This paper explores the use of ambient radio frequency (RF) signals for human presence detection through deep learning. Using WiFi signal as an example, we demonstrate that the channel state information (CSI) obtained at the receiver contains rich information about the propagation environment. Through judicious pre-processing of the estimated CSI followed by deep learning, reliable presence detection can be achieved. Several challenges in passive RF sensing are addressed. With presence detection, how to collect training data with human presence can have a significant impact on the performance. This is in contrast to activity detection when a specific motion pattern is of interest. A second challenge is that RF signals are complex-valued. Handling complex-valued input in deep learning requires careful data representation and network architecture design. Finally, human presence affects CSI variation along multiple dimensions; such variation, however, is often masked by system impediments such as timing or frequency offset. Addressing these challenges, the proposed learning system uses pre-processing to preserve human motion induced channel variation while insulating against other impairments. A convolutional neural network (CNN) properly trained with both magnitude and phase information is then designed to achieve reliable presence detection. Extensive experiments are conducted. Using off-the-shelf WiFi devices, the proposed deep learning based RF sensing achieves near perfect presence detection during multiple extended periods of test and exhibits superior performance compared with leading edge passive infrared sensors. Comparison with existing RF based human presence detection also demonstrates its robustness in performance, especially when deployed in a completely new environment.
Using the raw data from consumer-level RGB-D cameras as input, we propose a deep-learning based approach to efficiently generate RGB-D images with completed information in high resolution. To process the input images in low resolution with missing regions, new operators for adaptive convolution are introduced in our deep-learning network that consists of three cascaded modules -- the completion module, the refinement module and the super-resolution module. The completion module is based on an architecture of encoder-decoder, where the features of input raw RGB-D will be automatically extracted by the encoding layers of a deep neural-network. The decoding layers are applied to reconstruct the completed depth map, which is followed by a refinement module to sharpen the boundary of different regions. For the super-resolution module, we generate RGB-D images in high resolution by multiple layers for feature extraction and a layer for up-sampling. Benefited from the adaptive convolution operators newly proposed in this paper, our results outperform the existing deep-learning based approaches for RGB-D image complete and super-resolution. As an end-to-end approach, high fidelity RGB-D images can be generated efficiently at the rate of around 21 frames per second.
Stochastic gradient descent with momentum (SGDm) is one of the most popular optimization algorithms in deep learning. While there is a rich theory of SGDm for convex problems, the theory is considerably less developed in the context of deep learning where the problem is non-convex and the gradient noise might exhibit a heavy-tailed behavior, as empirically observed in recent studies. In this study, we consider a \emph{continuous-time} variant of SGDm, known as the underdamped Langevin dynamics (ULD), and investigate its asymptotic properties under heavy-tailed perturbations. Supported by recent studies from statistical physics, we argue both theoretically and empirically that the heavy-tails of such perturbations can result in a bias even when the step-size is small, in the sense that \emph{the optima of stationary distribution} of the dynamics might not match \emph{the optima of the cost function to be optimized}. As a remedy, we develop a novel framework, which we coin as \emph{fractional} ULD (FULD), and prove that FULD targets the so-called Gibbs distribution, whose optima exactly match the optima of the original cost. We observe that the Euler discretization of FULD has noteworthy algorithmic similarities with \emph{natural gradient} methods and \emph{gradient clipping}, bringing a new perspective on understanding their role in deep learning. We support our theory with experiments conducted on a synthetic model and neural networks.
The knowledge of potentially druggable binding sites on proteins is an important preliminary step towards the discovery of novel drugs. The computational prediction of such areas can be boosted by following the recent major advances in the deep learning field and by exploiting the increasing availability of proper data. In this paper, a novel computational method for the prediction of potential binding sites is proposed, called DeepSurf. DeepSurf combines a surface-based representation, where a number of 3D voxelized grids are placed on the protein's surface, with state-of-the-art deep learning architectures. After being trained on the large database of scPDB, DeepSurf demonstrates superior results on three diverse testing datasets, by surpassing all its main deep learning-based competitors, while attaining competitive performance to a set of traditional non-data-driven approaches.
As the popularity of mobile photography is growing constantly, lots of efforts are being invested now into building complex hand-crafted camera ISP solutions. In this work, we demonstrate that even the most sophisticated ISP pipelines can be replaced with a single end-to-end deep learning model trained without any prior knowledge about the sensor and optics used in a particular device. For this, we present PyNET, a novel pyramidal CNN architecture designed for fine-grained image restoration that implicitly learns to perform all ISP steps such as image demosaicing, denoising, white balancing, color and contrast correction, demoireing, etc. The model is trained to convert RAW Bayer data obtained directly from mobile camera sensor into photos captured with a professional high-end DSLR camera, making the solution independent of any particular mobile ISP implementation. To validate the proposed approach on the real data, we collected a large-scale dataset consisting of 10 thousand full-resolution RAW-RGB image pairs captured in the wild with the Huawei P20 cameraphone (12.3 MP Sony Exmor IMX380 sensor) and Canon 5D Mark IV DSLR. The experiments demonstrate that the proposed solution can easily get to the level of the embedded P20's ISP pipeline that, unlike our approach, is combining the data from two (RGB + B/W) camera sensors. The dataset, pre-trained models and codes used in this paper are available on the project website.
We present a deep learning based methodology for extracting the singing voice signal from a musical mixture based on the underlying linguistic content. Our model follows an encoder decoder architecture and takes as input the magnitude component of the spectrogram of a musical mixture with vocals. The encoder part of the model is trained via knowledge distillation using a teacher network to learn a content embedding, which is decoded to generate the corresponding vocoder features. Using this methodology, we are able to extract the unprocessed raw vocal signal from the mixture even for a processed mixture dataset with singers not seen during training. While the nature of our system makes it incongruous with traditional objective evaluation metrics, we use subjective evaluation via listening tests to compare the methodology to state-of-the-art deep learning based source separation algorithms. We also provide sound examples and source code for reproducibility.
fastai is a deep learning library which provides practitioners with high-level components that can quickly and easily provide state-of-the-art results in standard deep learning domains, and provides researchers with low-level components that can be mixed and matched to build new approaches. It aims to do both things without substantial compromises in ease of use, flexibility, or performance. This is possible thanks to a carefully layered architecture, which expresses common underlying patterns of many deep learning and data processing techniques in terms of decoupled abstractions. These abstractions can be expressed concisely and clearly by leveraging the dynamism of the underlying Python language and the flexibility of the PyTorch library. fastai includes: a new type dispatch system for Python along with a semantic type hierarchy for tensors; a GPU-optimized computer vision library which can be extended in pure Python; an optimizer which refactors out the common functionality of modern optimizers into two basic pieces, allowing optimization algorithms to be implemented in 4-5 lines of code; a novel 2-way callback system that can access any part of the data, model, or optimizer and change it at any point during training; a new data block API; and much more. We have used this library to successfully create a complete deep learning course, which we were able to write more quickly than using previous approaches, and the code was more clear. The library is already in wide use in research, industry, and teaching. NB: This paper covers fastai v2, which is currently in pre-release at http://dev.fast.ai/
Recurrent and convolutional neural networks are the most common architectures used for time series forecasting in deep learning literature. These networks use parameter sharing by repeating a set of fixed architectures with fixed parameters over time or space. The result is that the overall architecture is time-invariant (shift-invariant in the spatial domain) or stationary. We argue that time-invariance can reduce the capacity to perform multi-step-ahead forecasting, where modelling the dynamics at a range of scales and resolutions is required. We propose ForecastNet which uses a deep feed-forward architecture to provide a time-variant model. An additional novelty of ForecastNet is interleaved outputs, which we show assist in mitigating vanishing gradients. ForecastNet is demonstrated to outperform statistical and deep learning benchmark models on several datasets.
The difficulty of deploying various deep learning (DL) models on diverse DL hardware has boosted the research and development of DL compilers in the community. Several DL compilers have been proposed from both industry and academia such as Tensorflow XLA and TVM. Similarly, the DL compilers take the DL models described in different DL frameworks as input, and then generate optimized codes for diverse DL hardware as output. However, none of the existing survey has analyzed the unique design architecture of the DL compilers comprehensively. In this paper, we perform a comprehensive survey of existing DL compilers by dissecting the commonly adopted design in details, with emphasis on the DL oriented multi-level IRs, and frontend/backend optimizations. Specifically, we provide a comprehensive comparison among existing DL compilers from various aspects. In addition, we present detailed analysis on the design of multi-level IRs and illustrate the commonly adopted optimization techniques. Finally, several insights are highlighted as the potential research directions of DL compiler. This is the first survey paper focusing on the design architecture of DL compilers, which we hope can pave the road for future research towards DL compiler.
During the past five years the Bayesian deep learning community has developed increasingly accurate and efficient approximate inference procedures that allow for Bayesian inference in deep neural networks. However, despite this algorithmic progress and the promise of improved uncertainty quantification and sample efficiency there are---as of early 2020---no publicized deployments of Bayesian neural networks in industrial practice. In this work we cast doubt on the current understanding of Bayes posteriors in popular deep neural networks: we demonstrate through careful MCMC sampling that the posterior predictive induced by the Bayes posterior yields systematically worse predictions compared to simpler methods including point estimates obtained from SGD. Furthermore, we demonstrate that predictive performance is improved significantly through the use of a "cold posterior" that overcounts evidence. Such cold posteriors sharply deviate from the Bayesian paradigm but are commonly used as heuristic in Bayesian deep learning papers. We put forward several hypotheses that could explain cold posteriors and evaluate the hypotheses through experiments. Our work questions the goal of accurate posterior approximations in Bayesian deep learning: If the true Bayes posterior is poor, what is the use of more accurate approximations? Instead, we argue that it is timely to focus on understanding the origin of the improved performance of cold posteriors.
The literature on structured prediction for NLP describes a rich collection of distributions and algorithms over sequences, segmentations, alignments, and trees; however, these algorithms are difficult to utilize in deep learning frameworks. We introduce Torch-Struct, a library for structured prediction designed to take advantage of and integrate with vectorized, auto-differentiation based frameworks. Torch-Struct includes a broad collection of probabilistic structures accessed through a simple and flexible distribution-based API that connects to any deep learning model. The library utilizes batched, vectorized operations and exploits auto-differentiation to produce readable, fast, and testable code. Internally, we also include a number of general-purpose optimizations to provide cross-algorithm efficiency. Experiments show significant performance gains over fast baselines and case-studies demonstrate the benefits of the library. Torch-Struct is available at https://github.com/harvardnlp/pytorch-struct.
Half of wavefunction information is undetected by conventional transmission electron microscopy (CTEM) as only the intensity, and not the phase, of an image is recorded. Following successful applications of deep learning to optical hologram phase recovery, we have developed neural networks to recover phases from CTEM intensities for new datasets containing 98340 exit wavefunctions. Wavefunctions were simulated with clTEM multislice propagation for 12789 materials from the Crystallography Open Database. Our networks can recover 224x224 wavefunctions in ~25 ms for a large range of physical hyperparameters and materials, and we demonstrate that performance improves as the distribution of wavefunctions is restricted. Phase recovery with deep learning overcomes the limitations of traditional methods: it is live, not susceptible to distortions, does not require microscope modification or multiple images, and can be applied to any imaging regime. This paper introduces multiple approaches to CTEM phase recovery with deep learning, and is intended to establish starting points to be improved upon by future research. Source code and links to our new datasets and pre-trained models are available at https://github.com/Jeffrey-Ede/one-shot
Successful biomedical relation extraction can provide evidence to researchers and clinicians about possible unknown associations between biomedical entities, advancing the current knowledge we have about those entities and their inherent mechanisms. Most biomedical relation extraction systems do not resort to external sources of knowledge, such as domain-specific ontologies. However, using deep learning methods, along with biomedical ontologies, has been recently shown to effectively advance the biomedical relation extraction field. To perform relation extraction, our deep learning system, BiOnt, employs four types of biomedical ontologies, namely, the Gene Ontology, the Human Phenotype Ontology, the Human Disease Ontology, and the Chemical Entities of Biological Interest, regarding gene-products, phenotypes, diseases, and chemical compounds, respectively. We tested our system with three data sets that represent three different types of relations of biomedical entities. BiOnt achieved, in F-score, an improvement of 4.93 percentage points for drug-drug interactions (DDI corpus), 4.99 percentage points for phenotype-gene relations (PGR corpus), and 2.21 percentage points for chemical-induced disease relations (BC5CDR corpus), relatively to the state-of-the-art. The code supporting this system is available at https://github.com/lasigeBioTM/BiOnt.
Hippocampus segmentation on magnetic resonance imaging is of key importance for the diagnosis, treatment decision and investigation of neuropsychiatric disorders. Automatic segmentation is an active research field, with many recent models using deep learning. Most current state-of-the art hippocampus segmentation methods train their methods on healthy or Alzheimer's disease patients from public datasets. This raises the question whether these methods are capable of recognizing the hippocampus on a different domain, that of epilepsy patients with hippocampus resection. In this paper we present a state-of-the-art, open source, ready-to-use, deep learning based hippocampus segmentation method. It uses an extended 2D multi-orientation approach, with automatic pre-processing and orientation alignment. The methodology was developed and validated using HarP, a public Alzheimer's disease hippocampus segmentation dataset. We test this methodology alongside other recent deep learning methods, in two domains: The HarP test set and an in-house epilepsy dataset, containing hippocampus resections, named HCUnicamp. We show that our method, while trained only in HarP, surpasses others from the literature in both the HarP test set and HCUnicamp in Dice. Additionally, Results from training and testing in HCUnicamp volumes are also reported separately, alongside comparisons between training and testing in epilepsy and Alzheimer's data and vice versa. Although current state-of-the-art methods, including our own, achieve upwards of 0.9 Dice in HarP, all tested methods, including our own, produced false positives in HCUnicamp resection regions, showing that there is still room for improvement for hippocampus segmentation methods when resection is involved.
Person re-identification (Re-ID) aims at retrieving a person of interest across multiple non-overlapping cameras. With the advancement of deep neural networks and increasing demand of intelligent video surveillance, it has gained significantly increased interest in the computer vision community. By dissecting the involved components in developing a person Re-ID system, we categorize it into the closed-world and open-world settings. The widely studied closed-world setting is usually applied under various research-oriented assumptions, and has achieved inspiring success using deep learning techniques on a number of datasets. We first conduct a comprehensive overview with in-depth analysis for closed-world person Re-ID from three different perspectives, including deep feature representation learning, deep metric learning and ranking optimization. With the performance saturation under closed-world setting, the research focus for person Re-ID has recently shifted to the open-world setting, facing more challenging issues. This setting is closer to practical applications under specific scenarios. We summarize the open-world Re-ID in terms of five different aspects. By analyzing the advantages of existing methods, we design a powerful AGW baseline, achieving state-of-the-art or at least comparable performance on twelve datasets for FOUR different Re-ID tasks. Meanwhile, we introduce a new evaluation metric (mINP) for person Re-ID, indicating the cost for finding all the correct matches, which provides an additional criteria to evaluate the Re-ID system for real applications. Finally, some important yet under-investigated open issues are discussed.
Limiting failures of machine learning systems is of paramount importance for safety-critical applications. In order to improve the robustness of machine learning systems, Distributionally Robust Optimization (DRO) has been proposed as a generalization of Empirical Risk Minimization (ERM). However, its use in deep learning has been severely restricted due to the relative inefficiency of the optimizers available for DRO in comparison to the wide-spread variants of Stochastic Gradient Descent (SGD) optimizers for ERM. We propose SGD with hardness weighted sampling, a principled and efficient optimization method for DRO in machine learning that is particularly suited in the context of deep learning. Similar to a hard example mining strategy in practice, the proposed algorithm is straightforward to implement and computationally as efficient as SGD-based optimizers used for deep learning, requiring minimal overhead computation. In contrast to typical ad hoc hard mining approaches, we prove the convergence of our DRO algorithm for over-parameterized deep learning networks with ReLU activation and a finite number of layers and parameters. Our experiments on fetal brain 3D MRI segmentation and brain tumor segmentation in MRI demonstrate the feasibility and the usefulness of our approach. Using our hardness weighted sampling for training a state-of-the-art deep learning pipeline leads to improved robustness to anatomical variabilities in automatic fetal brain 3D MRI segmentation using deep learning and to improved robustness to the image protocol variations in brain tumor segmentation. Our code is available at https://github.com/LucasFidon/HardnessWeightedSampler.
Free-hand sketches are highly illustrative, and have been widely used by humans to depict objects or stories from ancient times to the present. The recent prevalence of touchscreen devices has made sketch creation a much easier task than ever and consequently made sketch-oriented applications increasingly popular. The progress of deep learning has immensely benefited free-hand sketch research and applications. This paper presents a comprehensive survey of the deep learning techniques oriented at free-hand sketch data, and the applications that they enable. The main contents of this survey include: (i) A discussion of the intrinsic traits and unique challenges of free-hand sketch, to highlight the essential differences between sketch data and other data modalities, e.g., natural photos. (ii) A review of the developments of free-hand sketch research in the deep learning era, by surveying existing datasets, research topics, and the state-of-the-art methods through a detailed taxonomy and experimental evaluation. (iii) Promotion of future work via a discussion of bottlenecks, open problems, and potential research directions for the community.
Rich phenomena from complex systems have long intrigued researchers, and yet modeling system micro-dynamics and inferring the forms of interaction remain challenging for conventional data-driven approaches, being generally established by human scientists. In this study, we propose AgentNet, a model-free data-driven framework consisting of deep neural networks to reveal and analyze the hidden interactions in complex systems from observed data alone. AgentNet utilizes a graph attention network with novel variable-wise attention to model the interaction between individual agents, and employs various encoders and decoders that can be selectively applied to any desired system. Our model successfully captured a wide variety of simulated complex systems, namely cellular automata (discrete), the Vicsek model (continuous), and active Ornstein--Uhlenbeck particles (non-Markovian) in which, notably, AgentNet's visualized attention values coincided with the true interaction strength and exhibited collective behavior that was absent in the training data. A demonstration with empirical data from a flock of birds showed that AgentNet could identify hidden interaction ranges exhibited by real birds, which cannot be detected by conventional velocity correlation analysis. We expect our framework to open a novel path to investigating complex systems and to provide insight into general process-driven modeling.
Deep learning as represented by the artificial deep neural networks (DNNs) has achieved great success in many important areas that deal with text, images, videos, graphs, and so on. However, the black-box nature of DNNs has become one of the primary obstacles for their wide acceptance in mission-critical applications such as medical diagnosis and therapy. Due to the huge potential of deep learning, interpreting neural networks has recently attracted much research attention. In this paper, based on our comprehensive taxonomy, we systematically review recent studies in understanding the mechanism of neural networks, describe applications of interpretability especially in medicine, and discuss future directions of interpretability research, such as in relation to fuzzy logic and brain science.
Automated medical image segmentation is an important step in many medical procedures. Recently, deep learning networks have been widely used for various medical image segmentation tasks, with U-Net and generative adversarial nets (GANs) being some of the commonly used ones. Foreground-background class imbalance is a common occurrence in medical images, and U-Net has difficulty in handling class imbalance because of its cross entropy (CE) objective function. Similarly, GAN also suffers from class imbalance because the discriminator looks at the entire image to classify it as real or fake. Since the discriminator is essentially a deep learning classifier, it is incapable of correctly identifying minor changes in small structures. To address these issues, we propose a novel context based CE loss function for U-Net, and a novel architecture Seg-GLGAN. The context based CE is a linear combination of CE obtained over the entire image and its region of interest (ROI). In Seg-GLGAN, we introduce a novel context discriminator to which the entire image and its ROI are fed as input, thus enforcing local context. We conduct extensive experiments using two challenging unbalanced datasets: PROMISE12 and ACDC. We observe that segmentation results obtained from our methods give better segmentation metrics as compared to various baseline methods.