Digital Commons Network^™

Improving Chinese Hate Speech Detection With Bert-Fasttext Fusion And Bert-Bilstm Fusion, Methini Ma

Chulalongkorn University Theses and Dissertations (Chula ETD)

Hate speech detection is an essential technique in the online environment, especially on social media platforms. This technique helps to create a safer space and reduce the risk of real-world harm. In Chinese, this task is particularly challenging because of unique linguistic structures and the frequent use of indirect expressions, sarcasm, homophones, character variants, and abbreviations. This study investigates how to improve Chinese hate speech detection by combining BERT with FastText and BERT with BiLSTM. There are six model variants that are configured: frozen BERT and fine-tuned BERT, each further extended with either FastText sentence embeddings or a clause-level BiLSTM. …

Multi-Lingual And Cross-Domain Frontiers In Machine-Generated Content Detection, Gurunameh Singh Chhatwal

Theses and Dissertations (Comprehensive)

The rapid advancement of generative artificial intelligence, particularly Large Language Models (LLMs) such as GPT-4 and their multilingual capabilities, has significantly blurred the distinction between human-authored and machine-generated content. This technological evolution introduces critical challenges concerning the detection and attribution of textual authenticity and authorship, exacerbating societal issues like misinformation proliferation and compromising academic and professional integrity. Traditional detection methodologies, predominantly monolingual and heuristic-based, have demonstrated inadequate generalizability and efficacy against the sophisticated, multilingual capabilities of contemporary generative models.

This thesis addresses two major problems arising from these advancements. Firstly, it introduces novel multilingual detection methodologies explicitly designed to differentiate …

Sparse Transformer For Anomaly Detection In Mobile Crowdsensing, Sanjeev Shrestha

Graduate Theses/Dissertations

Mobile Crowdsensing (MCS) is a sensing paradigm that leverages mobile devices to conduct a large-scale data collection. However, due to its openness and mobility nature, it is highly vulnerable to security issues such as injection attacks of malicious workers and fake tasks that can severely affect the platform’s normal functioning. To address this problem, the arrival of workers and task submission process is represented as a multivariate time series, and a two-stage framework is proposed. In the first step, we propose a novel transformer-based model, DozerAnomaly, that can efficiently detect anomalies in multivariate time series. We integrated a sparse attention …

การคาดการณ์ค่าดัชนีความแตกต่างของพืชและความชื้น โดยใช้เทคนิคการเรียนรู้ของเครื่องอย่างชาญฉลาด เพื่อการติดตามด้านการเกษตรและสิ่งแวดล้อม, ณุทยา เข็มเจริญ

Chulalongkorn University Theses and Dissertations (Chula ETD)

ปัจจุบัน ภัยแล้งและการเปลี่ยนแปลงสภาพภูมิอากาศส่งผลกระทบต่อผลผลิตทางการเกษตร โดยเฉพาะพืชเศรษฐกิจของไทยอย่างทุเรียน ซึ่งต้องอาศัยการติดตามสุขภาพพืชและการจัดการน้ำอย่างใกล้ชิด ข้อมูลจากภาพถ่ายดาวเทียม ดัชนีความแตกต่างของพืชและความชื้น จึงเป็นเครื่องมือสำคัญในการวิเคราะห์สภาพพื้นที่เพาะปลูก โครงงานนี้พัฒนาแบบจำลองการเรียนรู้ของเครื่องเพื่อพยากรณ์ทั้งสองดัชนีรายสัปดาห์ โดยใช้การเฉลี่ยและเติมค่าข้อมูลที่ขาดหายด้วยการถดถอยเชิงเส้น ก่อนนำไปฝึกแบบจำลองผสมระหว่างโครงข่ายคอนโวลูชัน และหน่วยความจำระยะสั้นแบบยาว ระบบถูกออกแบบให้ทำงานอัตโนมัติภายใต้แนวคิดการปฏิบัติการด้านการเรียนรู้ของเครื่อง แสดงผลผ่านเว็บแอปพลิเคชัน ติดตั้งและทดสอบบนเครื่องเซิร์ฟเวอร์

Thyroid Carcinoma Recurrence Prediction Using Artificial Intelligence On Prognosis Parameter, Supawat Suntornlimsiri

Chulalongkorn University Theses and Dissertations (Chula ETD)

Papillary thyroid carcinoma (PTC) is the most common subtype of thyroid malignancy and is increasingly diagnosed worldwide. Although typically localised, PTC exhibits a notable risk of bilateral involvement, with contralateral disease occurring in up to 44 percent of cases. While completion thyroidectomy is recommended in selected high-risk scenarios, it carries potential complications, making accurate prediction of contralateral involvement essential. This study investigates the utility of classical statistical methods and machine learning (ML) algorithms—Support Vector Machines (SVM), Extreme Gradient Boosting (XGBoost), and Random Forest (RF)—in predicting contralateral PTC using a retrospective dataset of 122 lobectomy patients. ML models, especially RF with …

Design And Evaluation Of A Thai Speech Emotion Recognition Corpus With Ambiguous Annotations, Chompakorn Chaksangchaichot

Chulalongkorn University Theses and Dissertations (Chula ETD)

THAI-SER is the first large-scale Thai speech emotion recognition corpus, comprising 41.6 hours (27,854 utterances) from 100 recordings across diverse environments (Zoom and studio). The data includes both scripted and improvised speech by 200 professional actors (112 females, 88 males, aged 18–55), covering five emotions: neutral, angry, happy, sad, and frustrated. Utterances were labeled via crowdsourcing, with rigorous quality control ensuring a majority agreement score above 0.71. Annotation reliability, measured by Krippendorff’s alpha, reached 0.692 (above the 0.667 threshold), and human emotion recognition accuracy reached 0.772 after filtering. We also report benchmark results from models trained and evaluated on both …

ปัจจัยที่ส่งผลต่อความตั้งใจใช้งานโมไบล์แบงก์กิ้งอย่างต่อเนื่อง, รตน รัตนพรเจริญ

Chulalongkorn University Theses and Dissertations (Chula ETD)

การใช้บริการโมไบล์แบงก์กิ้ง ในประเทศไทยมีอัตราเติบโตอย่างต่อเนื่องภายใต้การสนับสนุนจากภาครัฐ แต่ในช่วงหลายปีที่ผ่านมากลับพบว่าผู้ใช้จำนวนมากยังคงใช้งานเพียงบางฟังก์ชันพื้นฐาน เช่น การโอน-จ่าย-รับเงิน เท่านั้น (ธนาคารแห่งประเทศไทย, 2567) และประสบการณ์กับมิจฉาชีพและความเชื่อมั่นในระบบก็ยังคงเป็นประเด็นสำคัญที่สะท้อนให้เห็นถึงพฤติกรรมหลังการยอมรับและใช้งานโมไบล์แบงกิ้งกับการเปิดใจใช้บริการการเงินดิจิทัลอื่น ๆ ของผู้ใช้บริการบริโภค ซึ่งยังขาดการศึกษาในเชิงลึกโดยเฉพาะด้านปัจจัยที่ส่งผลต่อความตั้งใจใช้งานอย่างต่อเนื่อง และการเปิดรับฟังก์ชันใหม่ทั้งจากผู้ให้บริการโมไบล์แบงกิ้งและผู้ให้บริการอื่น ๆ ภายใต้ บริบทของการรับรู้ประโยชน์, และคุณภาพการบริการงานวิจัยนี้มีวัตถุประสงค์เพื่อศึกษาปัจจัยที่ส่งผลต่อความตั้งใจใช้งานโมไบล์แบงก์กิ้งอย่างต่อเนื่อง ความตั้งใจใช้ฟังก์ชันที่ไม่เคยใช้งานและความตั้งใจใช้งานร่วมกับแอปพลิเคชันทางการเงินอื่น โดยอิงกรอบแนวคิดจากทฤษฎีความตั้งใจใช้งานต่อเนื่อง (Continuance Intention Model) รวมถึงการวิเคราะห์ผลจากประสบการณ์กับมิจฉาชีพต่างกัน กลุ่มตัวอย่างทั้งหมดจำนวน 434 คนถูกรวบรวมโดยการสุ่มแบบเจาะจง (Purposive Sampling) จากผู้ใช้โมไบล์แบงก์กิ้งในกรุงเทพมหานครและปริมณฑล ข้อมูลถูกวิเคราะห์ด้วยสถิติเชิงพรรณนาและการถดถอยเชิงพหุ (Multiple Regression Analysis)โดยผลการศึกษาพบว่า ปัจจัยด้านการรับรู้ประโยชน์ ความพึงพอใจ และความเชื่อมั่นในผู้ให้บริการ ส่งผลเชิงบวกต่อความตั้งใจในการใช้งานต่อเนื่องและการเปิดรับฟังก์ชันใหม่ทั้งจากธนาคารและผู้ให้บริการภายนอก ในขณะที่ประสบการณ์กับมิจฉาชีพมีแนวโน้มส่งผลให้ความตั้งใจใช้งาน แต่ไม่ส่งผลต่อการเปิดรับฟังก์ชันใหม่ การศึกษาปัจจัยด้านคุณภาพสารสนเทศทั้งสามมิติพบผลกระทบเชิงบวกต่อความพึงพอใจของผู้ใช้ และการศึกษาด้านความเชื่อมั่นพบผลกระทบเชิงบวกจากความสามารถ ความหวังดี และความซื่อสัตย์ของผู้ให้บริการ ผลการวิจัยนี้สามารถนำไปใช้พัฒนาแนวทางในการส่งเสริมการใช้งานฟังก์ชันของแอปพลิเคชันทางการเงินให้ครอบคลุมยิ่งขึ้น ทั้งในมิติการออกแบบระบบและผลิตภัณฑ์ทางการเงิน การสร้างความเชื่อมั่น และการออกนโยบายจัดการความเสี่ยงจากมิจฉาชีพ อีกทั้งยังสามารถขยายผลเชิงวิชาการในด้านการประยุกต์ใช้โมเดลพฤติกรรมผู้บริโภคในบริบทเทคโนโลยีทางการเงิน ซึ่งยังเป็นพื้นที่วิจัยที่เปิดกว้างและท้าทายในประเทศไทย

Robust Palm Print For Mobile Authentication System, Son Nguyen

Chulalongkorn University Theses and Dissertations (Chula ETD)

Smartphones are gateways to financial, health, and personal data; consequently, mobile authentication must be accurate, fast, privacy-preserving, and scalable. This thesis presents an end-to-end palmprint authentication framework that addresses a practical trilemma: label-efficient learning, on-device efficiency, and cloud-scale identification. We pretrain a ResNet-18 encoder with self-supervised contrastive learning on unlabeled palm images, distill its representation to a lightweight MobileNetV3 student for real-time inference on phones, and support both 1:1 on-device verification and 1:N cloud identification using FAISS/HNSW. On public datasets, the system attains 99.2% accuracy, a 0.15% equal-error rate (EER), and ~87 ms end-to-end latency on iPhone-class hardware. FAISS scales …

Scam Slayer : A Gamification For Creating A Cyber Scam Awareness, Xin Lyu

Chulalongkorn University Theses and Dissertations (Chula ETD)

This study explores how gamified learning can enhance the ability to recognize and prevent online fraud. The research team developed a role-playing educational game, Scam Slayer, using Ren’Py, integrating five common fraud scenarios and an Anti-Fraud Assistant to create an immersive and reflective learning experience. Analysis of a pre-game survey with 400 participants and a post-game survey with 40 players shows significant improvements in fraud recognition and alertness, with no demographic differences in learning outcomes. Participants also reported high satisfaction with the game’s narrative and educational value, suggesting that gamified learning can serve as an effective complement to traditional online …

การเพิ่มประสิทธิภาพการจัดตารางการอัดประจุรถยนต์ไฟฟ้าโดยอาศัยข้อมูลโปรไฟล์แบตเตอรี่และเทคนิคการเรียนรู้ของเครื่อง, กนกภรณ์ ปัญจันทร์สิงห์

Chulalongkorn University Theses and Dissertations (Chula ETD)

งานวิจัยนี้มุ่งเพิ่มประสิทธิภาพการจัดตารางการอัดประจุรถยนต์ไฟฟ้าโดยอาศัยข้อมูลโปรไฟล์แบตเตอรี่และเทคนิคการเรียนรู้ของเครื่อง เพื่อแก้ปัญหาความไม่สอดคล้องระหว่างแผนการอัดประจุกับพฤติกรรมการอัดประจุจริงของผู้ใช้ โดยเริ่มจากการพัฒนาแบบจำลองการทำนาย (Predictive Model) เพื่อคาดการณ์ความต้องการพลังงาน ระยะเวลาการอัดประจุ และสถานะการประจุเริ่มต้นของแบตเตอรี่ (State of Charge: SOC) จากข้อมูลการอัดประจุในอดีต จากนั้นได้ทำการสร้างโปรไฟล์การอัดประจุของแบตเตอรี่ที่เป็นตัวแทนของรถยนต์แต่ละรุ่น เพื่อสะท้อนลักษณะการอัดประจุจริงทั้งในกรณีของเครื่องอัดประจุแบบกระแสสลับ (AC) และกระแสตรง (DC) แล้วนำข้อมูลดังกล่าวมาใช้ในการเพิ่มประสิทธิภาพการจัดตารางการอัดประจุของสถานีอัดประจุ ผลการจำลองแสดงให้เห็นว่าวิธีการที่เสนอ (Optimization with Battery Profile) ให้ประสิทธิภาพสูงกว่าวิธีการทั่วไปที่ไม่ใช้ข้อมูลโปรไฟล์แบตเตอรี่ โดยสามารถเพิ่มกำไรของสถานีได้สูงสุด 3,145.3 บาท และเพิ่มอัตราการใช้พลังงานของระบบได้ร้อยละ 4.8 ในกรณีที่มีสัดส่วนเครื่องอัดประจุ DC แตกต่างกัน อีกทั้งยังให้ผลลัพธ์ที่ดีอย่างต่อเนื่องภายใต้ข้อจำกัดกำลังพีคของสถานี โดยเพิ่มกำไรได้สูงสุด 3,649.25 บาท และเพิ่มอัตราการใช้ประโยชน์ของระบบได้ร้อยละ 9.6 สะท้อนถึงประสิทธิภาพของแนวทางที่เสนอในการจัดการพลังงานอย่างชาญฉลาดสำหรับสถานีอัดประจุรถยนต์ไฟฟ้าในอนาคต

Secure Federated Learning Via Neural Cryptography With Homomorphic Operations, Espen Sele, Ferhat Ozgur Catak, Jungwon Seo, Murat Kuzlu

Engineering Technology Faculty Publications

This study examines neural cryptography with homomorphic operations as an alternative secure aggregation method for federated learning (FL). It proposes a novel neural cryptographic system supporting homomorphic addition on fixed-point encrypted data, and consisting of three networks, namely (1) an encryption network (Alice), (2) a homomorphic network (HO), and (3) a decryption network (Bob), along with an adversarial Eve network. Using the MNIST dataset, the proposed Neural Homomorphic Operation System (NHOS) is evaluated against a plaintext baseline and the CKKS scheme, a widely used public-key homomorphic encryption method. The results show that the proposed NHOS approach offers a satisfying performance, …

Machine Learning Models For Pancreatic Cancer Survival Prediction: A Multi-Model Analysis Across Stages And Treatments Using The Surveillance, Epidemiology, And End Results (Seer) Database, Aditya Chakraborty, Mohan D. Pant

Epidemiology, Biostatistics, & Environmental Health Faculty Publications

Background: Pancreatic cancer is among the most lethal malignancies, with poor prognosis and limited survival despite treatment advances. Accurate survival modeling is critical for prognostication and clinical decision-making. This study had three primary aims: (1) to determine the best-fitting survival distribution among patients diagnosed and deceased from pancreatic cancer across stages and treatment types; (2) to construct and compare predictive risk classification models; and (3) to evaluate survival probabilities using parametric, semi-parametric, non-parametric, machine learning, and deep learning methods for Stage IV patients receiving both chemotherapy and radiation. Methods: Using data from the SEER database, parametric models (Generalized Extreme Value, …

An Integrated Machine Learning Approach For Identifying Emergency Rescue Messages On Social Media During Natural Disasters, Wael Khallouli, Jiang Li, Jingwei Huang, Ghaith Rabadi, Samuel Kovacic

School of Cybersecurity Faculty Publications

During large-scale disasters, emergency call centers are often overwhelmed by the large volume of rescue requests and calls for help. Consequently, people are turning to social media platforms to seek assistance. Rescue information posted on these platforms is extremely valuable for first responders to make informed rescue decisions. Therefore, the automatic identification of these requests from the vast amount of data posted on social media during crises is critical yet challenging. This work presents our ongoing research on applying deep learning techniques to extract actionable rescue information from social media during crises. We proposed a novel deep learning model that …

Fostering Critically Conscious Lesson Planning In A Generative Artificial Intelligence Era, Derek Riddle, Paula Cristina Azevedo, Catharyn Shelton, Jaime Colwell, Jori Beck

Teaching & Learning Faculty Publications

Teacher candidates (TCs) use digital resources and social media to plan and develop learning material, and with publically accessible generative artificial intelligence (GAI), TCs are able to generate lesson plans within seconds rather than hours or days. While there is research on how to support TCs' evaluation of reliable digital media, there is little known on how to prepare TCs for GAI content. Using the complementary frameworks of Freire’s (1970) critical consciousness and Jonnasen’s (1991) theory of constructivism, this in-progress design based research seeks to develop an adaptable framework that addresses the evolving nature of technology, specifically GAI, and the …

An Analytical Model Of Motion Artifacts In A Measured Arterial Pulse Signal—Part I: Accelerometers And Ppg Sensors, Md Mahfuzur Rahman, Subodh Toraskar, Mamun Hasan, Zhili Hao

Mechanical & Aerospace Engineering Faculty Publications

This paper, the first of two parts, presents an analytical model of motion artifacts (MAs) in measured pulse signals by accelerometers and photoplethysmography (PPG) sensors. As the transmission path from the true pulse signal in an artery to the sensor output (measured pulse signal), the tissue-contact-sensor (TCS) stack is modeled as a 1DOF (degree-of-freedom) system. MAs cause baseline drift of the mass and simultaneously time-varying system parameters (TVSPs) of the TCS stack. With arterial wall displacement and pulsatile pressure serving separately as the true pulse signal, an analytical model is developed to mathematically relate baseline drift and TVSP to a …

An Analytical Model Of Motion Artifacts In A Measured Arterial Pulse Signal—Part Ii: Tactile Sensors, Md Mahfuzur Rahman, Subodh Toraskar, Mamun Hasan, Zhili Hao

Mechanical & Aerospace Engineering Faculty Publications

This paper, the second of two parts, presents an analytical model of motion artifacts (MA) in measured pulse signals by a tactile sensor, which contains a deformable microstructure sitting on a substrate. While the tissue-contact-sensor (TCS) stack and the sensor are both treated as a 1DOF (degree-of-freedom) system, tissue–sensor contact joins their mass together to form a 1DOF system with springs and dampers on both sides. MA on the sensor substrate causes baseline drift and time-varying system parameters (TVSP) of the TCS stack simultaneously. An analytical model is developed to mathematically relate baseline drift and TVSP to a measured pulse …

Motion Artifacts Removal From Measured Arterial Pulse Signals At Rest: A Generalized Sdof-Model-Based Time-Frequency Method, Zhili Hao

Mechanical & Aerospace Engineering Faculty Publications

Motion artifacts (MA) are a key factor affecting the accuracy of a measured arterial pulse signal at rest. This paper presents a generalized time–frequency method for MA removal that is built upon a single-degree-of-freedom (SDOF) model of MA, where MA is manifested as time-varying system parameters (TVSPs) of the SDOF system for the tissue–contact-sensor (TCS) stack between an artery and a sensor. This model distinguishes the effects of MA and respiration on the instant parameters of harmonics in a measured pulse signal. Accordingly, a generalized SDOF-model-based time–frequency (SDOF-TF) method is developed to obtain the instant parameters of each harmonic in …

Quantifying Multidimensional Effects Of Physicochemical Parameters On Pfas Adsorption Using A Hybrid Response Surface Methodology-Machine Learning Approach, Harsh V. Patel, Jazmin Green, John Park, Stephanie Luster-Teasley Pass, Renzun Zhao

Engineering Management & Systems Engineering Faculty Publications

Per- and polyfluoroalkyl substances (PFAS) contamination has posed a significant environmental and public health challenge due to their ubiquitous nature. Adsorption has emerged as a promising remediation technique, yet optimizing adsorption efficiency remains complex due to the diverse physicochemical properties of PFAS and the wide range of adsorbent materials. Traditional modeling approaches, such as response surface methodology (RSM), struggled to capture nonlinear interactions, while standalone machine learning (ML) models required extensive datasets. This study addressed these limitations by developing hybrid RSM-ML models to improve the prediction and optimization of PFAS adsorption. A comprehensive dataset was constructed using experimental adsorption data, …

Zero Trust Architecture As A Risk Countermeasure In Small-Medium Enterprises And Advanced Technology Systems, Ahmed M. Abdelmagid, Rafael Diaz

Engineering Management & Systems Engineering Faculty Publications

The growing sophistication of cyberattacks exposes small- and medium-sized businesses (SMBs) to a widening range of security risks. As these threats evolve in complexity, the need for advanced security measures becomes increasingly pressing. This necessitates a proactive approach to defending against potential cyber intrusions. Emerging technologies, such as blockchain, artificial intelligence, and Zero Trust security framework, offer crucial tools for strengthening the digital infrastructure of SMBs. The Zero Trust architecture (ZTA) holds significant promise as a critical strategy for protecting SMBs. While existing literature explores the implementation of ZTA in various business settings, discussions specifically addressing the financial, human resource, …

Not Here, Go There: Analyzing Redirection Patterns On The Web, Kritika Garg, Sawood Alam, Dietrich Ayala, Michele C. Weigle, Michael L. Nelson

Computer Science Faculty Publications

URI redirections are integral to web management, supporting structural changes, SEO optimization, and security. However, their complexities affect usability, SEO performance, and digital preservation. This study analyzed 11 million unique redirecting URIs, following redirections up to 10 hops per URI, to uncover patterns and implications of redirection practices. Our findings revealed that 50% of the URIs terminated successfully, while 50% resulted in errors, including 0.06% exceeding 10 hops. Canonical redirects, such as HTTP to HTTPS transitions, were prevalent, reflecting adherence to SEO best practices. Non-canonical redirects, often involving domain or path changes, highlighted significant web migrations, rebranding, and security risks. …

Gramseq-Dta: A Grammar-Based Drug-Target Affinity Prediction Approach Fusing Gene Expression Information, Kasul Debnath, Pratip Rana, Preetam Ghosh

Computer Science Faculty Publications

Drug–target affinity (DTA) prediction is a critical aspect of drug discovery. The meaningful representation of drugs and targets is crucial for accurate prediction. Using 1D string-based representations for drugs and targets is a common approach that has demonstrated good results in drug–target affinity prediction. However, these approach lacks information on the relative position of the atoms and bonds. To address this limitation, graph-based representations have been used to some extent. However, solely considering the structural aspect of drugs and targets may be insufficient for accurate DTA prediction. Integrating the functional aspect of these drugs at the genetic level can enhance …

The Invisible Influencer In Information Infrastructure, Herbert Van De Sompel, Michael L. Nelson

Computer Science Faculty Publications

The UPS Prototype was a proof-of-concept web portal built in preparation for the Universal Preprint Service Meeting held in October 1999 in Santa Fe, New Mexico. The portal provided search functionality for a set of metadata records that had been aggregated from a range of repositories that hosted preprints, working papers, and technical reports. Every search result was overlaid with a dynamically generated menu, called an SFX-menu, that provided a selection of value-adding links for the described scholarly work. The meeting eventually led to the Open Archives Initiative and its Protocol for Metadata Harvesting (OAI-PMH), which remains widely used in …

Github Repository Complexity Leads To Diminished Web Archive Availability, David Calano, Michael Nelson, Michele Weigle

Computer Science Faculty Publications

Software is often developed using versioned controlled software, such as Git, and hosted on centralized Web hosts, such as GitHub and GitLab. These Web hosted software repositories are made available to users in the form of traditional HTML Web pages for each source file and directory, as well as a presentational home page and various descriptive pages. We examined more than 12,000 Web hosted Git repository project home pages, primarily from GitHub, to measure how well their presentational components are preserved in the Internet Archive, as well as the source trees of the collected GitHub repositories to assess the extent …

Coldstartcpi: Induced-Fit Theory-Guided Dti Predictive Model With Improved Generalization Performance, Qichang Zhao, Haochen Zhao, Linyuan Gao, Kai Zheng, Yajie Li, Qiao Ling, Jing Tang, Yaohang Li, Jianxin Wang

Computer Science Faculty Publications

Predicting compound-protein interactions (CPIs) plays a crucial role in drug discovery. Traditional methods, based on the key-lock theory and rigid docking, often fail with novel compounds and proteins due to their inability to account for molecular flexibility and the high sparsity of CPI data. Here, we introduce ColdstartCPI, a framework inspired by induced-fit theory, which leverages unsupervised pre-training features and a Transformer module to learn both compound and protein characteristics. ColdstartCPI treats proteins and compounds as flexible molecules during inference, aligning with biological insights. It outperforms state-of-the-art sequence-based models, particularly for unseen compounds and proteins, and shows strong generalization capability …

Optiselect And Enshap: Integrating Machine Learning And Game Theory For Ischemic Stroke Prediction, Pritam Chakraborty, Anjan Bandyopadhyay, Sricheta Parul, Sujata Swain, Partha Sarathy Banerjee, Tapas Si, Hong Qin, Saurav Mallik

Computer Science Faculty Publications

Stroke analysis using game theory and machine learning techniques. The study investigates the use of the Shapley value in predictive ischemic brain stroke analysis. Initially, preference algorithms identify the most important features in various machine learning models, including logistic regression, K-nearest neighbor, decision tree, support vector machine (linear kernel), support vector machine ( RBF kernel), neural networks, etc. For each sample, the top 3, 4, and 5 features are evaluated and selected to evaluate their performance. The Shapley value method was used to rank the models using their best four features based on their predictive capabilities. As a result, better-performing …

Sting: A Stealthy Backdoor Attack On Gnn-Based Malicious Domain Detection Via Dns Perturbations, Muhammad Anan, Mahmoud Nazzal, Abdallah Khreishah, Issa Khalil, Nhathai Phan, Ahmad Sawalmeh

Computer Science Faculty Publications

Detecting malicious Internet domains is essential for safeguarding against various online threats. The current approach to detecting malicious domains (MDD) employs a graph neural network (GNN) method, which uses DNS logs to construct heterogeneous graphs for determining the maliciousness of unknown domains. Despite its success, this method is vulnerable to data poisoning attacks where an adversary can manipulate specific graph nodes to implant a backdoor into the model during training. To showcase the vulnerability, we propose a stealthy trigger injection attack on node features and graph structure in MDD, dubbed (STING). The attacker carefully manipulates selected features and edges of …

Normalizing Images In Various Weather And Lighting Conditions Using Colorpix2pix Generative Adversarial Network, Sanjida Tasnim, Ashif Mahmud Mostafa, Azmain Morshed, Namreen Shaiyaz, Shakib Mahmud Dipto, Saad Aloteibi, Mohammad Ali Moni, Md. Golam Rabiul Alam, Md. Ashraful Alam

Computer Science Faculty Publications

Autonomous vehicles (AVs) are widely regarded as the future of transportation due to their tremendous benefits and user comfort. However, the AVs have been struggling with very crucial challenges, such as achieving reliable accuracy in object detection as well as faster computation required for quick decision-making. In recent years, perception systems in driverless cars have been significantly enhanced, mainly due to advances in deep-learning-based object detection systems. However, these perception systems are still heavily affected by environmental variables, such as changes in illumination, refractive interference, and adverse weather conditions, which may compromise their reliability and safety. This research proposes an …

A Survey On Deep Learning For Drug-Target Binding Prediction: Models, Benchmarks, Evaluation, And Case Studies, Kusal Debnath, Pratip Rana, Preetam Ghosh

Computer Science Faculty Publications

Conventional drug discovery is expensive, time-consuming, and prone to failure. Artificial intelligence has become a potent substitute over the last decade, providing strong answers to challenging biological issues in this field. Among these difficulties, drug-target binding (DTB) is a key component of drug discovery techniques. In this context, drug-target affinity and drug–target interaction are complementary and essential frameworks that work together to improve our comprehension of DTB dynamics. In this work, we thoroughly analyze the most recent deep learning models, popular benchmark datasets, and assessment metrics for DTB prediction. We look at the paradigm shift in the development of drug …

Insights In Adaptation: Examining Self-Reflection Strategies Of Job Seekers With Visual Impairments In India, Akshay Kolgar Nayak, Yash Prakash, Sampath Jayarathna, Hae-Na Lee, Vikas Ashok

Computer Science Faculty Publications

Significant changes in the digital employment landscape, driven by rapid technological advancements and the COVID-19 pandemic, have introduced new opportunities for blind and visually impaired (BVI) individuals in developing countries like India. However, a significant portion of the BVI population in India remains unemployed despite extensive accessibility advancements and job search interventions. Therefore, we conducted semi-structured interviews with 20 BVI persons who were either pursuing or recently sought employment in the digital industry. Our findings reveal that despite gaining digital literacy and extensive training, BVI individuals struggle to meet industry requirements for fulfilling job openings. While they engage in self-reflection …

Coming Back Differently: An Exploratory Case Study Of Near Death Experiences Of Webpages, Lesley Frew, Michael L. Nelson, Michele Weigle

Computer Science Faculty Publications

In this case study, we use web archives to analyze 8,824 webpages that were taken offline and subsequently put back online, thus experiencing a “near death experience.” We enumerate the stages of a webpage’s near death experience, including the change from a successful HTTP status code to non-successful and back, the intermediate stage with markers such as an under construction banner, and an analysis of how the pages came back differently.