PyTorchKR Papers

최신 논문

총 8,922편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2101.04073

Deeplite Neutrino: An End-to-End Framework for Constrained Deep Learning Model Optimization

Designing deep learning-based solutions is becoming a race for training deeper models with a greater number of layers. While a large-size deeper model could provide competitive accuracy, it creates a lot of logistical challenges and unreasonable resource requirements during development and deployment. This has been one of the key reasons for deep learning models not being excessively used in various production environments, especially in edge devices. There is an immediate requirement for optimizing and compressing these deep learning models, to enable on-device intelligence. In this research, we introduce a black-box framework, Deeplite Neutrino for production-ready optimization of deep learning models. The framework provides an easy mechanism for the end-users to provide constraints such as a tolerable drop in accuracy or target size of the optimized models, to guide the whole optimization process. The framework is easy to include in an existing production pipeline and is available as a Python Package, supporting PyTorch and Tensorflow libraries. The optimization performance of the framework is shown across multiple benchmark datasets and popular deep learning models. Further, the framework is currently used in production and the results and testimonials from several clients are summarized.

arXiv:2101.04012

Evaluating Deep Learning Approaches for Covid19 Fake News Detection

Social media platforms like Facebook, Twitter, and Instagram have enabled connection and communication on a large scale. It has revolutionized the rate at which information is shared and enhanced its reach. However, another side of the coin dictates an alarming story. These platforms have led to an increase in the creation and spread of fake news. The fake news has not only influenced people in the wrong direction but also claimed human lives. During these critical times of the Covid19 pandemic, it is easy to mislead people and make them believe in fatal information. Therefore it is important to curb fake news at source and prevent it from spreading to a larger audience. We look at automated techniques for fake news detection from a data mining perspective. We evaluate different supervised text classification algorithms on Contraint@AAAI 2021 Covid-19 Fake news detection dataset. The classification algorithms are based on Convolutional Neural Networks (CNN), Long Short Term Memory (LSTM), and Bidirectional Encoder Representations from Transformers (BERT). We also evaluate the importance of unsupervised learning in the form of language model pre-training and distributed word representations using unlabelled covid tweets corpus. We report the best accuracy of 98.41\% on the Covid-19 Fake news detection dataset.

arXiv:2101.03973

Load Encoding for Learning AC-OPF

The AC Optimal Power Flow (AC-OPF) problem is a core building block in electrical transmission system. It seeks the most economical active and reactive generation dispatch to meet demands while satisfying transmission operational limits. It is often solved repeatedly, especially in regions with large penetration of wind farms to avoid violating operational and physical limits. Recent work has shown that deep learning techniques have huge potential in providing accurate approximations of AC-OPF solutions. However, deep learning approaches often suffer from scalability issues, especially when applied to real life power grids. This paper focuses on the scalability limitation and proposes a load compression embedding scheme to reduce training model sizes using a 3-step approach. The approach is evaluated experimentally on large-scale test cases from the PGLib, and produces an order of magnitude improvements in training convergence and prediction accuracy.

arXiv:2101.02888공식 코드 · Papers with Code

Predicting Semen Motility using three-dimensional Convolutional Neural Networks

Manual and computer aided methods to perform semen analysis are time-consuming, requires extensive training and prone to human error. The use of classical machine learning and deep learning based methods using videos to perform semen analysis have yielded good results. The state-of-the-art method uses regular convolutional neural networks to perform quality assessments on a video of the provided sample. In this paper we propose an improved deep learning based approach using three-dimensional convolutional neural networks to predict sperm motility from microscopic videos of the semen sample. We make use of the VISEM dataset that consists of video and tabular data of semen samples collected from 85 participants. We were able to achieve good results from significantly less data points. Our models indicate that deep learning based automatic semen analysis may become a valuable and effective tool in fertility and IVF labs.

arXiv:2101.02644

Data Poisoning Attacks to Deep Learning Based Recommender Systems

Recommender systems play a crucial role in helping users to find their interested information in various web services such as Amazon, YouTube, and Google News. Various recommender systems, ranging from neighborhood-based, association-rule-based, matrix-factorization-based, to deep learning based, have been developed and deployed in industry. Among them, deep learning based recommender systems become increasingly popular due to their superior performance. In this work, we conduct the first systematic study on data poisoning attacks to deep learning based recommender systems. An attacker's goal is to manipulate a recommender system such that the attacker-chosen target items are recommended to many users. To achieve this goal, our attack injects fake users with carefully crafted ratings to a recommender system. Specifically, we formulate our attack as an optimization problem, such that the injected ratings would maximize the number of normal users to whom the target items are recommended. However, it is challenging to solve the optimization problem because it is a non-convex integer programming problem. To address the challenge, we develop multiple techniques to approximately solve the optimization problem. Our experimental results on three real-world datasets, including small and large datasets, show that our attack is effective and outperforms existing attacks. Moreover, we attempt to detect fake users via statistical analysis of the rating patterns of normal and fake users. Our results show that our attack is still effective and outperforms existing attacks even if such a detector is deployed.

arXiv:2101.02144

Combining Deep Learning and Mathematical Morphology for Historical Map Segmentation

The digitization of historical maps enables the study of ancient, fragile, unique, and hardly accessible information sources. Main map features can be retrieved and tracked through the time for subsequent thematic analysis. The goal of this work is the vectorization step, i.e., the extraction of vector shapes of the objects of interest from raster images of maps. We are particularly interested in closed shape detection such as buildings, building blocks, gardens, rivers, etc. in order to monitor their temporal evolution. Historical map images present significant pattern recognition challenges. The extraction of closed shapes by using traditional Mathematical Morphology (MM) is highly challenging due to the overlapping of multiple map features and texts. Moreover, state-of-the-art Convolutional Neural Networks (CNN) are perfectly designed for content image filtering but provide no guarantee about closed shape detection. Also, the lack of textural and color information of historical maps makes it hard for CNN to detect shapes that are represented by only their boundaries. Our contribution is a pipeline that combines the strengths of CNN (efficient edge detection and filtering) and MM (guaranteed extraction of closed shapes) in order to achieve such a task. The evaluation of our approach on a public dataset shows its effectiveness for extracting the closed boundaries of objects in historical maps.

arXiv:2101.02118공식 코드 · Papers with Code

Do We Really Need Deep Learning Models for Time Series Forecasting?

Time series forecasting is a crucial task in machine learning, as it has a wide range of applications including but not limited to forecasting electricity consumption, traffic, and air quality. Traditional forecasting models rely on rolling averages, vector auto-regression and auto-regressive integrated moving averages. On the other hand, deep learning and matrix factorization models have been recently proposed to tackle the same problem with more competitive performance. However, one major drawback of such models is that they tend to be overly complex in comparison to traditional techniques. In this paper, we report the results of prominent deep learning models with respect to a well-known machine learning baseline, a Gradient Boosting Regression Tree (GBRT) model. Similar to the deep neural network (DNN) models, we transform the time series forecasting task into a window-based regression problem. Furthermore, we feature-engineered the input and output structure of the GBRT model, such that, for each training window, the target values are concatenated with external features, and then flattened to form one input instance for a multi-output GBRT model. We conducted a comparative study on nine datasets for eight state-of-the-art deep-learning models that were presented at top-level conferences in the last years. The results demonstrate that the window-based input transformation boosts the performance of a simple GBRT model to levels that outperform all state-of-the-art DNN models evaluated in this paper.

arXiv:2101.01352

Development of a Respiratory Sound Labeling Software for Training a Deep Learning-Based Respiratory Sound Analysis Model

Respiratory auscultation can help healthcare professionals detect abnormal respiratory conditions if adventitious lung sounds are heard. The state-of-the-art artificial intelligence technologies based on deep learning show great potential in the development of automated respiratory sound analysis. To train a deep learning-based model, a huge number of accurate labels of normal breath sounds and adventitious sounds are needed. In this paper, we demonstrate the work of developing a respiratory sound labeling software to help annotators identify and label the inhalation, exhalation, and adventitious respiratory sound more accurately and quickly. Our labeling software integrates six features from MATLAB Audio Labeler, and one commercial audio editor, RX7. As of October, 2019, we have labeled 9,765 15-second-long audio files of breathing lung sounds, and accrued 34,095 inhalation labels,18,349 exhalation labels, 13,883 continuous adventitious sounds (CASs) labels and 15,606 discontinuous adventitious sounds (DASs) labels, which are significantly larger than previously published studies. The trained convolutional recurrent neural networks based on these labels showed good performance with F1-scores of 86.0% on inhalation event detection, 51.6% on CASs event detection and 71.4% on DASs event detection. In conclusion, our results show that our proposed respiratory sound labeling software could easily pre-define a label, perform one-click labeling, and overall facilitate the process of accurately labeling. This software helps develop deep learning-based models that require a huge amount of labeled acoustic data.

arXiv:2101.01310

Research on Fast Text Recognition Method for Financial Ticket Image

Currently, deep learning methods have been widely applied in and thus promoted the development of different fields. In the financial accounting field, the rapid increase in the number of financial tickets dramatically increases labor costs; hence, using a deep learning method to relieve the pressure on accounting is necessary. At present, a few works have applied deep learning methods to financial ticket recognition. However, first, their approaches only cover a few types of tickets. In addition, the precision and speed of their recognition models cannot meet the requirements of practical financial accounting systems. Moreover, none of the methods provides a detailed analysis of both the types and content of tickets. Therefore, this paper first analyzes the different features of 482 kinds of financial tickets, divides all kinds of financial tickets into three categories and proposes different recognition patterns for each category. These recognition patterns can meet almost all types of financial ticket recognition needs. Second, regarding the fixed format types of financial tickets (accounting for 68.27\% of the total types of tickets), we propose a simple yet efficient network named the Financial Ticket Faster Detection network (FTFDNet) based on a Faster RCNN. Furthermore, according to the characteristics of the financial ticket text, in order to obtain higher recognition accuracy, the loss function, Region Proposal Network (RPN), and Non-Maximum Suppression (NMS) are improved to make FTFDNet focus more on text. Finally, we perform a comparison with the best ticket recognition model from the ICDAR2019 invoice competition. The experimental results illustrate that FTFDNet increases the processing speed by 50\% while maintaining similar precision.

arXiv:2101.01214공식 코드 · Papers with Code

Reconstructing Patchy Reionization with Deep Learning

The precision anticipated from next-generation cosmic microwave background (CMB) surveys will create opportunities for characteristically new insights into cosmology. Secondary anisotropies of the CMB will have an increased importance in forthcoming surveys, due both to the cosmological information they encode and the role they play in obscuring our view of the primary fluctuations. Quadratic estimators have become the standard tools for reconstructing the fields that distort the primary CMB and produce secondary anisotropies. While successful for lensing reconstruction with current data, quadratic estimators will be sub-optimal for the reconstruction of lensing and other effects at the expected sensitivity of the upcoming CMB surveys. In this paper we describe a convolutional neural network, ResUNet-CMB, that is capable of the simultaneous reconstruction of two sources of secondary CMB anisotropies, gravitational lensing and patchy reionization. We show that the ResUNet-CMB network significantly outperforms the quadratic estimator at low noise levels and is not subject to the lensing-induced bias on the patchy reionization reconstruction that would be present with a straightforward application of the quadratic estimator.

arXiv:2101.01178

Advances in Electron Microscopy with Deep Learning

This doctoral thesis covers some of my advances in electron microscopy with deep learning. Highlights include a comprehensive review of deep learning in electron microscopy; large new electron microscopy datasets for machine learning, dataset search engines based on variational autoencoders, and automatic data clustering by t-distributed stochastic neighbour embedding; adaptive learning rate clipping to stabilize learning; generative adversarial networks for compressed sensing with spiral, uniformly spaced and other fixed sparse scan paths; recurrent neural networks trained to piecewise adapt sparse scan paths to specimens by reinforcement learning; improving signal-to-noise; and conditional generative adversarial networks for exit wavefunction reconstruction from single transmission electron micrographs. This thesis adds to my publications by presenting their relationships, reflections, and holistic conclusions. This version of my thesis is typeset for online dissemination to improve readability, whereas the thesis submitted to the University of Warwick in support of my application for the degree of Doctor of Philosophy in Physics is typeset for physical printing and binding.

arXiv:2101.01115

Analysis of Filter Size Effect In Deep Learning

With the use of deep learning in many areas, how to improve this technology or how to develop the structure used more effectively and in a shorter time is an issue that is of interest to many people working in this field. Many studies are carried out on this subject, it is aimed to reduce the duration of the operation and the processing power required, except to obtain the best result with the changes made in the variables, functions and data in the models used. In this study, in the leaf classification made using Mendeley data set consisting of leaf images with a fixed background, all other variables such as layer number, iteration, number of layers in the model and pooling process were kept constant, except for the filter dimensions of the convolution layers in the determined model. Convolution layers in 3 different filter sizes and in addition to this, many results obtained in 2 different structures, increasing and decreasing, and 3 different image sizes were examined. In the literature, it is seen that different uses of pooling layers, changes due to increase or decrease in the number of layers, the difference in the size of the data used, and the results of many functions used with different parameters are evaluated. In the leaf classification of the determined data set with CNN, the change in the filter size of the convolution layer together with the change in different filter combinations and in different sized images was focused. Using the data set and data reproduction methods, it was aimed to make the differences in filter sizes and image sizes more distinct. Using the fixed number of iterations, model and data set, the effect of different filter sizes has been observed.

arXiv:2101.00982공식 코드 · Papers with Code

Uncertainty-Wizard: Fast and User-Friendly Neural Network Uncertainty Quantification

Uncertainty and confidence have been shown to be useful metrics in a wide variety of techniques proposed for deep learning testing, including test data selection and system supervision.We present uncertainty-wizard, a tool that allows to quantify such uncertainty and confidence in artificial neural networks. It is built on top of the industry-leading tf.keras deep learning API and it provides a near-transparent and easy to understand interface. At the same time, it includes major performance optimizations that we benchmarked on two different machines and different configurations.

arXiv:2101.00073

A Multi-modal Deep Learning Model for Video Thumbnail Selection

Thumbnail is the face of online videos. The explosive growth of videos both in number and variety underpins the importance of a good thumbnail because it saves potential viewers time to choose videos and even entice them to click on them. A good thumbnail should be a frame that best represents the content of a video while at the same time capturing viewers' attention. However, the techniques and models in the past only focus on frames within a video, and we believe such narrowed focus leave out much useful information that are part of a video. In this paper, we expand the definition of content to include title, description, and audio of a video and utilize information provided by these modalities in our selection model. Specifically, our model will first sample frames uniformly in time and return the top 1,000 frames in this subset with the highest aesthetic scores by a Double-column Convolutional Neural Network, to avoid the computational burden of processing all frames in downstream task. Then, the model incorporates frame features extracted from VGG16, text features from ELECTRA, and audio features from TRILL. These models were selected because of their results on popular datasets as well as their competitive performances. After feature extraction, the time-series features, frames and audio, will be fed into Transformer encoder layers to return a vector representing their corresponding modality. Each of the four features (frames, title, description, audios) will pass through a context gating layer before concatenation. Finally, our model will generate a vector in the latent space and select the frame that is most similar to this vector in the latent space. To the best of our knowledge, we are the first to propose a multi-modal deep learning model to select video thumbnail, which beats the result from the previous State-of-The-Art models.

arXiv:2101.00337

Biologically Inspired Hexagonal Deep Learning for Hexagonal Image Generation

Whereas conventional state-of-the-art image processing systems of recording and output devices almost exclusively utilize square arranged methods, biological models, however, suggest an alternative, evolutionarily-based structure. Inspired by the human visual perception system, hexagonal image processing in the context of machine learning offers a number of key advantages that can benefit both researchers and users alike. The hexagonal deep learning framework Hexnet leveraged in this contribution serves therefore the generation of hexagonal images by utilizing hexagonal deep neural networks (H-DNN). As the results of our created test environment show, the proposed models can surpass current approaches of conventional image generation. While resulting in a reduction of the models' complexity in the form of trainable parameters, they furthermore allow an increase of test rates in comparison to their square counterparts.

arXiv:2101.00218

An iterative K-FAC algorithm for Deep Learning

Kronecker-factored Approximate Curvature (K-FAC) method is a high efficiency second order optimizer for the deep learning. Its training time is less than SGD(or other first-order method) with same accuracy in many large-scale problems. The key of K-FAC is to approximates Fisher information matrix (FIM) as a block-diagonal matrix where each block is an inverse of tiny Kronecker factors. In this short note, we present CG-FAC -- an new iterative K-FAC algorithm. It uses conjugate gradient method to approximate the nature gradient. This CG-FAC method is matrix-free, that is, no need to generate the FIM matrix, also no need to generate the Kronecker factors A and G. We prove that the time and memory complexity of iterative CG-FAC is much less than that of standard K-FAC algorithm.

arXiv:2101.00747

Frequency Principle in Deep Learning Beyond Gradient-descent-based Training

Frequency perspective recently makes progress in understanding deep learning. It has been widely verified in both empirical and theoretical studies that deep neural networks (DNNs) often fit the target function from low to high frequency, namely Frequency Principle (F-Principle). F-Principle sheds light on the strength and the weakness of DNNs and inspires a series of subsequent works, including theoretical studies, empirical studies and the design of efficient DNN structures etc. Previous works examine the F-Principle in gradient-descent-based training. It remains unclear whether gradient-descent-based training is a necessary condition for the F-Principle. In this paper, we show that the F-Principle exists stably in the training process of DNNs with non-gradient-descent-based training, including optimization algorithms with gradient information, such as conjugate gradient and BFGS, and algorithms without gradient information, such as Powell's method and Particle Swarm Optimization. These empirical studies show the universality of the F-Principle and provide hints for further study of F-Principle.

arXiv:2101.01121

Local Competition and Stochasticity for Adversarial Robustness in Deep Learning

This work addresses adversarial robustness in deep learning by considering deep networks with stochastic local winner-takes-all (LWTA) activations. This type of network units result in sparse representations from each model layer, as the units are organized in blocks where only one unit generates a non-zero output. The main operating principle of the introduced units lies on stochastic arguments, as the network performs posterior sampling over competing units to select the winner. We combine these LWTA arguments with tools from the field of Bayesian non-parametrics, specifically the stick-breaking construction of the Indian Buffet Process, to allow for inferring the sub-part of each layer that is essential for modeling the data at hand. Then, inference is performed by means of stochastic variational Bayes. We perform a thorough experimental evaluation of our model using benchmark datasets. As we show, our method achieves high robustness to adversarial perturbations, with state-of-the-art performance in powerful adversarial attack schemes.

arXiv:2012.15391

Generalized Operating Procedure for Deep Learning: an Unconstrained Optimal Design Perspective

Deep learning (DL) has brought about remarkable breakthrough in processing images, video and speech due to its efficacy in extracting highly abstract representation and learning very complex functions. However, there is seldom operating procedure reported on how to make it for real use cases. In this paper, we intend to address this problem by presenting a generalized operating procedure for DL from the perspective of unconstrained optimal design, which is motivated by a simple intension to remove the barrier of using DL, especially for those scientists or engineers who are new but eager to use it. Our proposed procedure contains seven steps, which are project/problem statement, data collection, architecture design, initialization of parameters, defining loss function, computing optimal parameters, and inference, respectively. Following this procedure, we build a multi-stream end-to-end speaker verification system, in which the input speech utterance is processed by multiple parallel streams within different frequency range, so that the acoustic modeling can be more robust resulting from the diversity of features. Trained with VoxCeleb dataset, our experimental results verify the effectiveness of our proposed operating procedure, and also show that our multi-stream framework outperforms single-stream baseline with 20 % relative reduction in minimum decision cost function (minDCF).

arXiv:2012.15330

Sequential Deep Learning for Credit Risk Monitoring with Tabular Financial Data

Machine learning plays an essential role in preventing financial losses in the banking industry. Perhaps the most pertinent prediction task that can result in billions of dollars in losses each year is the assessment of credit risk (i.e., the risk of default on debt). Today, much of the gains from machine learning to predict credit risk are driven by gradient boosted decision tree models. However, these gains begin to plateau without the addition of expensive new data sources or highly engineered features. In this paper, we present our attempts to create a novel approach to assessing credit risk using deep learning that does not rely on new model inputs. We propose a new credit card transaction sampling technique to use with deep recurrent and causal convolution-based neural networks that exploits long historical sequences of financial data without costly resource requirements. We show that our sequential deep learning approach using a temporal convolutional network outperformed the benchmark non-sequential tree-based model, achieving significant financial savings and earlier detection of credit risk. We also demonstrate the potential for our approach to be used in a production environment, where our sampling technique allows for sequences to be stored efficiently in memory and used for fast online learning and inference.