Dialogue Discourse-Aware Graph Model and Data Augmentation for Meeting Summarization

Meeting summarization is a challenging task due to its dynamic interaction nature among multiple speakers and lack of sufficient training data. Existing methods view the meeting as a linear sequence of utterances while ignoring the diverse relations between each utterance. Besides, the limited labeled data further hinders the ability of data-hungry neural models. In this paper, we try to mitigate the above challenges by introducing dialogue-discourse relations. First, we present a Dialogue Discourse-Dware Meeting Summarizer (DDAMS) to explicitly model the interaction between utterances in a meeting by modeling different discourse relations. The core module is a relational graph encoder, where the utterances and discourse relations are modeled in a graph interaction manner. Moreover, we devise a Dialogue Discourse-Aware Data Augmentation (DDADA) strategy to construct a pseudo-summarization corpus from existing input meetings, which is 20 times larger than the original dataset and can be used to pretrain DDAMS. Experimental results on AMI and ICSI meeting datasets show that our full system can achieve SOTA performance. Our codes and outputs are available at https://github.com/xcfcode/DDAMS/.

Download Full-text

MIGAN: Malware Image Synthesis Using GANs

Proceedings of the AAAI Conference on Artificial Intelligence ◽

10.1609/aaai.v33i01.330110033 ◽

2019 ◽

Vol 33 ◽

pp. 10033-10034 ◽

Cited By ~ 1

Author(s):

Abhishek Singh ◽

Debojyoti Dutta ◽

Amit Saha

Keyword(s):

Language Processing ◽

Domain Knowledge ◽

Data Augmentation ◽

Image Synthesis ◽

Substantial Improvement ◽

Training Data ◽

Malware Analysis ◽

Training Procedure ◽

Original Dataset ◽

Augmentation Techniques

Majority of the advancement in Deep learning (DL) has occurred in domains such as computer vision, and natural language processing, where abundant training data is available. A major obstacle in leveraging DL techniques for malware analysis is the lack of sufficiently big, labeled datasets. In this paper, we take the first steps towards building a model which can synthesize labeled dataset of malware images using GAN. Such a model can be utilized to perform data augmentation for training a classifier. Furthermore, the model can be shared publicly for community to reap benefits of dataset without sharing the original dataset. First, we show the underlying idiosyncrasies of malware images and why existing data augmentation techniques as well as traditional GAN training fail to produce quality artificial samples. Next, we propose a new method for training GAN where we explicitly embed prior domain knowledge about the dataset into the training procedure. We show improvements in training stability and sample quality assessed on different metrics. Our experiments show substantial improvement on baselines and promise for using such a generative model for malware visualization systems.

Download Full-text

Data augmentation techniques for transfer learning improvement in drill wear classification using convolutional neural network

Machine Graphics and Vision ◽

10.22630/mgv.2019.28.1.1 ◽

2019 ◽

Vol 28 (1) ◽

pp. 3-12

Author(s):

Jarosław Kurek ◽

Joanna Aleksiejuk-Gawron ◽

Izabella Antoniuk ◽

Jarosław Górski ◽

Albina Jegorowa ◽

...

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

Transfer Learning ◽

Data Augmentation ◽

Good Condition ◽

Training Data ◽

Time Data ◽

Learning Improvement ◽

Original Dataset ◽

Augmentation Techniques

This paper presents an improved method for recognizing the drill state on the basis of hole images drilled in a laminated chipboard, using convolutional neural network (CNN) and data augmentation techniques. Three classes were used to describe the drill state: red -- for drill that is worn out and should be replaced, yellow -- for state in which the system should send a warning to the operator, indicating that this element should be checked manually, and green -- denoting the drill that is still in good condition, which allows for further use in the production process. The presented method combines the advantages of transfer learning and data augmentation methods to improve the accuracy of the received evaluations. In contrast to the classical deep learning methods, transfer learning requires much smaller training data sets to achieve acceptable results. At the same time, data augmentation customized for drill wear recognition makes it possible to expand the original dataset and to improve the overall accuracy. The experiments performed have confirmed the suitability of the presented approach to accurate class recognition in the given problem, even while using a small original dataset.

Download Full-text

Analysis of the Fuzziness of Image Caption Generation Models due to Data Augmentation Techniques

International Journal of Recent Technology and Engineering - 2 ◽

10.35940/ijrte.c6439.0910321 ◽

2021 ◽

Vol 10 (3) ◽

pp. 131-139

Author(s):

Kota Akshith Reddy ◽

◽

Satish C J ◽

Jahnavi Polsani ◽

Teja Naveen Chintapalli ◽

...

Keyword(s):

Deep Learning ◽

Data Augmentation ◽

Training Data ◽

Image Captioning ◽

The Core ◽

Augmentation Techniques ◽

Evaluation Metric ◽

Learning Data ◽

Image Caption Generation ◽

Image Caption

Automatic Image Caption Generation is one of the core problems in the field of Deep Learning. Data Augmentation is a technique which helps in increasing the amount of data at hand and this is done by augmenting the training data using various techniques like flipping, rotating, Zooming, Brightening, etc. In this work, we create an Image Captioning model and check its robustness on all the major types of Image Augmentation techniques. The results show the fuzziness of the model while working with the same image but a different augmentation technique and because of this, a different caption is produced every time a different data augmentation technique is employed. We also show the change in the performance of the model after applying these augmentation techniques. Flickr8k dataset is used for this study along with BLEU score as the evaluation metric for the image captioning model.

Download Full-text

Sensitivity Analysis and Selection of Check Index of Signal Intersection Simulation Model Based on VISSIM

智能城市应用 ◽

10.33142/sca.v2i5.818 ◽

2019 ◽

Vol 2 (5) ◽

Author(s):

Mengda Zhang ◽

Chenjing Zhou ◽

Tian-tian Zhang ◽

Yan Han

Keyword(s):

Sensitivity Analysis ◽

Simulation Model ◽

Optimal Process ◽

Parameter Calibration ◽

Index Set ◽

The Core ◽

Core Module ◽

Simulation Parameter ◽

Simulation Parameters ◽

Selection Of

Selecting check index quantitatively is the core of the calibration of micro traffic simulation parameters at signal intersection. Five indexes in the node (intersection) module of VISSIM were selected as the check index set. Twelve simulation parameters in the core module were selected as the simulation parameters set. Optimal process of parameter calibration was proposed and model of the intersection of Huangcun west street and Xinghua street in Beijing was built in VISSIM to verify it. The sensitivity analysis between each check index and simulation parameter in their own set was conducted respectively. Sensitive parameter sets of different check indices were obtained and compared. The results show that different indexes have different size of set, and average vehicle delay's is maximum, so it's necessary to select index quantitatively. The results can provide references for scientific selection of the check indexes and improve the study efficiency of parameter calibration.

Download Full-text

Building Damage Detection from Post-Event Aerial Imagery Using Single Shot Multibox Detector

Applied Sciences ◽

10.3390/app9061128 ◽

2019 ◽

Vol 9 (6) ◽

pp. 1128 ◽

Cited By ~ 12

Author(s):

Yundong Li ◽

Wei Hu ◽

Han Dong ◽

Xueyan Zhang

Keyword(s):

Machine Learning ◽

Data Augmentation ◽

Hurricane Sandy ◽

Training Data ◽

Aerial Images ◽

Detection Methods ◽

Single Shot ◽

Data Set ◽

Augmentation Strategies ◽

Post Disaster

Using aerial cameras, satellite remote sensing or unmanned aerial vehicles (UAV) equipped with cameras can facilitate search and rescue tasks after disasters. The traditional manual interpretation of huge aerial images is inefficient and could be replaced by machine learning-based methods combined with image processing techniques. Given the development of machine learning, researchers find that convolutional neural networks can effectively extract features from images. Some target detection methods based on deep learning, such as the single-shot multibox detector (SSD) algorithm, can achieve better results than traditional methods. However, the impressive performance of machine learning-based methods results from the numerous labeled samples. Given the complexity of post-disaster scenarios, obtaining many samples in the aftermath of disasters is difficult. To address this issue, a damaged building assessment method using SSD with pretraining and data augmentation is proposed in the current study and highlights the following aspects. (1) Objects can be detected and classified into undamaged buildings, damaged buildings, and ruins. (2) A convolution auto-encoder (CAE) that consists of VGG16 is constructed and trained using unlabeled post-disaster images. As a transfer learning strategy, the weights of the SSD model are initialized using the weights of the CAE counterpart. (3) Data augmentation strategies, such as image mirroring, rotation, Gaussian blur, and Gaussian noise processing, are utilized to augment the training data set. As a case study, aerial images of Hurricane Sandy in 2012 were maximized to validate the proposed method’s effectiveness. Experiments show that the pretraining strategy can improve of 10% in terms of overall accuracy compared with the SSD trained from scratch. These experiments also demonstrate that using data augmentation strategies can improve mAP and mF1 by 72% and 20%, respectively. Finally, the experiment is further verified by another dataset of Hurricane Irma, and it is concluded that the paper method is feasible.

Download Full-text

Automated Ventricular System Segmentation in Paediatric Patients Treated for Hydrocephalus Using Deep Learning Methods

BioMed Research International ◽

10.1155/2019/3059170 ◽

2019 ◽

Vol 2019 ◽

pp. 1-9 ◽

Cited By ~ 2

Author(s):

Michał Klimont ◽

Mateusz Flieger ◽

Jacek Rzeszutek ◽

Joanna Stachera ◽

Aleksandra Zakrzewska ◽

...

Keyword(s):

Neural Network ◽

Data Augmentation ◽

Ct Images ◽

Policy Transfer ◽

Training Data ◽

Intraobserver Variability ◽

Practical Applications ◽

Brain Scans ◽

Rate Policy ◽

Ct Brain

Hydrocephalus is a common neurological condition that can have traumatic ramifications and can be lethal without treatment. Nowadays, during therapy radiologists have to spend a vast amount of time assessing the volume of cerebrospinal fluid (CSF) by manual segmentation on Computed Tomography (CT) images. Further, some of the segmentations are prone to radiologist bias and high intraobserver variability. To improve this, researchers are exploring methods to automate the process, which would enable faster and more unbiased results. In this study, we propose the application of U-Net convolutional neural network in order to automatically segment CT brain scans for location of CSF. U-Net is a neural network that has proven to be successful for various interdisciplinary segmentation tasks. We optimised training using state of the art methods, including “1cycle” learning rate policy, transfer learning, generalized dice loss function, mixed float precision, self-attention, and data augmentation. Even though the study was performed using a limited amount of data (80 CT images), our experiment has shown near human-level performance. We managed to achieve a 0.917 mean dice score with 0.0352 standard deviation on cross validation across the training data and a 0.9506 mean dice score on a separate test set. To our knowledge, these results are better than any known method for CSF segmentation in hydrocephalic patients, and thus, it is promising for potential practical applications.

Download Full-text

Deep Learning-Based Differentiation between Mucinous Cystic Neoplasm and Serous Cystic Neoplasm in the Pancreas Using Endoscopic Ultrasonography

Diagnostics ◽

10.3390/diagnostics11061052 ◽

2021 ◽

Vol 11 (6) ◽

pp. 1052

Author(s):

Leang Sim Nguon ◽

Kangwon Seo ◽

Jung-Hyun Lim ◽

Tae-Jun Song ◽

Sung-Hyun Cho ◽

...

Keyword(s):

Decision Making ◽

Deep Learning ◽

Network Model ◽

Endoscopic Ultrasonography ◽

Data Augmentation ◽

Clinical Information ◽

Training Data ◽

Fine Tuning ◽

Cystic Neoplasm ◽

Cystic Neoplasms

Mucinous cystic neoplasms (MCN) and serous cystic neoplasms (SCN) account for a large portion of solitary pancreatic cystic neoplasms (PCN). In this study we implemented a convolutional neural network (CNN) model using ResNet50 to differentiate between MCN and SCN. The training data were collected retrospectively from 59 MCN and 49 SCN patients from two different hospitals. Data augmentation was used to enhance the size and quality of training datasets. Fine-tuning training approaches were utilized by adopting the pre-trained model from transfer learning while training selected layers. Testing of the network was conducted by varying the endoscopic ultrasonography (EUS) image sizes and positions to evaluate the network performance for differentiation. The proposed network model achieved up to 82.75% accuracy and a 0.88 (95% CI: 0.817–0.930) area under curve (AUC) score. The performance of the implemented deep learning networks in decision-making using only EUS images is comparable to that of traditional manual decision-making using EUS images along with supporting clinical information. Gradient-weighted class activation mapping (Grad-CAM) confirmed that the network model learned the features from the cyst region accurately. This study proves the feasibility of diagnosing MCN and SCN using a deep learning network model. Further improvement using more datasets is needed.

Download Full-text

Estimating the phase volume fraction of multi-phase steel via unsupervised deep learning

Scientific Reports ◽

10.1038/s41598-021-85407-y ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Sung Wook Kim ◽

Seong-Hoon Kang ◽

Se-Jong Kim ◽

Seungchul Lee

Keyword(s):

High Performance ◽

Materials Science ◽

Volume Fraction ◽

Training Data ◽

Phase Fraction ◽

Phase Volume ◽

Generative Adversarial Network ◽

Phase Volume Fraction ◽

Original Dataset ◽

Multi Phase

AbstractAdvanced high strength steel (AHSS) is a steel of multi-phase microstructure that is processed under several conditions to meet the current high-performance requirements from the industry. Deep neural network (DNN) has emerged as a promising tool in materials science for the task of estimating the phase volume fraction of these steels. Despite its advantages, one of its major drawbacks is its requirement of a sufficient amount of training data with correct labels to the network. This often comes as a challenge in many areas where obtaining data and labeling it is extremely labor-intensive. To overcome this challenge, an unsupervised way of learning DNN, which does not require any manual labeling, is proposed. Information maximizing generative adversarial network (InfoGAN) is used to learn the underlying probability distribution of each phase and generate realistic sample points with class labels. Then, the generated data is used for training an MLP classifier, which in turn predicts the labels for the original dataset. The result shows a mean relative error of 4.53% at most, while it can be as low as 0.73%, which implies the estimated phase fraction closely matches the true phase fraction. This presents the high feasibility of using the proposed methodology for fast and precise estimation of phase volume fraction in both industry and academia.

Download Full-text

Robust Approach to Supervised Deep Neural Network Training for Real-Time Object Classification in Cluttered Indoor Environment

Applied Sciences ◽

10.3390/app11157148 ◽

2021 ◽

Vol 11 (15) ◽

pp. 7148

Author(s):

Bedada Endale ◽

Abera Tullu ◽

Hayoung Shi ◽

Beom-Soo Kang

Keyword(s):

Neural Network ◽

Deep Learning ◽

Real Time ◽

Network Architecture ◽

Input Data ◽

Deep Neural Network ◽

Data Augmentation ◽

Object Classification ◽

Training Data ◽

Gradient Descent Algorithm

Unmanned aerial vehicles (UAVs) are being widely utilized for various missions: in both civilian and military sectors. Many of these missions demand UAVs to acquire artificial intelligence about the environments they are navigating in. This perception can be realized by training a computing machine to classify objects in the environment. One of the well known machine training approaches is supervised deep learning, which enables a machine to classify objects. However, supervised deep learning comes with huge sacrifice in terms of time and computational resources. Collecting big input data, pre-training processes, such as labeling training data, and the need for a high performance computer for training are some of the challenges that supervised deep learning poses. To address these setbacks, this study proposes mission specific input data augmentation techniques and the design of light-weight deep neural network architecture that is capable of real-time object classification. Semi-direct visual odometry (SVO) data of augmented images are used to train the network for object classification. Ten classes of 10,000 different images in each class were used as input data where 80% were for training the network and the remaining 20% were used for network validation. For the optimization of the designed deep neural network, a sequential gradient descent algorithm was implemented. This algorithm has the advantage of handling redundancy in the data more efficiently than other algorithms.

Download Full-text

Rethinking the Random Cropping Data Augmentation Method Used in the Training of CNN-Based SAR Image Ship Detector

Remote Sensing ◽

10.3390/rs13010034 ◽

2020 ◽

Vol 13 (1) ◽

pp. 34

Author(s):

Rong Yang ◽

Robert Wang ◽

Yunkai Deng ◽

Xiaoxue Jia ◽

Heng Zhang

Keyword(s):

Neural Network ◽

Data Augmentation ◽

Back Propagation ◽

Detection Performance ◽

Training Data ◽

Sar Image ◽

Optical Images ◽

The Neural Network ◽

Effective Training ◽

Standard Configuration

The random cropping data augmentation method is widely used to train convolutional neural network (CNN)-based target detectors to detect targets in optical images (e.g., COCO datasets). It can expand the scale of the dataset dozens of times while consuming only a small amount of calculations when training the neural network detector. In addition, random cropping can also greatly enhance the spatial robustness of the model, because it can make the same target appear in different positions of the sample image. Nowadays, random cropping and random flipping have become the standard configuration for those tasks with limited training data, which makes it natural to introduce them into the training of CNN-based synthetic aperture radar (SAR) image ship detectors. However, in this paper, we show that the introduction of traditional random cropping methods directly in the training of the CNN-based SAR image ship detector may generate a lot of noise in the gradient during back propagation, which hurts the detection performance. In order to eliminate the noise in the training gradient, a simple and effective training method based on feature map mask is proposed. Experiments prove that the proposed method can effectively eliminate the gradient noise introduced by random cropping and significantly improve the detection performance under a variety of evaluation indicators without increasing inference cost.

Download Full-text