3rd International Conference on Machine Learning, NLP and Data Mining (MLDA 2024)

<!DOCTYPE html> <html> <head>  <link href="https://fonts.googleapis.com/icon?family=Material+Icons" rel="stylesheet"> <link href="https://fonts.googleapis.com/css?family=Roboto+Condensed" rel="stylesheet">  <link type="text/css" rel="stylesheet" href="css/materialize.min.css" media="screen,projection" /> <link rel="stylesheet" href="https://use.fontawesome.com/releases/v5.0.13/css/all.css" integrity="sha384-DNOHZ68U8hZfKXOrtjWvjxusGo9WQnrNx2sqG0tfsghAvtVlRW3tvkXWZh58N9jp" crossorigin="anonymous"> <link type="text/css" rel="stylesheet" href="css/main.css" /> <meta charset="UTF-8">  <meta name="viewport" content="width=device-width, initial-scale=1.0" /> <title>3rd International Conference on Machine Learning, NLP and Data Mining (MLDA 2024)</title> <link rel="icon" type="image/ico" href="img/logo.ico"> </head> <body>  <header class="main-header"> <nav class="transparent"> <div class="container"> <div class="nav-wrapper"> <a href="#" class="brand-logo">MLDA </a> <a href="#" data-activates="mobile-nav" class="button-collapse"> <i class="fa fa-bars"></i> </a> <ul class="right hide-on-med-and-down"> <li> <a href="index">HOME</a> </li> <li> <a href="papersubmission">PAPER SUBMISSION</a> </li> <li> <a href="committee">PROGRAM COMMITTEE</a> </li> <li> <a class="active-link" href="#">ACCEPTED PAPERS</a> </li> <li> <a href="contact">CONTACT US</a> </li> </ul> <ul class="side-nav grey darken-1 white-text" id="mobile-nav"> <h4 class="center">MLDA 2024</h4> <li> <div class="divider"></div> </li> <li> <a href="index"> <i class="fa fa-home white-text"></i>Home </a> </li> <li> <a href="papersubmission"> <i class="fa fa-user white-text"></i>Paper Submission </a> </li> <li> <a href="committee"> <i class="fa fa-user white-text"></i>Program Committee </a> </li> <li> <a class="active-link" href="papers"> <i class="fa fa-newspaper white-text"></i>Accepted Papers </a> </li> <li> <a href="contact"> <i class="fa fa-phone white-text"></i>Contact Us </a> </li> <li> <div class="divider"></div> </li> <li> <a href="/submission/index.php" target="blank" class="btn grey waves-effect waves-light">Paper Submission</a> </li> </ul> </div> </div> </nav>  <div class="showcase container"> <div class="row"> <div class="col s12 m10 offset-m1 center grey-text text-darken-3"> <h5>Welcome to MLDA 2024</h5> <h2>3<sup>rd</sup> International Conference on Machine Learning, NLP and Data Mining (MLDA 2024)</h2> <p>July 13 ~ 14, 2024, Virtual Conference</p> <br> <br> </div> </div> </div> </header> <section class="section section-icons "> <div class="container"> <div class="row"> <div class="col s12 m12"> <div class="card-panel grey darken-2 z-depth-3 white-text center"> <i class="fa fa-paper-plane fa-3x"></i> <h5>Accepted Papers</h5> </div> </div> <div class="col s12 m12"> <div class="card-panel white z-depth-3 "> <h6 style="color:black;font-family:classic wide,sans-cserif;font-size:20px"><b>AI-Powered Solutions for Missing Data in Pipeline Risk Assessments</b></h6> <p style="color:black;text-align:justify;font-size: 15px;">Syed JehanzebAdeel Haider, Enbridge Energy Inc.,Houston, USA</p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>ABSTRACT</b></h6> <p style="color:black;text-align:justify;">The use of Artificial Intelligence (AI) and Machine Learning (ML) in the oil and gas pipeline industry has shown significant promise, particularly in addressing challenges posed by incomplete datasets. This paper explores the application of AI in filling missing data for risk assessments, with a focus on safety-critical scenarios. Through a detailed process flow, this paper illustrates the potential pitfalls and risks associated with relying solely on AI-generated data. This paper also suggests strategies to balance AI reliance with real data acquisition, emphasizing the importance of consequence analysis, cost-benefit considerations, and a hybrid approach to ensure the safety and reliability of operations across the pipeline and broader oil and gas industry. </p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>Keywords</b></h6> <p style="color:black;text-align:justify;">Artificial Intelligence (AI), Machine Learning (ML), Risk Assessment, Pipeline Safety, ALARP</p> <br> <h6 style="color:black;font-family:classic wide,sans-cserif;font-size:20px"><b>Detecting Droplets for Crop Spraying Systems Using Machine Learning</b></h6> <p style="color:black;text-align:justify;font-size: 15px;">Debmalya Ray, India</p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>ABSTRACT</b></h6> <p style="color:black;text-align:justify;">Agricultural Development combined with technology has made great progress in recent years, making it possible to improve the yield for farmers. This project combines deep learning algorithms with spraying technology to design a machine vision precision real-time targeting spraying system for field scenarios. Highly efficient mechanized nozzles are used to spray and apply nutrients and pesticides to crops so that farmers can increase production and mitigate the gap between supplies and demands. We employ high-speed visualization [8] to quantify the impact and evaporation of a droplet on a solid surface. This will also help us to identify the density/area covered by a single spray at a time and correct the delta part left to be covered at first work. This paper is focused on using image classification techniques with a computer vision algorithm to extract the parameters required from a single image at a time and convert it into structured data so that an unsupervised algorithm can cluster the regions based on density. </p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>Keywords</b></h6> <p style="color:black;text-align:justify;">machine vision, image processing methods, unsupervised learning, droplets impact.</p> <br> <h6 style="color:black;font-family:classic wide,sans-cserif;font-size:20px"><b>Web-based Automation Testing and Tools Leveraging AI and ML</b></h6> <p style="color:black;text-align:justify;font-size: 15px;">Narendar Kumar Ale, University Of Cumberlands, United States of America</p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>ABSTRACT</b></h6> <p style="color:black;text-align:justify;">Software testing remains an essential phase of the software development lifecycle particularly for web-based applications. The integration of AI and ML automation testing has reached new heights in efficiency accuracy and coverage. This paper discusses the latest advancements in web automation testing tools that leverage AI and ML providing insights into their benefits and selection criteria. </p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>Keywords</b></h6> <p style="color:black;text-align:justify;">Automation Testing, AI, ML, Web Applications.</p> <br> <h6 style="color:black;font-family:classic wide,sans-cserif;font-size:20px"><b>Robust Multi-modal Face Anti-spoofing: Handling Missing Modalities and Fusion Techniques</b></h6> <p style="color:black;text-align:justify;font-size: 15px;">Zain Ul Abideen<sup>1</sup>, Muhammad Asim<sup>2,3</sup>, and Mohammed A. ELAffendi<sup>3</sup>, <sup>1</sup>School of Computer Science and Engineering, Central South University, Changsha 410083, China, <sup>2</sup>School of Computer Science and Technology, Guangdong University of Technology, Guangzhou 510006, China, <sup>3</sup>EIAS Data Science Lab, College of Computer and Information Sciences, and Center of Excellence in Quantum and Intelligent Computing, Prince Sultan University, Riyadh 11586, Saudi Arabia</p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>ABSTRACT</b></h6> <p style="color:black;text-align:justify;"> Recent advancements in multi-modal learning have significantly enhanced face antispoofing systems. Despite these improvements, real-world applications often face the challenge of missing modalities from different imaging sensors. Previous studies have largely ignored this issue or have increased model complexity without effectively addressing it. This study presents a robust yet straightforward methodology utilizing a multi-modal face anti-spoofing architecture with spatial-temporal encoders and a dedicated fusion unit. The spatial-temporal encoders extract features from each modality using ResNet34 and Transformer architectures, while augmentation and regularization techniques further enhance model performance. Various fusion methods are assessed for their effectiveness in managing missing modalities. Additionally, we present FaceMAE, a modular autoencoder designed to predict and reconstruct missing-modalities. FaceMAE functions via a dual-phase process: encoding detected modalities to produce latent representations and subsequently decoding them to reconstruct missing modalities. Through the incorporation of transformer encoders and a flexible fusion module, FaceMAE enhances the ability to differentiate between live and spoof facial images. Evaluations on datasets such as CASIA-SURF, CASIA-SURF CeFA, and WMCA indicate that our method achieves competitive results. </p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>Keywords</b></h6> <p style="color:black;text-align:justify;">Artificial intelligence,Computer vision, multi-modal learning, face anti-spoofing, missing modality scenarios, face attack detection, Data augmentation, spatial-temporal encoders.</p> <br> <h6 style="color:black;font-family:classic wide,sans-cserif;font-size:20px"><b>A Combine Cnn-rnn Based Approach for Augmenting the Performance of Speech Emotions Recognition</b></h6> <p style="color:black;text-align:justify;font-size: 15px;">Afaq Ahmed, Muaaz Bin Kaleem, Department of Computer Science, Central South University, Changsha, China</p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>ABSTRACT</b></h6> <p style="color:black;text-align:justify;">Due to the advancement of neural networks and the increasing demand for accurate and real-time Speech Emotion Recognition (SER) in human-computer interactions, it is necessary to com-pare existing methods and databases in SER in order to arrive at feasible solutions and a complete understanding of this open-ended problem in SER. To detect and recognize the emotions expressed in speech, various techniques have been used in the literature, including well-established speech analysis and classification techniques. These techniques, including speech analysis and classification, have been used to extract emotions from signals. In this study, we propose a novel method for analysing signals called Wavelet-Scaled Spectrogram which com-bines the frequency and scale spectrum of a signal using wavelet transform. This method is effective in analysing signals at different scales and frequency content. In order to train models for speech emotion identification, a large number of handcrafted features and intermediary depictions i.e., frequency-time plot that have traditionally been utilized in data compilation, collection, and analysis. The development of end-to-end models which extract characteristics and learn directly from raw speech signals to improve speech recognition has recently been studied by re-searchers following the emergence of deep learning. After training and evaluation on the famous speech databases EmoDB, RAVDESS and IEMOCAP, the proposed model is evaluated on various speakers in both speaker-independent and speaker-dependent modes and on a variety of different voices. When advanced preprocessing techniques or data augmentation are omitted from the proposed architecture, the results demonstrate that it can produce products comparable to those produced by the current state of the art. Three concurrent CNN pipelines and a series of modified local features learning blocks (LFLBs) achieved the highest classification accuracy attainable. </p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>Keywords</b></h6> <p style="color:black;text-align:justify;">Artificial Intelligence, Network Intrusion Detection Systems, Machine Learning, Ensemble Models, Cybersecurity, Feature Selection.</p> <br> <h6 style="color:black;font-family:classic wide,sans-cserif;font-size:20px"><b>Enhancing Sentiment Analysis for Low-resource Pashto Language: a Bert-infused Lstm Framework</b></h6> <p style="color:black;text-align:justify;font-size: 15px;">Abdul Hamid Azizi<sup>1</sup>, Muhammad Asim<sup>2</sup>, and Mudasir Ahmad Wani<sup>3</sup>, <sup>1</sup>School of Computer Science and Engineering, Central South University, Changsha 410083 P.R. China, <sup>2</sup>EIAS Data Science Lab, College of Computer and Information Sciences, Prince Sultan University, Riyadh 11586, Saudi Arabia, <sup>3</sup>School of Computer Science and Technology, Guangdong University of Technology, Guangzhou 510006, China</p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>ABSTRACT</b></h6> <p style="color:black;text-align:justify;">In the domain of Natural Language Processing (NLP), sentiment analysis plays a pivotal role, aiding in applications ranging from user feedback interpretation to customer sentiment tracking. However, the focus of sentiment analysis has predominantly centered on major languages such as English, side-lining the requirements of less commonly used languages. In this context, low-resource languages like Pashto encounter a considerable deficiency in sentiment analysis tools and resources, limiting their potential in the field of NLP. To address this critical gap, the proposed work is focused on Pashto sentiment analysis, a low-resource language, and utilizes Romanized Pashto data collected from Twitter using Tweepy, a Python library for accessing the Twitter API. To enhance the data quality, a sequence of pre-processing steps are implemented, encompassing the elimination of extraneous information, stemming, and vectorization. Additionally, in light of the absence of a Pashto language-specific stop words list and stemming dictionary, previous research endeavours have faced limitations. In response, this study also proposed a proactive approach by crafting a comprehensive list of Pashto stop words and constructing a stemming dictionary called PasLex (Pashto Lexicon) with the guidance of domain experts. Using BERT as a tokenizer and combining it with Long Short-Term Memory (LSTM) network leads to improved sentiment detection. We also check the validity of the proposed model by combining BERT with Support Vector Machine (SVM) for sentiment classification. In both cases the proposed approach has outperformed the state-of-the-art methods by achieving highest accuracy, precision, recall, and F1-measure of 92.33%, 92.5%, 92.33%, and 92.32%, respectively. </p> <h6 style="color:black;font-family:classic wide,sans-serif;"><b>Keywords</b></h6> <p style="color:black;text-align:justify;">Sentiment Analysis, Low Resource Languages, Pashto, BERT, LSTM, SVM.</p> <br> </div> </div> </div> </div> </section>   <footer class="page-footer grey lighten-1"> <div class="container"> <div class="row"> <div class="col s12 m6"> <h5 class="grey-text lighten-3"> <font color="#FFF">Contact Us</font> </h5> <a href="mailto:mlda@mlda2024.org" style="color:#000">mlda@mlda2024.org</a> </div> </div> </div> <div class="footer-copyright grey darken-2"> <div class="container center"> Copyright © MLDA 2024 </div> </div> </footer>  <script type="text/javascript" src="https://code.jquery.com/jquery-3.2.1.min.js"></script> <script type="text/javascript" src="js/materialize.min.js"></script> <script> $(document).ready(function() { // Custom JS & jQuery here $('.button-collapse').sideNav(); }); </script> </body> </html>

CINXE.COM

3rd International Conference on Machine Learning, NLP and Data Mining (MLDA 2024)