PyTorchKR Papers

최신 논문

총 8,922편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2112.00849공식 코드 · Papers with Code

Interpretable Deep Learning-Based Forensic Iris Segmentation and Recognition

Iris recognition of living individuals is a mature biometric modality that has been adopted globally from governmental ID programs, border crossing, voter registration and de-duplication, to unlocking mobile phones. On the other hand, the possibility of recognizing deceased subjects with their iris patterns has emerged recently. In this paper, we present an end-to-end deep learning-based method for postmortem iris segmentation and recognition with a special visualization technique intended to support forensic human examiners in their efforts. The proposed postmortem iris segmentation approach outperforms the state of the art and in addition to iris annulus, as in case of classical iris segmentation methods - detects abnormal regions caused by eye decomposition processes, such as furrows or irregular specular highlights present on the drying and wrinkling cornea. The method was trained and validated with data acquired from 171 cadavers, kept in mortuary conditions, and tested on subject-disjoint data acquired from 259 deceased subjects. To our knowledge, this is the largest corpus of data used in postmortem iris recognition research to date. The source code of the proposed method are offered with the paper. The test data will be available through the National Archive of Criminal Justice Data (NACJD) archives.

arXiv:2112.01525

Co-domain Symmetry for Complex-Valued Deep Learning

We study complex-valued scaling as a type of symmetry natural and unique to complex-valued measurements and representations. Deep Complex Networks (DCN) extends real-valued algebra to the complex domain without addressing complex-valued scaling. SurReal takes a restrictive manifold view of complex numbers, adopting a distance metric to achieve complex-scaling invariance while losing rich complex-valued information. We analyze complex-valued scaling as a co-domain transformation and design novel equivariant and invariant neural network layer functions for this special transformation. We also propose novel complex-valued representations of RGB images, where complex-valued scaling indicates hue shift or correlated changes across color channels. Benchmarked on MSTAR, CIFAR10, CIFAR100, and SVHN, our co-domain symmetric (CDS) classifiers deliver higher accuracy, better generalization, robustness to co-domain transformations, and lower model bias and variance than DCN and SurReal with far fewer parameters.

arXiv:2112.01423공식 코드 · Papers with Code

Training Efficiency and Robustness in Deep Learning

Deep Learning has revolutionized machine learning and artificial intelligence, achieving superhuman performance in several standard benchmarks. It is well-known that deep learning models are inefficient to train; they learn by processing millions of training data multiple times and require powerful computational resources to process large batches of data in parallel at the same time rather than sequentially. Deep learning models also have unexpected failure modes; they can be fooled into misbehaviour, producing unexpectedly incorrect predictions. In this thesis, we study approaches to improve the training efficiency and robustness of deep learning models. In the context of learning visual-semantic embeddings, we find that prioritizing learning on more informative training data increases convergence speed and improves generalization performance on test data. We formalize a simple trick called hard negative mining as a modification to the learning objective function with no computational overhead. Next, we seek improvements to optimization speed in general-purpose optimization methods in deep learning. We show that a redundancy-aware modification to the sampling of training data improves the training speed and develops an efficient method for detecting the diversity of training signal, namely, gradient clustering. Finally, we study adversarial robustness in deep learning and approaches to achieve maximal adversarial robustness without training with additional data. For linear models, we prove guaranteed maximal robustness achieved only by appropriate choice of the optimizer, regularization, or architecture.

arXiv:2112.00534

Empirical evaluation of shallow and deep learning classifiers for Arabic sentiment analysis

This work presents a detailed comparison of the performance of deep learning models such as convolutional neural networks (CNN), long short-term memory (LSTM), gated recurrent units (GRU), their hybrids, and a selection of shallow learning classifiers for sentiment analysis of Arabic reviews. Additionally, the comparison includes state-of-the-art models such as the transformer architecture and the araBERT pre-trained model. The datasets used in this study are multi-dialect Arabic hotel and book review datasets, which are some of the largest publicly available datasets for Arabic reviews. Results showed deep learning outperforming shallow learning for binary and multi-label classification, in contrast with the results of similar work reported in the literature. This discrepancy in outcome was caused by dataset size as we found it to be proportional to the performance of deep learning models. The performance of deep and shallow learning techniques was analyzed in terms of accuracy and F1 score. The best performing shallow learning technique was Random Forest followed by Decision Tree, and AdaBoost. The deep learning models performed similarly using a default embedding layer, while the transformer model performed best when augmented with araBERT.

arXiv:2112.00686

CYBORG: Blending Human Saliency Into the Loss Improves Deep Learning

Can deep learning models achieve greater generalization if their training is guided by reference to human perceptual abilities? And how can we implement this in a practical manner? This paper proposes a training strategy to ConveY Brain Oversight to Raise Generalization (CYBORG). This new approach incorporates human-annotated saliency maps into a loss function that guides the model's learning to focus on image regions that humans deem salient for the task. The Class Activation Mapping (CAM) mechanism is used to probe the model's current saliency in each training batch, juxtapose this model saliency with human saliency, and penalize large differences. Results on the task of synthetic face detection, selected to illustrate the effectiveness of the approach, show that CYBORG leads to significant improvement in accuracy on unseen samples consisting of face images generated from six Generative Adversarial Networks across multiple classification network architectures. We also show that scaling to even seven times the training data, or using non-human-saliency auxiliary information, such as segmentation masks, and standard loss cannot beat the performance of CYBORG-trained models. As a side effect of this work, we observe that the addition of explicit region annotation to the task of synthetic face detection increased human classification accuracy. This work opens a new area of research on how to incorporate human visual saliency into loss functions in practice. All data, code and pre-trained models used in this work are offered with this paper.

arXiv:2112.00324

Optimizing for In-memory Deep Learning with Emerging Memory Technology

In-memory deep learning computes neural network models where they are stored, thus avoiding long distance communication between memory and computation units, resulting in considerable savings in energy and time. In-memory deep learning has already demonstrated orders of magnitude higher performance density and energy efficiency. The use of emerging memory technology promises to increase the gains in density, energy, and performance even further. However, emerging memory technology is intrinsically unstable, resulting in random fluctuations of data reads. This can translate to non-negligible accuracy loss, potentially nullifying the gains. In this paper, we propose three optimization techniques that can mathematically overcome the instability problem of emerging memory technology. They can improve the accuracy of the in-memory deep learning model while maximizing its energy efficiency. Experiments show that our solution can fully recover most models' state-of-the-art accuracy, and achieves at least an order of magnitude higher energy efficiency than the state-of-the-art.

arXiv:2112.00190

Is the use of Deep Learning and Artificial Intelligence an appropriate means to locate debris in the ocean without harming aquatic wildlife?

With the global issue of plastic debris ever expanding, it is about time that the technology industry stepped in. This study aims to assess whether deep learning can successfully distinguish between marine life and man-made debris underwater. The aim is to find if we are safely able to clean up our oceans with Artificial Intelligence without disrupting the delicate balance of the aquatic ecosystems. The research explores the use of Convolutional Neural Networks from the perspective of protecting the ecosystem, rather than primarily collecting rubbish. We did this by building a custom-built, deep learning model, with an original database including 1,644 underwater images and used a binary classification to sort synthesised material from aquatic life. We concluded that although it is possible to safely distinguish between debris and life, further exploration with a larger database and stronger CNN structure has the potential for much more promising results.

arXiv:2111.15613

The MIS Check-Dam Dataset for Object Detection and Instance Segmentation Tasks

Deep learning has led to many recent advances in object detection and instance segmentation, among other computer vision tasks. These advancements have led to wide application of deep learning based methods and related methodologies in object detection tasks for satellite imagery. In this paper, we introduce MIS Check-Dam, a new dataset of check-dams from satellite imagery for building an automated system for the detection and mapping of check-dams, focusing on the importance of irrigation structures used for agriculture. We review some of the most recent object detection and instance segmentation methods and assess their performance on our new dataset. We evaluate several single stage, two-stage and attention based methods under various network configurations and backbone architectures. The dataset and the pre-trained models are available at https://www.cse.iitb.ac.in/gramdrishti/.

arXiv:2111.15518

Detecting Adversaries, yet Faltering to Noise? Leveraging Conditional Variational AutoEncoders for Adversary Detection in the Presence of Noisy Images

With the rapid advancement and increased use of deep learning models in image identification, security becomes a major concern to their deployment in safety-critical systems. Since the accuracy and robustness of deep learning models are primarily attributed from the purity of the training samples, therefore the deep learning architectures are often susceptible to adversarial attacks. Adversarial attacks are often obtained by making subtle perturbations to normal images, which are mostly imperceptible to humans, but can seriously confuse the state-of-the-art machine learning models. What is so special in the slightest intelligent perturbations or noise additions over normal images that it leads to catastrophic classifications by the deep neural networks? Using statistical hypothesis testing, we find that Conditional Variational AutoEncoders (CVAE) are surprisingly good at detecting imperceptible image perturbations. In this paper, we show how CVAEs can be effectively used to detect adversarial attacks on image classification networks. We demonstrate our results over MNIST, CIFAR-10 dataset and show how our method gives comparable performance to the state-of-the-art methods in detecting adversaries while not getting confused with noisy images, where most of the existing methods falter.

arXiv:2111.15409

Fully Automatic Deep Learning Framework for Pancreatic Ductal Adenocarcinoma Detection on Computed Tomography

Early detection improves prognosis in pancreatic ductal adenocarcinoma (PDAC) but is challenging as lesions are often small and poorly defined on contrast-enhanced computed tomography scans (CE-CT). Deep learning can facilitate PDAC diagnosis, however current models still fail to identify small (<2cm) lesions. In this study, state-of-the-art deep learning models were used to develop an automatic framework for PDAC detection, focusing on small lesions. Additionally, the impact of integrating surrounding anatomy was investigated. CE-CT scans from a cohort of 119 pathology-proven PDAC patients and a cohort of 123 patients without PDAC were used to train a nnUnet for automatic lesion detection and segmentation (nnUnet_T). Two additional nnUnets were trained to investigate the impact of anatomy integration: (1) segmenting the pancreas and tumor (nnUnet_TP), (2) segmenting the pancreas, tumor, and multiple surrounding anatomical structures (nnUnet_MS). An external, publicly available test set was used to compare the performance of the three networks. The nnUnet_MS achieved the best performance, with an area under the receiver operating characteristic curve of 0.91 for the whole test set and 0.88 for tumors <2cm, showing that state-of-the-art deep learning can detect small PDAC and benefits from anatomy information.

arXiv:2111.15317

AutoDrop: Training Deep Learning Models with Automatic Learning Rate Drop

Modern deep learning (DL) architectures are trained using variants of the SGD algorithm that is run with a $\textit{manually}$ defined learning rate schedule, i.e., the learning rate is dropped at the pre-defined epochs, typically when the training loss is expected to saturate. In this paper we develop an algorithm that realizes the learning rate drop $\textit{automatically}$. The proposed method, that we refer to as AutoDrop, is motivated by the observation that the angular velocity of the model parameters, i.e., the velocity of the changes of the convergence direction, for a fixed learning rate initially increases rapidly and then progresses towards soft saturation. At saturation the optimizer slows down thus the angular velocity saturation is a good indicator for dropping the learning rate. After the drop, the angular velocity "resets" and follows the previously described pattern - it increases again until saturation. We show that our method improves over SOTA training approaches: it accelerates the training of DL models and leads to a better generalization. We also show that our method does not require any extra hyperparameter tuning. AutoDrop is furthermore extremely simple to implement and computationally cheap. Finally, we develop a theoretical framework for analyzing our algorithm and provide convergence guarantees.

arXiv:2111.15090

The Geometric Occam's Razor Implicit in Deep Learning

In over-parameterized deep neural networks there can be many possible parameter configurations that fit the training data exactly. However, the properties of these interpolating solutions are poorly understood. We argue that over-parameterized neural networks trained with stochastic gradient descent are subject to a Geometric Occam's Razor; that is, these networks are implicitly regularized by the geometric model complexity. For one-dimensional regression, the geometric model complexity is simply given by the arc length of the function. For higher-dimensional settings, the geometric model complexity depends on the Dirichlet energy of the function. We explore the relationship between this Geometric Occam's Razor, the Dirichlet energy and other known forms of implicit regularization. Finally, for ResNets trained on CIFAR-10, we observe that Dirichlet energy measurements are consistent with the action of this implicit Geometric Occam's Razor.

arXiv:2111.14955

Privacy-Preserving Serverless Edge Learning with Decentralized Small Data

In the last decade, data-driven algorithms outperformed traditional optimization-based algorithms in many research areas, such as computer vision, natural language processing, etc. However, extensive data usages bring a new challenge or even threat to deep learning algorithms, i.e., privacy-preserving. Distributed training strategies have recently become a promising approach to ensure data privacy when training deep models. This paper extends conventional serverless platforms with serverless edge learning architectures and provides an efficient distributed training framework from the networking perspective. This framework dynamically orchestrates available resources among heterogeneous physical units to efficiently fulfill deep learning objectives. The design jointly considers learning task requests and underlying infrastructure heterogeneity, including last-mile transmissions, computation abilities of mobile devices, edge and cloud computing centers, and devices battery status. Furthermore, to significantly reduce distributed training overheads, small-scale data training is proposed by integrating with a general, simple data classifier. This low-load enhancement can seamlessly work with various distributed deep models to improve communications and computation efficiencies during the training phase. Finally, open challenges and future research directions encourage the research community to develop efficient distributed deep learning techniques.

arXiv:2111.15111

Automatic tracing of mandibular canal pathways using deep learning

There is an increasing demand in medical industries to have automated systems for detection and localization which are manually inefficient otherwise. In dentistry, it bears great interest to trace the pathway of mandibular canals accurately. Proper localization of the position of the mandibular canals, which surrounds the inferior alveolar nerve (IAN), reduces the risk of damaging it during dental implantology. Manual detection of canal paths is not an efficient way in terms of time and labor. Here, we propose a deep learning-based framework to detect mandibular canals from CBCT data. It is a 3-stage process fully automatic end-to-end. Ground truths are generated in the preprocessing stage. Instead of using commonly used fixed diameter tubular-shaped ground truth, we generate centerlines of the mandibular canals and used them as ground truths in the training process. A 3D U-Net architecture is used for model training. An efficient post-processing stage is developed to rectify the initial prediction. The precision, recall, F1-score, and IoU are measured to analyze the voxel-level segmentation performance. However, to analyze the distance-based measurements, mean curve distance (MCD) both from ground truth to prediction and prediction to ground truth is calculated. Extensive experiments are conducted to demonstrate the effectiveness of the model.

arXiv:2111.15015

Neural Attention for Image Captioning: Review of Outstanding Methods

Image captioning is the task of automatically generating sentences that describe an input image in the best way possible. The most successful techniques for automatically generating image captions have recently used attentive deep learning models. There are variations in the way deep learning models with attention are designed. In this survey, we provide a review of literature related to attentive deep learning models for image captioning. Instead of offering a comprehensive review of all prior work on deep image captioning models, we explain various types of attention mechanisms used for the task of image captioning in deep learning models. The most successful deep learning models used for image captioning follow the encoder-decoder architecture, although there are differences in the way these models employ attention mechanisms. Via analysis on performance results from different attentive deep models for image captioning, we aim at finding the most successful types of attention mechanisms in deep models for image captioning. Soft attention, bottom-up attention, and multi-head attention are the types of attention mechanism widely used in state-of-the-art attentive deep learning models for image captioning. At the current time, the best results are achieved from variants of multi-head attention with bottom-up attention.

arXiv:2111.14931

How Facial Features Convey Attention in Stationary Environments

Awareness detection technologies have been gaining traction in a variety of enterprises; most often used for driver fatigue detection, recent research has shifted towards using computer vision technologies to analyze user attention in environments such as online classrooms. This paper aims to extend previous research on distraction detection by analyzing which visual features contribute most to predicting awareness and fatigue. We utilized the open source facial analysis toolkit OpenFace in order to analyze visual data of subjects at varying levels of attentiveness. Then, using a Support-Vector Machine (SVM) we created several prediction models for user attention and identified Histogram of Oriented Gradients (HOG) and Action Units to be the greatest predictors of the features we tested. We also compared the performance of this SVM to deep learning approaches that utilize Convolutional and/or Recurrent neural networks (CNN's and CRNN's). Interestingly, CRNN's did not appear to perform significantly better than their CNN counterparts. While deep learning methods achieved greater prediction accuracy, SVMs utilized less resources and, using certain parameters, were able to approach the performance of deep learning methods.

arXiv:2111.14377

Collective Intelligence for Deep Learning: A Survey of Recent Developments

In the past decade, we have witnessed the rise of deep learning to dominate the field of artificial intelligence. Advances in artificial neural networks alongside corresponding advances in hardware accelerators with large memory capacity, together with the availability of large datasets enabled practitioners to train and deploy sophisticated neural network models that achieve state-of-the-art performance on tasks across several fields spanning computer vision, natural language processing, and reinforcement learning. However, as these neural networks become bigger, more complex, and more widely used, fundamental problems with current deep learning models become more apparent. State-of-the-art deep learning models are known to suffer from issues that range from poor robustness, inability to adapt to novel task settings, to requiring rigid and inflexible configuration assumptions. Collective behavior, commonly observed in nature, tends to produce systems that are robust, adaptable, and have less rigid assumptions about the environment configuration. Collective intelligence, as a field, studies the group intelligence that emerges from the interactions of many individuals. Within this field, ideas such as self-organization, emergent behavior, swarm optimization, and cellular automata were developed to model and explain complex systems. It is therefore natural to see these ideas incorporated into newer deep learning methods. In this review, we will provide a historical context of neural network research's involvement with complex systems, and highlight several active areas in modern deep learning research that incorporate the principles of collective intelligence to advance its current capabilities. We hope this review can serve as a bridge between the complex systems and deep learning communities.

arXiv:2111.14110공식 코드 · Papers with Code

Enhancing Identification of Structure Function of Academic Articles Using Contextual Information

With the enrichment of literature resources, researchers are facing the growing problem of information explosion and knowledge overload. To help scholars retrieve literature and acquire knowledge successfully, clarifying the semantic structure of the content in academic literature has become the essential research question. In the research on identifying the structure function of chapters in academic articles, only a few studies used the deep learning model and explored the optimization for feature input. This limits the application, optimization potential of deep learning models for the research task. This paper took articles of the ACL conference as the corpus. We employ the traditional machine learning models and deep learning models to construct the classifiers based on various feature input. Experimental results show that (1) Compared with the chapter content, the chapter title is more conducive to identifying the structure function of academic articles. (2) Relative position is a valuable feature for building traditional models. (3) Inspired by (2), this paper further introduces contextual information into the deep learning models and achieved significant results. Meanwhile, our models show good migration ability in the open test containing 200 sampled non-training samples. We also annotated the ACL main conference papers in recent five years based on the best practice performing models and performed a time series analysis of the overall corpus. This work explores and summarizes the practical features and models for this task through multiple comparative experiments and provides a reference for related text classification tasks. Finally, we indicate the limitations and shortcomings of the current model and the direction of further optimization.

arXiv:2111.13901

Fast Sampling from Time-Integrated Bridges using Deep Learning

We propose a methodology to sample from time-integrated stochastic bridges, namely random variables defined as $\int_{t_1}^{t_2} f(Y(t))dt$ conditioned on $Y(t_1)\!=\!a$ and $Y(t_2)\!=\!b$, with $a,b\in R$. The Stochastic Collocation Monte Carlo sampler and the Seven-League scheme are applied for this purpose. Notably, the distribution of the time-integrated bridge is approximated utilizing a polynomial chaos expansion built on a suitable set of stochastic collocation points. Furthermore, artificial neural networks are employed to learn the collocation points. The result is a robust, data-driven procedure for the Monte Carlo sampling from conditional time-integrated processes, which guarantees high accuracy and generates thousands of samples in milliseconds. Applications, with a focus on finance, are presented here as well.

arXiv:2111.14761

Adaptive First- and Second-Order Algorithms for Large-Scale Machine Learning

In this paper, we consider both first- and second-order techniques to address continuous optimization problems arising in machine learning. In the first-order case, we propose a framework of transition from deterministic or semi-deterministic to stochastic quadratic regularization methods. We leverage the two-phase nature of stochastic optimization to propose a novel first-order algorithm with adaptive sampling and adaptive step size. In the second-order case, we propose a novel stochastic damped L-BFGS method that improves on previous algorithms in the highly nonconvex context of deep learning. Both algorithms are evaluated on well-known deep learning datasets and exhibit promising performance.