning and natural language processing models. Understanding them and visualizing them is essential for gathering insights about the information they capture and the behavior of the models. State of the art in analyzing embeddings consists in projecting them in two-dimensional planes without any interpretable semantics associated to the axes of the projection, which makes detailed analyses and comparison among multiple sets of embeddings challenging. In this work, we propose to use explicit axes defined as algebraic formulae over embeddings to project them into a lower dimensional, but semantically meaningful subspace, as a simple yet effective analysis and visualization methodology. This methodology assigns an interpretable semantics to the measures of variability and the axes of visualizations, allowing for both comparisons among different sets of embeddings and fine-grained inspection of the embedding spaces. We demonstrate the power of the proposed methodology through a series of case studies that make use of visualizations constructed around the underlying methodology and through a user study. The results show how the methodology is effective at providing more profound insights than classical projection methods and how it is widely applicable to many other use cases.
http://arxiv.org/abs/1905.12099
Often times, we specify tasks for a robot using temporal language that can also span different levels of abstraction. The example command go to the kitchen before going to the second floor'' contains spatial abstraction, given thatfloor’’ consists of individual rooms that can also be referred to in isolation (“kitchen”, for example). There is also a temporal ordering of events, defined by the word “before”. Previous works have used Linear Temporal Logic (LTL) to interpret temporal language (such as “before”), and Abstract Markov Decision Processes (AMDPs) to interpret hierarchical abstractions (such as “kitchen” and “second floor”), separately. To handle both types of commands at once, we introduce the Abstract Product Markov Decision Process (AP-MDP), a novel approach capable of representing non-Markovian reward functions at different levels of abstractions. The AP-MDP framework translates LTL into its corresponding automata, creates a product Markov Decision Process (MDP) of the LTL specification and the environment MDP, and decomposes the problem into subproblems to enable efficient planning with abstractions. AP-MDP performs faster than a non-hierarchical method of solving LTL problems in over 95% of tasks, and this number only increases as the size of the environment domain increases. We also present a neural sequence-to-sequence model trained to translate language commands into LTL expression, and a new corpus of non-Markovian language commands spanning different levels of abstraction. We test our framework with the collected language commands on a drone, demonstrating that our approach enables a robot to efficiently solve temporal commands at different levels of abstraction.
http://arxiv.org/abs/1905.12096
Facial expression recognition is a major problem in the domain of artificial intelligence. One of the best ways to solve this problem is the use of convolutional neural networks (CNNs). However, a large amount of data is required to train properly these networks but most of the datasets available for facial expression recognition are relatively small. A common way to circumvent the lack of data is to use CNNs trained on large datasets of different domains and fine-tuning the layers of such networks to the target domain. However, the fine-tuning process does not preserve the memory integrity as CNNs have the tendency to forget patterns they have learned. In this paper, we evaluate different strategies of fine-tuning a CNN with the aim of assessing the memory integrity of such strategies in a cross-dataset scenario. A CNN pre-trained on a source dataset is used as the baseline and four adaptation strategies have been evaluated: fine-tuning its fully connected layers; fine-tuning its last convolutional layer and its fully connected layers; retraining the CNN on a target dataset; and the fusion of the source and target datasets and retraining the CNN. Experimental results on four datasets have shown that the fusion of the source and the target datasets provides the best trade-off between accuracy and memory integrity.
http://arxiv.org/abs/1905.12082
While imitation learning is often used in robotics, the approach frequently suffers from data mismatch and compounding errors. DAgger is an iterative algorithm that addresses these issues by aggregating training data from both the expert and novice policies, but does not consider the impact of safety. We present a probabilistic extension to DAgger, which attempts to quantify the confidence of the novice policy as a proxy for safety. Our method, EnsembleDAgger, approximates a Gaussian Process using an ensemble of neural networks. Using the variance as a measure of confidence, we compute a decision rule that captures how much we doubt the novice, thus determining when it is safe to allow the novice to act. With this approach, we aim to maximize the novice’s share of actions, while constraining the probability of failure. We demonstrate improved safety and learning performance compared to other DAgger variants and classic imitation learning on an inverted pendulum and in the MuJoCo HalfCheetah environment.
http://arxiv.org/abs/1807.08364
Knowledge bases (KBs) are the backbone of many ubiquitous applications and are thus required to exhibit high precision. However, for KBs that store subjective attributes of entities, e.g., whether a movie is “kid friendly”, simply estimating precision is complicated by the inherent ambiguity in measuring subjective phenomena. In this work, we develop a method for constructing KBs with tunable precision–i.e., KBs that can be made to operate at a specific false positive rate, despite storing both difficult-to-evaluate subjective attributes and more traditional factual attributes. The key to our approach is probabilistically modeling user consensus with respect to each entity-attribute pair, rather than modeling each pair as either True or False. Uncertainty in the model is explicitly represented and used to control the KB’s precision. We propose three neural networks for fitting the consensus model and evaluate each one on data from Google Maps–a large KB of locations and their subjective and factual attributes. The results demonstrate that our learned models are well-calibrated and thus can successfully be used to control the KB’s precision. Moreover, when constrained to maintain 95% precision, the best consensus model matches the F-score of a baseline that models each entity-attribute pair as a binary variable and does not support tunable precision. When unconstrained, our model dominates the same baseline by 12% F-score. Finally, we perform an empirical analysis of attribute-attribute correlations and show that leveraging them effectively contributes to reduced uncertainty and better performance in attribute prediction.
http://arxiv.org/abs/1905.12807
A recent strategy to circumvent the exploding and vanishing gradient problem in RNNs, and to allow the stable propagation of signals over long time scales, is to constrain recurrent connectivity matrices to be orthogonal or unitary. This ensures eigenvalues with unit norm and thus stable dynamics and training. However this comes at the cost of reduced expressivity due to the limited variety of orthogonal transformations. We propose a novel connectivity structure based on the Schur decomposition and a splitting of the Schur form into normal and non-normal parts. This allows to parametrize matrices with unit-norm eigenspectra without orthogonality constraints on eigenbases. The resulting architecture ensures access to a larger space of spectrally constrained matrices, of which orthogonal matrices are a subset. This crucial difference retains the stability advantages and training speed of orthogonal RNNs while enhancing expressivity, especially on tasks that require computations over ongoing input sequences.
http://arxiv.org/abs/1905.12080
We introduce a new method for category-level pose estimation which produces a distribution over predicted poses by integrating 3D shape estimates from a generative object model with segmentation information. Given an input depth-image of an object, our variable-time method uses a mixture density network architecture to produce a multi-modal distribution over 3DOF poses; this distribution is then combined with a prior probability encouraging silhouette agreement between the observed input and predicted object pose. Our approach significantly outperforms the current state-of-the-art in category-level 3DOF pose estimation—which outputs a point estimate and does not explicitly incorporate shape and segmentation information—as measured on the Pix3D and ShapeNet datasets.
http://arxiv.org/abs/1905.12079
Generalized planning is about finding plans that solve collections of planning instances, often infinite collections, rather than single instances. Recently it has been shown how to reduce the planning problem for generalized planning to the planning problem for a qualitative numerical problem; the latter being a reformulation that simultaneously captures all the instances in the collection. An important thread of research thus consists in finding such reformulations, or abstractions, automatically. A recent proposal learns the abstractions inductively from a finite and small sample of transitions from instances in the collection. However, as in all inductive processes, the learned abstraction is not guaranteed to be correct for the whole collection. In this work we address this limitation by performing an analysis of the abstraction with respect to the collection, and show how to obtain formal guarantees for generalization. These guarantees, in the form of first-order formulas, may be used to 1) define subcollections of instances on which the abstraction is guaranteed to be sound, 2) obtain necessary conditions for generalization under certain assumptions, and 3) do automated synthesis of complex invariants for planning problems. Our framework is general, it can be extended or combined with other approaches, and it has applications that go beyond generalized planning.
http://arxiv.org/abs/1905.12071
Abstract Meaning Representation (AMR) is a recently designed semantic representation language intended to capture the meaning of a sentence, which may be represented as a single-rooted directed acyclic graph with labeled nodes and edges. The automatic evaluation of this structure plays an important role in the development of better systems, as well as for semantic annotation. Despite there is one available metric, smatch, it has some drawbacks. For instance, smatch creates a self-relation on the root of the graph, has weights for different error types, and does not take into account the dependence of the elements in the AMR structure. With these drawbacks, smatch masks several problems of the AMR parsers and distorts the evaluation of the AMRs. In view of this, in this paper, we introduce an extended metric to evaluate AMR parsers, which deals with the drawbacks of the smatch metric. Finally, we compare both metrics, using four well-known AMR parsers, and we argue that our metric is more refined, robust, fairer, and faster than smatch.
http://arxiv.org/abs/1905.12069
Most work on sense disambiguation presumes that one knows beforehand – e.g. from a thesaurus – a set of polysemous terms. But published lists invariably give only partial coverage. For example, the English word tan has several obvious senses, but one may overlook the abbreviation for tangent. In this paper, we present an algorithm for identifying interesting polysemous terms and measuring their degree of polysemy, given an unlabeled corpus. The algorithm involves: (i) collecting all terms within a k-term window of the target term; (ii) computing the inter-term distances of the contextual terms, and reducing the multi-dimensional distance space to two dimensions using standard methods; (iii) converting the two-dimensional representation into radial coordinates and using isotonic/antitonic regression to compute the degree to which the distribution deviates from a single-peak model. The amount of deviation is the proposed polysemy index
http://arxiv.org/abs/1905.12065
In recent years, cyber-physical system (CPS) security as applied to robotic systems has become a popular research area. Mainly because robotics systems have traditionally emphasized the completion of a specific objective and lack security oriented design. Our previous work, HoneyBot \cite{celine}, presented the concept and prototype of the first software hybrid interaction honeypot specifically designed for networked robotic systems. The intuition behind HoneyBot was that it would be a remotely accessible robotic system that could simulate unsafe actions and physically perform safe actions to fool attackers. Unassuming attackers would think they were connected to an ordinary robotic system, believing their exploits were being successfully executed. All the while, the HoneyBot is logging all communications and exploits sent to be used for attacker attribution and threat model creation. In this paper, we present findings from the result of a user study performed to evaluate the effectiveness of the HoneyBot framework and architecture as it applies to real robotic systems. The user study consisted of 40 participants, was conducted over the course of several weeks, and drew from a wide range of participants aged between 18-60 with varying level of technical expertise. From the study we found that research subjects could not tell the difference between the simulated sensor values and the real sensor values coming from the HoneyBot, meaning the HoneyBot convincingly spoofed communications.
http://arxiv.org/abs/1905.12061
We present an information-theoretic approach to registration of DWI with explicit optimization over the orientational scale, with an additional focus on normalized mutual information as a robust information-theoretic similarity measure for DWI. The framework is an extension of the LOR-DWI density-based hierarchical scale-space model, that varies and optimizes over the integration, spatial, directional, and intensity scales. We extend the model to non-rigid deformations and show that the formulation provides intrinsic regularization through the orientational information. Our experiments illustrate that the proposed model deforms ODFs correctly and is capable of handling the classic complex challenges in DWI-registrations, such as the registration of fiber-crossings along with kissing, fanning and interleaving fibers. Our results clearly illustrate a novel promising regularizing effect, that comes from the nonlinear orientation-based cost function. We illustrate the properties of the different image scales, and show that including orientational information in our model make the model better at retrieving deformations compared to standard scalar-based registration.
http://arxiv.org/abs/1905.12056
Recent developments in deep domain adaptation have allowed knowledge transfer from a labeled source domain to an unlabeled target domain at the level of intermediate features or input pixels. We propose that advantages may be derived by combining them, in the form of different insights that lead to a novel design and complementary properties that result in better performance. At the feature level, inspired by insights from semi-supervised learning, we propose a classification-aware domain adversarial neural network that brings target examples into more classifiable regions of source domain. Next, we posit that computer vision insights are more amenable to injection at the pixel level. In particular, we use 3D geometry and image synthesis based on a generalized appearance flow to preserve identity across pose transformations, while using an attribute-conditioned CycleGAN to translate a single source into multiple target images that differ in lower-level properties such as lighting. Besides standard UDA benchmark, we validate on a novel and apt problem of car recognition in unlabeled surveillance images using labeled images from the web, handling explicitly specified, nameable factors of variation through pixel-level and implicit, unspecified factors through feature-level adaptation.
http://arxiv.org/abs/1803.00068
Though reinforcement learning has greatly benefited from the incorporation of neural networks, the inability to verify the correctness of such systems limits their use. Current work in explainable deep learning focuses on explaining only a single decision in terms of input features, making it unsuitable for explaining a sequence of decisions. To address this need, we introduce Abstracted Policy Graphs, which are Markov chains of abstract states. This representation concisely summarizes a policy so that individual decisions can be explained in the context of expected future transitions. Additionally, we propose a method to generate these Abstracted Policy Graphs for deterministic policies given a learned value function and a set of observed transitions, potentially off-policy transitions used during training. Since no restrictions are placed on how the value function is generated, our method is compatible with many existing reinforcement learning methods. We prove that the worst-case time complexity of our method is quadratic in the number of features and linear in the number of provided transitions, $O(|F|^2 |tr_samples|)$. By applying our method to a family of domains, we show that our method scales well in practice and produces Abstracted Policy Graphs which reliably capture relationships within these domains.
http://arxiv.org/abs/1905.12044
Despite remarkable success in image-to-image translation that celebrates the advancements of generative adversarial networks (GANs), very limited attempts are known for video domain translation. We study the task of video-to-video translation in the context of visual speech generation, where the goal is to transform an input video of any spoken word to an output video of a different word. This is a multi-domain translation, where each word forms a domain of videos uttering this word. Adaptation of the state-of-the-art image-to-image translation model (StarGAN) to this setting falls short with a large vocabulary size. Instead we propose to use character encodings of the words and design a novel character-based GANs architecture for video-to-video translation called Visual Speech GAN (ViSpGAN). We are the first to demonstrate video-to-video translation with a vocabulary of 500 words.
http://arxiv.org/abs/1905.12043
The process of identifying changes or transformations in a scene along with the ability of reasoning about their causes and effects, is a key aspect of intelligence. In this work we go beyond recent advances in computational perception, and introduce a more challenging task, Image-based Event-Sequencing (IES). In IES, the task is to predict a sequence of actions required to rearrange objects from the configuration in an input source image to the one in the target image. IES also requires systems to possess inductive generalizability. Motivated from evidence in cognitive development, we compile the first IES dataset, the Blocksworld Image Reasoning Dataset (BIRD) which contains images of wooden blocks in different configurations, and the sequence of moves to rearrange one configuration to the other. We first explore the use of existing deep learning architectures and show that these end-to-end methods under-perform in inferring temporal event-sequences and fail at inductive generalization. We then propose a modular two-step approach: Visual Perception followed by Event-Sequencing, and demonstrate improved performance by combining learning and reasoning. Finally, by showing an extension of our approach on natural images, we seek to pave the way for future research on event sequencing for real world scenes.
http://arxiv.org/abs/1905.12042
We use a quantum annealing D-Wave 2X computer to obtain solutions to NP-hard sparse coding problems. To reduce the dimensionality of the sparse coding problem to fit on the quantum D-Wave 2X hardware, we passed downsampled MNIST images through a bottleneck autoencoder. To establish a benchmark for classification performance on this reduced dimensional data set, we used an AlexNet-like architecture implemented in TensorFlow, obtaining a classification score of $94.54 \pm 0.7 \%$. As a control, we showed that the same AlexNet-like architecture produced near-state-of-the-art classification performance $(\sim 99\%)$ on the original MNIST images. To obtain a set of optimized features for inferring sparse representations of the reduced dimensional MNIST dataset, we imprinted on a random set of $47$ image patches followed by an off-line unsupervised learning algorithm using stochastic gradient descent to optimize for sparse coding. Our single-layer of sparse coding matched the stride and patch size of the first convolutional layer of the AlexNet-like deep neural network and contained $47$ fully-connected features, $47$ being the maximum number of dictionary elements that could be embedded onto the D-Wave $2$X hardware. Recent work suggests that the optimal level of sparsity corresponds to a critical value of the trade-off parameter associated with a putative second order phase transition, an observation supported by a free energy analysis of D-Wave energy states. When the sparse representations inferred by the D-Wave $2$X were passed to a linear support vector machine, we obtained a classification score of $95.68\%$. Thus, on this problem, we find that a single-layer of quantum inference is able to outperform a standard deep neural network architecture.
http://arxiv.org/abs/1905.13215
Gradient descent finds a global minimum in training deep neural networks despite the objective function being non-convex. The current paper proves gradient descent achieves zero training loss in polynomial time for a deep over-parameterized neural network with residual connections (ResNet). Our analysis relies on the particular structure of the Gram matrix induced by the neural network architecture. This structure allows us to show the Gram matrix is stable throughout the training process and this stability implies the global optimality of the gradient descent algorithm. We further extend our analysis to deep residual convolutional neural networks and obtain a similar convergence result.
http://arxiv.org/abs/1811.03804
Despite the great achievements of deep neural networks (DNNs), the vulnerability of state-of-the-art DNNs raises security concerns of DNNs in many application domains requiring high reliability.We propose the fault sneaking attack on DNNs, where the adversary aims to misclassify certain input images into any target labels by modifying the DNN parameters. We apply ADMM (alternating direction method of multipliers) for solving the optimization problem of the fault sneaking attack with two constraints: 1) the classification of the other images should be unchanged and 2) the parameter modifications should be minimized. Specifically, the first constraint requires us not only to inject designated faults (misclassifications), but also to hide the faults for stealthy or sneaking considerations by maintaining model accuracy. The second constraint requires us to minimize the parameter modifications (using L0 norm to measure the number of modifications and L2 norm to measure the magnitude of modifications). Comprehensive experimental evaluation demonstrates that the proposed framework can inject multiple sneaking faults without losing the overall test accuracy performance.
http://arxiv.org/abs/1905.12032
Image alignment across domains has recently become one of the realistic and popular topics in the research community. In this problem, a deep learning-based image alignment method is usually trained on an available largescale database. During the testing steps, this trained model is deployed on unseen images collected under different camera conditions and modalities. The delivered deep network models are unable to be updated, adapted or fine-tuned in these scenarios. Thus, recent deep learning techniques, e.g. domain adaptation, feature transferring, and fine-tuning, are unable to be deployed. This paper presents a novel deep learning based approach to tackle the problem of across unseen modalities. The proposed network is then applied to image alignment as an illustration. The proposed approach is designed as an end-to-end deep convolutional neural network to optimize the deep models to improve the performance. The proposed network has been evaluated in digit recognition when the model is trained on MNIST and then tested on unseen domain MNIST-M. Finally, the proposed method is benchmarked in image alignment problem when training on RGB images and testing on Depth and X-Ray images.
http://arxiv.org/abs/1905.12028
Recognition across domains has recently become an active topic in the research community. However, it has been largely overlooked in the problem of recognition in new unseen domains. Under this condition, the delivered deep network models are unable to be updated, adapted or fine-tuned. Therefore, recent deep learning techniques, such as: domain adaptation, feature transferring, and fine-tuning, cannot be applied. This paper presents a novel approach to the problem of domain generalization in the context of deep learning. The proposed method is evaluated on different datasets in various problems, i.e. (i) digit recognition on MNIST, SVHN and MNIST-M, (ii) face recognition on Extended Yale-B, CMU-PIE and CMU-MPIE, and (iii) pedestrian recognition on RGB and Thermal image datasets. The experimental results show that our proposed method consistently improves the performance accuracy. It can be also easily incorporated with any other CNN frameworks within an end-to-end deep network design for object detection and recognition problems to improve their performance.
http://arxiv.org/abs/1905.13040
We introduce a unified probabilistic approach for deep continual learning based on variational Bayesian inference with open set recognition. Our model combines a probabilistic encoder with a generative model and a generative linear classifier that get shared across tasks. The open set recognition bounds the approximate posterior by fitting regions of high density on the basis of correctly classified data points and balances open-space risk with recognition errors. Catastrophic inference for both generative models is significantly alleviated through generative replay, where the open set recognition is used to sample from high density areas of the class specific posterior and reject statistical outliers. Our approach naturally allows for forward and backward transfer while maintaining past knowledge without the necessity of storing old data, regularization or inferring task labels. We demonstrate compelling results in the challenging scenario of incrementally expanding the single-head classifier for both class incremental visual and audio classification tasks, as well as incremental learning of datasets across modalities.
http://arxiv.org/abs/1905.12019
In this working notes paper, we describe IBM Research AI (Almaden) team’s participation in the ImageCLEF 2019 VQA-Med competition. The challenge consists of four question-answering tasks based on radiology images. The diversity of imaging modalities, organs and disease types combined with a small imbalanced training set made this a highly complex problem. To overcome these difficulties, we implemented a modular pipeline architecture that utilized transfer learning and multi-task learning. Our findings led to the development of a novel model called Supporting Facts Network (SFN). The main idea behind SFN is to cross-utilize information from upstream tasks to improve the accuracy on harder downstream ones. This approach significantly improved the scores achieved in the validation set (18 point improvement in F-1 score). Finally, we submitted four runs to the competition and were ranked seventh.
http://arxiv.org/abs/1905.12008
We present a framework for autonomously learning a portable representation that describes a collection of low-level continuous environments. We show that these abstract representations can be learned in a task-independent egocentric space specific to the agent that, when grounded with problem-specific information, are provably sufficient for planning. We demonstrate transfer in two different domains, where an agent learns a portable, task-independent symbolic vocabulary, as well as rules expressed in that vocabulary, and then learns to instantiate those rules on a per-task basis. This reduces the number of samples required to learn a representation of a new task.
http://arxiv.org/abs/1905.12006
Biopsies are the gold standard for breast cancer diagnosis. This task can be improved by the use of Computer Aided Diagnosis (CAD) systems, reducing the time of diagnosis and reducing the inter and intra-observer variability. The advances in computing have brought this type of system closer to reality. However, datasets of Histopathological Images (HI) from biopsies are quite small and unbalanced what makes difficult to use modern machine learning techniques such as deep learning. In this paper we propose a compact architecture based on texture filters that has fewer parameters than traditional deep models but is able to capture the difference between malignant and benign tissues with relative accuracy. The experimental results on the BreakHis dataset have show that the proposed texture CNN achieves almost 90% of accuracy for classifying benign and malignant tissues.
http://arxiv.org/abs/1905.12005
In this paper, the concept of representation learning based on deep neural networks is applied as an alternative to the use of handcrafted features in a method for automatic visual inspection of corroded thermoelectric metallic pipes. A texture convolutional neural network (TCNN) replaces handcrafted features based on Local Phase Quantization (LPQ) and Haralick descriptors (HD) with the advantage of learning an appropriate textural representation and the decision boundaries into a single optimization process. Experimental results have shown that it is possible to reach the accuracy of 99.20% in the task of identifying different levels of corrosion in the internal surface of thermoelectric pipe walls, while using a compact network that requires much less effort in tuning parameters when compared to the handcrafted approach since the TCNN architecture is compact regarding the number of layers and connections. The observed results open up the possibility of using deep neural networks in real-time applications such as the automatic inspection of thermoelectric metal pipes.
http://arxiv.org/abs/1905.12003
In this work, we develop a novel regularizer to improve the learning of long-range dependency of sequence data. Applied on language modelling, our regularizer expresses the inductive bias that sequence variables should have high mutual information even though the model might not see abundant observations for complex long-range dependency. We show how the `next sentence prediction (classification)’ heuristic can be derived in a principled way from our mutual information estimation framework, and be further extended to maximize the mutual information of sequence variables. The proposed approach not only is effective at increasing the mutual information of segments under the learned model but more importantly, leads to a higher likelihood on holdout data, and improved generation quality.
https://arxiv.org/abs/1905.11978
Existing controllable text generation systems rely on annotated attributes, which greatly limits their capabilities and applications. In this work, we make the first successful attempt to use VAEs to achieve controllable text generation without supervision. We do so by decomposing the latent space of the VAE into two parts: one incorporates structural constraints to capture dominant global variations implicitly present in the data, e.g., sentiment or topic; the other is unstructured and is used for the reconstruction of the source sentences. With the enforced structural constraint, the underlying global variations will be discovered and disentangled during the training of the VAE. The structural constraint also provides a natural recipe for mitigating posterior collapse for the structured part, which cannot be fully resolved by the existing techniques. On the task of text style transfer, our unsupervised approach achieves significantly better performance than previous supervised approaches. By showcasing generation with finer-grained control including Cards-Against-Humanity-style topic transitions within a sentence, we demonstrate that our model can perform controlled text generation in a more flexible way than existing methods.
https://arxiv.org/abs/1905.11975
Deep neural networks are vulnerable to adversarial attacks. The literature is rich with algorithms that can easily craft successful adversarial examples. In contrast, the performance of defense techniques still lags behind. This paper proposes ME-Net, a defense method that leverages matrix estimation (ME). In ME-Net, images are preprocessed using two steps: first pixels are randomly dropped from the image; then, the image is reconstructed using ME. We show that this process destroys the adversarial structure of the noise, while re-enforcing the global structure in the original image. Since humans typically rely on such global structures in classifying images, the process makes the network mode compatible with human perception. We conduct comprehensive experiments on prevailing benchmarks such as MNIST, CIFAR-10, SVHN, and Tiny-ImageNet. Comparing ME-Net with state-of-the-art defense mechanisms shows that ME-Net consistently outperforms prior techniques, improving robustness against both black-box and white-box attacks.
https://arxiv.org/abs/1905.11971
Music Genre Classification is the problem of associating genre-related labels to digitized music tracks. It has applications in the organization of commercial and personal music collections. Often, music tracks are described as a set of timbre-inspired sound textures. In shallow-learning systems, the total number of sound textures per track is usually too high, and texture downsampling is necessary to make training tractable. Although previous work has solved this by linear downsampling, no extensive work has been done to evaluate how texture selection benefits genre classification in the context of the bag of frames track descriptions. In this paper, we evaluate the impact of frame selection on automatic music genre classification in a bag of frames scenario. We also present a novel texture selector based on K-Means aimed to identify diverse sound textures within each track. We evaluated texture selection in diverse datasets, four different feature sets, as well as its relationship to a univariate feature selection strategy. The results show that frame selection leads to significant improvement over the single vector baseline on datasets consisting of full-length tracks, regardless of the feature set. Results also indicate that the K-Means texture selector achieves significant improvements over the baseline, using fewer textures per track than the commonly used linear downsampling. The results also suggest that texture selection is complementary to the feature selection strategy evaluated. Our qualitative analysis indicates that texture variety within classes benefits model generalization. Our analysis shows that selecting specific audio excerpts can improve classification performance, and it can be done automatically.
https://arxiv.org/abs/1905.11959
Because of the rich dynamical structure of videos and their ubiquity in everyday life, it is a natural idea that video data could serve as a powerful unsupervised learning signal for training visual representations in deep neural networks. However, instantiating this idea, especially at large scale, has remained a significant artificial intelligence challenge. Here we present the Video Instance Embedding (VIE) framework, which extends powerful recent unsupervised loss functions for learning deep nonlinear embeddings to multi-stream temporal processing architectures on large-scale video datasets. We show that VIE-trained networks substantially advance the state of the art in unsupervised learning from video datastreams, both for action recognition in the Kinetics dataset, and object recognition in the ImageNet dataset. We show that a hybrid model with both static and dynamic processing pathways is optimal for both transfer tasks, and provide analyses indicating how the pathways differ. Taken in context, our results suggest that deep neural embeddings are a promising approach to unsupervised visual learning across a wide variety of domains.
https://arxiv.org/abs/1905.11954
Recent work addressing model reliability and generalization has resulted in a variety of methods that seek to proactively address differences between the training and unknown target environments. While most methods achieve this by finding distributions that will be invariant across environments, we will show they do not necessarily find the same distributions which has implications for performance. In this paper we unify existing work on prediction using stable distributions by relating environmental shifts to edges in the graph underlying a prediction problem, and characterize stable distributions as those which effectively remove these edges. We then quantify the effect of edge deletion on performance in the linear case and corroborate the findings in a simulated and real data experiment.
https://arxiv.org/abs/1905.11374
This work presents a deep object co-segmentation (DOCS) approach for segmenting common objects of the same class within a pair of images. This means that the method learns to ignore common, or uncommon, background stuff and focuses on objects. If multiple object classes are presented in the image pair, they are jointly extracted as foreground. To address this task, we propose a CNN-based Siamese encoder-decoder architecture. The encoder extracts high-level semantic features of the foreground objects, a mutual correlation layer detects the common objects, and finally, the decoder generates the output foreground masks for each image. To train our model, we compile a large object co-segmentation dataset consisting of image pairs from the PASCAL VOC dataset with common objects masks. We evaluate our approach on commonly used datasets for co-segmentation tasks and observe that our approach consistently outperforms competing methods, for both seen and unseen object classes.
http://arxiv.org/abs/1804.06423
Convolutional Neural Networks (ConvNets) are commonly developed at a fixed resource budget, and then scaled up for better accuracy if more resources are available. In this paper, we systematically study model scaling and identify that carefully balancing network depth, width, and resolution can lead to better performance. Based on this observation, we propose a new scaling method that uniformly scales all dimensions of depth/width/resolution using a simple yet highly effective compound coefficient. We demonstrate the effectiveness of this method on scaling up MobileNets and ResNet. To go even further, we use neural architecture search to design a new baseline network and scale it up to obtain a family of models, called EfficientNets, which achieve much better accuracy and efficiency than previous ConvNets. In particular, our EfficientNet-B7 achieves state-of-the-art 84.4% top-1 / 97.1% top-5 accuracy on ImageNet, while being 8.4x smaller and 6.1x faster on inference than the best existing ConvNet. Our EfficientNets also transfer well and achieve state-of-the-art accuracy on CIFAR-100 (91.7%), Flowers (98.8%), and 3 other transfer learning datasets, with an order of magnitude fewer parameters. Source code is at this https URL.
https://arxiv.org/abs/1905.11946
In recent years, histopathology images have been increasingly used as a diagnostic tool in the medical field. The process of accurately diagnosing a biopsy sample requires significant expertise in the field, and as such can be time-consuming and is prone to uncertainty and error. With the advent of digital pathology, using image recognition systems to highlight problem areas or locate similar images can aid pathologists in making quick and accurate diagnoses. In this paper, we specifically consider the encoded local projections (ELP) algorithm, which has previously shown some success as a tool for classification and recognition of histopathology images. We build on the success of the ELP algorithm as a means for image classification and recognition by proposing a modified algorithm which captures the local frequency information of the image. The proposed algorithm estimates local frequencies by quantifying the changes in multiple projections in local windows of greyscale images. By doing so we remove the need to store the full projections, thus significantly reducing the histogram size, and decreasing computation time for image retrieval and classification tasks. Furthermore, we investigate the effectiveness of applying our method to histopathology images which have been digitally separated into their hematoxylin and eosin stain components. The proposed algorithm is tested on the publicly available invasive ductal carcinoma (IDC) data set. The histograms are used to train an SVM to classify the data. The experiments showed that the proposed method outperforms the original ELP algorithm in image retrieval tasks. On classification tasks, the results are found to be comparable to state-of-the-art deep learning methods and better than many handcrafted features from the literature.
https://arxiv.org/abs/1905.11945
To generalize to novel visual scenes with new viewpoints and new object poses, a visual system needs representations of the shapes of the parts of an object that are invariant to changes in viewpoint or pose. 3D graphics representations disentangle visual factors such as viewpoints and lighting from object structure in a natural way. It is possible to learn to invert the process that converts 3D graphics representations into 2D images, provided the 3D graphics representations are available as labels. When only the unlabeled images are available, however, learning to derender is much harder. We consider a simple model which is just a set of free floating parts. Each part has its own relation to the camera and its own triangular mesh which can be deformed to model the shape of the part. At test time, a neural network looks at a single image and extracts the shapes of the parts and their relations to the camera. Each part can be viewed as one head of a multi-headed derenderer. During training, the extracted parts are used as input to a differentiable 3D renderer and the reconstruction error is backpropagated to train the neural net. We make the learning task easier by encouraging the deformations of the part meshes to be invariant to changes in viewpoint and invariant to the changes in the relative positions of the parts that occur when the pose of an articulated body changes. Cerberus, our multi-headed derenderer, outperforms previous methods for extracting 3D parts from single images without part annotations, and it does quite well at extracting natural parts of human figures.
https://arxiv.org/abs/1905.11940
What is the role of real-time control and learning on the formation of social conventions? To answer this question, we propose a computational model that matches human behavioral data in a social decision-making game that was analyzed both in discrete-time and continuous-time setups. Furthermore, unlike previous approaches, our model takes into account the role of sensorimotor control loops in embodied decision-making scenarios. For this purpose, we introduce the Control-based Reinforcement Learning (CRL) model. CRL is grounded in the Distributed Adaptive Control (DAC) theory of mind and brain, where low-level sensorimotor control is modulated through perceptual and behavioral learning in a layered structure. CRL follows these principles by implementing a feedback control loop handling the agent’s reactive behaviors (pre-wired reflexes), along with an adaptive layer that uses reinforcement learning to maximize long-term reward. We test our model in a multi-agent game-theoretic task in which coordination must be achieved to find an optimal solution. We show that CRL is able to reach human-level performance on standard game-theoretic metrics such as efficiency in acquiring rewards and fairness in reward distribution.
http://arxiv.org/abs/1802.06108
Adversarial training is a useful approach to promote the learning of transferable representations across the source and target domains, which has been widely applied for domain adaptation (DA) tasks based on deep neural networks. Until very recently, existing adversarial domain adaptation (ADA) methods ignore the useful information from the label space, which is an important factor accountable for the complicated data distributions associated with different semantic classes. Especially, the inter-class semantic relationships have been rarely considered and discussed in the current work of transfer learning. In this paper, we propose a novel relationship-aware adversarial domain adaptation (RADA) algorithm, which first utilizes a single multi-class domain discriminator to enforce the learning of inter-class dependency structure during domain-adversarial training and then aligns this structure with the inter-class dependencies that are characterized from training the label predictor on the source domain. Specifically, we impose a regularization term to penalize the structure discrepancy between the inter-class dependencies respectively estimated from domain discriminator and label predictor. Through this alignment, our proposed method makes the ADA aware of class relationships. Empirical studies show that the incorporation of class relationships significantly improves the performance on benchmark datasets.
https://arxiv.org/abs/1905.11931
Spiking neural networks (SNN) are artificial computational models that have been inspired by the brain’s ability to naturally encode and process information in the time domain. The added temporal dimension is believed to render them more computationally efficient than the conventional artificial neural networks, though their full computational capabilities are yet to be explored. Recently, computational memory architectures based on non-volatile memory crossbar arrays have shown great promise to implement parallel computations in artificial and spiking neural networks. In this work, we experimentally demonstrate for the first time, the feasibility to realize high-performance event-driven in-situ supervised learning systems using nanoscale and stochastic phase-change synapses. Our SNN is trained to recognize audio signals of alphabets encoded using spikes in the time domain and to generate spike trains at precise time instances to represent the pixel intensities of their corresponding images. Moreover, with a statistical model capturing the experimental behavior of the devices, we investigate architectural and systems-level solutions for improving the training and inference performance of our computational memory-based system. Combining the computational potential of supervised SNNs with the parallel compute power of computational memory, the work paves the way for next-generation of efficient brain-inspired systems.
https://arxiv.org/abs/1905.11929
In this work we present a data-driven approach for predicting the behavior of (i.e., profiling) a given non-linear audio signal processing effect (henceforth “audio effect”). Our objective is to learn a mapping function that maps the unprocessed audio to the processed by the audio effect to be profiled, using time-domain samples. To that aim, we employ a deep auto-encoder model that is conditioned on both time-domain samples and the control parameters of the target audio effect. As a test-case study, we focus on the offline profiling of two dynamic range compression audio effects, one software-based and the other analog. Compressors were chosen because they are a widely used and important set of effects and because their parameterized nonlinear time-dependent nature makes them a challenging problem for a system aiming to profile “general” audio effects. Results from our experimental procedure show that the primary functional and auditory characteristics of the compressors can be captured, however there is still sufficient audible noise to merit further investigation before such methods are applied to real-world audio processing workflows.
https://arxiv.org/abs/1905.11928
Convolution is a central operation in Convolutional Neural Networks (CNNs), which applies a kernel or mask to overlapping regions shifted across the image. In this work we show that the underlying kernels are trained with highly correlated data, which leads to co-adaptation of model weights. To address this issue we propose what we call network deconvolution, a procedure that aims to remove pixel-wise and channel-wise correlations before the data is fed into each layer. We show that by removing this correlation we are able to achieve better convergence rates during model training with superior results without the use of batch normalization on the CIFAR-10, CIFAR-100, MNIST, Fashion-MNIST datasets, as well as against reference models from “model zoo” on the ImageNet standard benchmark.
https://arxiv.org/abs/1905.11926
The GLUE benchmark (Wang et al., 2019b) is a suite of language understanding tasks which has seen dramatic progress in the past year, with average performance moving from 70.0 at launch to 83.9, state of the art at the time of writing (May 24, 2019). Here, we measure human performance on the benchmark, in order to learn whether significant headroom remains for further progress. We provide a conservative estimate of human performance on the benchmark through crowdsourcing: Our annotators are non-experts who must learn each task from a brief set of instructions and 20 examples. In spite of limited training, these annotators robustly outperform the state of the art on six of the nine GLUE tasks and achieve an average score of 87.1. Given the fast pace of progress however, the headroom we observe is quite limited. To reproduce the data-poor setting that our annotators must learn in, we also train the BERT model (Devlin et al., 2019) in limited-data regimes, and conclude that low-resource sentence classification remains a challenge for modern neural network approaches to text understanding.
https://arxiv.org/abs/1905.10425