ICAICTA2026 Program
Each presentation is allocated 15 minutes, consisting of 10 minutes for the presentation, 4 minutes for Q&A, and 1 minute for transition to the next presenter. Presenters are requested to adhere strictly to the allocated time to ensure the programme runs smoothly. All times in this program are set according to Malaysia Time (UTC+8).
General Schedule
| Time | Programme |
|---|---|
| Day 1 · Tuesday, 20 October 2026 | |
| 8:00 – 8:30 AM | Registration |
| 8:30 – 8:40 AM | Opening by MC |
| 8:40 – 8:50 AM | Welcoming RemarksProf. Ts. Dr. Liew Soung Yue, Chair, ICAICTA 2026 |
| 8:50 – 9:00 AM | Welcoming Guest SpeechDr. Lim Kok Cheng, IEEE Computer Society Malaysia Chapter Representative |
| 9:00 – 9:10 AM | Opening SpeechProf. Dr. Zuraidah Binti Abd Manaf, Vice President (R&D), UTAR |
| 9:10 – 9:25 AM | Gift Presentation and Photo Session |
| 9:25 – 10:25 AM | Keynote Speech 1Assoc. Prof. Ir. Dr. Rila Mandala |
| 10:25 – 10:45 AM | Refreshment Break at A009 |
| 10:45 – 11:45 AM | Keynote Speech 2Assoc. Prof. Dr. Kyoko Hine |
| 11:45 AM – 12:00 PM | Transition / Networking |
| 12:00 – 1:15 PM | Lunch Break at A009 |
| 1:15 – 3:00 PM | Presentation Parallel Sessions 1, 2 and 3Session 1: A001Session 2: A003Session 3: A011 Join online |
| 3:00 – 3:20 PM | Refreshment Break at A009 |
| 3:20 – 5:05 PM | Presentation Parallel Sessions 4, 5 and 6Session 4: A001Session 5: A003Session 6: A011 Join online |
| 5:05 – 5:15 PM | Day 1 Closing & Announcements |
| Day 2 · Wednesday, 21 October 2026 | |
| 8:00 – 8:30 AM | Registration |
| 8:30 – 9:30 AM | Keynote Speech 3Prof. Minoru Sasaki |
| 9:30 – 9:40 AM | Transition / Networking |
| 9:40 – 10:00 AM | Refreshment Break at A009 |
| 10:00 AM – 12:00 PM | Presentation Parallel Sessions 7, 8 and 9Session 7: A001Session 8: A003 Join onlineSession 9: A011 Join online |
| 12:00 – 1:15 PM | Lunch Break at A009 |
| 1:15 – 3:00 PM | Presentation Parallel Sessions 10, 11 and 12Session 10: A001Session 11: A003 Join onlineSession 12: A011 Join online |
| 3:00 – 3:20 PM | Refreshment Break at A009 |
| 3:20 – 4:50 PM | Presentation Parallel Sessions 13 and 14Session 13: A001Session 14: A011 Join online |
| 4:50 – 5:00 PM | End of Conference |
All times are Malaysia Time (GMT+8). Rooms are in Block A, Heritage Hall, Universiti Tunku Abdul Rahman (UTAR), Kampar. Refreshments and lunch are served in A009.
Parallel Sessions Schedule
| Time | Paper ID | Title |
|---|---|---|
| Day 1 · Tuesday, 20 October 2026 | ||
| 1:15 – 1:30 PM | 22 | Simulation-Based Evaluation of Portfolio Construction Strategies: Evidence for Robustness over Optimisation Complexity |
| 1:30 – 1:45 PM | 135 | Indonesian – Sundanese Code-Switching Speech Recognition and Normalization |
| 1:45 – 2:00 PM | 149 | Bridging Physical and Conversational Interaction: Design and Evaluation of an IoT- and RAG-Enabled Information Robot for Smart Campus Services |
| 2:00 – 2:15 PM | 103 | Maximizing Direct Excerpts towards Verifiable Abstractive Summarization of Local Assembly Debates |
| 2:15 – 2:30 PM | 69 | Explainable Traffic Volume Forecasting: A Comparison of Predefined and Random Search Optimization Strategies |
| 2:30 – 2:45 PM | 82 | Hyperspectral Anomaly Detection for Agri-food Industry Inspection using 3D Convolutional Autoencoder and PatchCore Algorithm |
| 2:45 – 3:00 PM | 61 | Multi-Agent Reinforcement Learning for Portfolio Management on the Indonesia Stock Exchange: A Microstructure-Aware Ablation Study |
| 1:15 – 1:30 PM | 118 | Missing-Part Detection and AR Overlay of Damaged Statues |
| 1:30 – 1:45 PM | 24 | Logo Variety Classification with Convolutional Neural Networks for Automated Brand Asset Management |
| 1:45 – 2:00 PM | 102 | Information-Gain-Based Question Selection for Pharmaceutical Search Dialogue Systems |
| 2:00 – 2:15 PM | 70 | Scaling Behavior and Accuracy Trade-offs of Distributed Data Parallel for Volumetric Medical Image Segmentation |
| 2:15 – 2:30 PM | 50 | Voice Authentication with Text-to-Speech: Enhancing Accessibility and Security in Biometric QR Payment Systems |
| 2:30 – 2:45 PM | 66 | Adapting MARS to Intraday Trading on the Indonesian Stock Exchange: Online Meta-Controller Collapse and a Variance-Floor Fix |
| 2:45 – 3:00 PM | 52 | Real-Time Automation of Economic Value Estimation of Inorganic Waste Using Instance Segmentation (Mask R-CNN) and Market Price Integration |
| 1:15 – 1:30 PM | 119 | Wearable IMU-Based TUG Phase Recognition: Temporal Formulation and Boundary-Centered Analysis |
| 1:30 – 1:45 PM | 84 | Dynamic Multimodal Emotion Recognition via Gated Cross-Attention Fusion |
| 1:45 – 2:00 PM | 178 | Comparative Feature-Fusion Strategies for LSTM Autoencoder–Based Indonesian Speech Deepfake Detection |
| 2:00 – 2:15 PM | 132 | Adapting Chain of Retrieval-Augmented Generation for Indonesian Fact-Checking |
| 2:15 – 2:30 PM | 49 | Mitigating Hallucination in Biomedical RAG-LLM using Acronym Expansion, Query Rewriting, and Context Reranking on a Biomedical Dataset |
| 2:30 – 2:45 PM | 30 | Hybrid Ensemble of Transformer and Classical Models for Cross-Domain Misinformation Classification |
| 2:45 – 3:00 PM | 140 | Detection of Font and Ink Color Manipulation in Digital Documents Using Hybrid Line Localization and Inter-Word Deviation Analysis |
| 3:20 – 3:35 PM | 181 | SIGMA: An ESP32-Based IoT System for Real-Time Soil Monitoring and Decision Support in Precision Shallot Farming |
| 3:35 – 3:50 PM | 153 | Designing A Decentralized Teleradiology System Using Adaptive DWT-SVD Steganography Based on Blockchain and Zero-Knowledge Proof for Medical Image Record Security |
| 3:50 – 4:05 PM | 165 | Investigating the Granularity of Named Entity Understanding in Large Language Models |
| 4:05 – 4:20 PM | 65 | Regime-Aware Multi-Agent Reinforcement Learning for Portfolio Management on the Indonesian Stock Exchange |
| 4:20 – 4:35 PM | 126 | Integrating Speech and Pointing Gestures Using a Vision Language Model for Geometry Problem Solving |
| 4:35 – 4:50 PM | 51 | Refusing with the Evidence in Context: A Decision-Boundary Account of Over-Refusal in a Deployed Scientific RAG System |
| 3:20 – 3:35 PM | 39 | Cosine Similarity of CLIP Embeddings for Automated Assessment of Brand Uniqueness and Visual Replication in AI-Generated Logos |
| 3:35 – 3:50 PM | 106 | NLP-Based Customer Service Chatbot with Generative AI Fallback for HRIS Platforms |
| 3:50 – 4:05 PM | 130 | Mapping the Conceptual Structure of Online Beauty Shopping Research Using Co-word Networks and Interpretive Synthesis |
| 4:05 – 4:20 PM | 129 | Enriching Sparse and Imbalanced HR Data for Employee Resignation Risk Detection through Feature Engineering and Resampling |
| 4:20 – 4:35 PM | 57 | Reducing Identity Switches in Soccer Player Tracking: Deep OC-SORT and the Primacy of Camera Motion Compensation |
| 4:35 – 4:50 PM | 124 | Optimal Site Selection of Data Center for Distribution Main Control Center and Disaster Recovery Center Infrastructure using Analytical Hierarchy Process and Geographic Information System Methods |
| 4:50 – 5:05 PM | 128 | Reusing Existing Bus CCTV for Passenger Flow Counting |
| 3:20 – 3:35 PM | 86 | Noise-Aware Domain-Adapted T5-Small Model for Abstractive Summarization of Social Media Comment Pools |
| 3:35 – 3:50 PM | 62 | Machine Learning and Deep Learning Benchmarking for Multi-Horizon Tide Level Forecasting at Dili Port, Timor-Leste |
| 3:50 – 4:05 PM | 123 | Overcoming Face Mask Occlusion for 3D Face Reconstruction Using Deep Learning |
| 4:05 – 4:20 PM | 63 | Grounded, Multilingual, and Multimodal Avatars for Immersive Historical Education: Evolving a Generative-AI Platform and Its Unity3D Integration |
| 4:20 – 4:35 PM | 112 | Detection of Potential Depressive Disorders on “X” Social Media Based on User-Level Analysis Using Multi-Modal Transformers Architecture |
| 4:35 – 4:50 PM | 12 | Comparative Analysis of Machine Learning Models for Indoor Carbon Monoxide Prediction in Assisted-Living Environments |
| Day 2 · Wednesday, 21 October 2026 | ||
| 10:00 – 10:15 AM | 28 | Forest Fire Risk Prediction Based on Historical Data Using Enhanced Feedforward Neural Network |
| 10:15 – 10:30 AM | 146 | RISC-V Instruction Set Extension for Improving Keccak-256 and Poseidon Hash Algorithm Execution in Blockchain Applications |
| 10:30 – 10:45 AM | 53 | A Non-Binary Genetic Algorithm with Rule-Based Fitness Evaluation for Generating Monophonic Pentatonic Melodies in Festive Chinese Folk Music |
| 10:45 – 11:00 AM | 10 | Variable-Dependent Performance of Statistical and Machine Learning Imputation Methods for Hydrological Time Series |
| 11:00 – 11:15 AM | 72 | From Simulation Software to Smartphone: AI-Generated Interactive Tools for OEE-Based Decision Support in Semiconductor Manufacturing |
| 11:15 – 11:30 AM | 89 | EA-Prompt: Explainable Rule-Induced Prompting for Zero-Shot Open-Set Plant Disease Recognition |
| 11:30 – 11:45 AM | 44 | An Integrated Preprocessing Framework for Hydrological Flood Frequency Analysis Using Stationary GEV |
| 11:45 AM – 12:00 PM | 170 | Battery-Aware Reference-Rate Hungarian Subchannel Assignment for OFDMA-SWIPT IoT Networks |
| 10:00 – 10:15 AM | 23 | From Detection to Prevention: A Comparative Study of Object Detection Techniques for Distracted Driving |
| 10:15 – 10:30 AM | 45 | Benchmarking and Developing Optimized Agentic AI Workflows for Enterprise Customer Service in Dual-Control Environments |
| 10:30 – 10:45 AM | 56 | CASS: Corpus-Adaptive Lexicon Induction for Measuring Investor-Forum Discourse in Taiwan |
| 10:45 – 11:00 AM | 48 | Implementation of Fundus Image Classification for Eye Disease Detection Using Deep Learning and Explainable Artificial Intelligence in an Android-Based Application |
| 11:00 – 11:15 AM | 17 | Artificial Intelligence and Corporate Financial Performance: The Moderating Role of Firm Size |
| 11:15 – 11:30 AM | 183 | Community Detection on Function Call Graphs with an Adapted DYNMOGA for Program Comprehension |
| 11:30 – 11:45 AM | 116 | Confidence-Aware and Context-Validated Tourism Itinerary Recommendation Across Multiple LLMs: A Case Study in East Java, Indonesia |
| 11:45 AM – 12:00 PM | 33 | Development of a Knowledge-Based System Using Knowledge Graph Representation for Lung Disease Diagnosis |
| 10:00 – 10:15 AM | 73 | Automated Classification of Knee Osteoarthritis using Feature Extraction and Lightweight Classifiers |
| 10:15 – 10:30 AM | 142 | An LLM-based Relation Extraction Pipeline for Automated Boxology Extended Annotation Model Construction from Scientific Publications |
| 10:30 – 10:45 AM | 162 | Process-Supervised Multi-Agent Retrieval Augmented Generation for Constraint Fulfillment in Tourist Itinerary Recommendation System: A Case Study of the Bandung Region |
| 10:45 – 11:00 AM | 74 | Development of a Web and Phone Camera-based Deep Learning System for Detection and Identification of Banana Plant Leaf Diseases |
| 11:00 – 11:15 AM | 93 | A Hybrid Retrieval-Augmented Generation Chatbot for Legal Knowledge Services on the Indonesian Criminal Code (KUHP) and Criminal Procedure Code (KUHAP) |
| 11:15 – 11:30 AM | 139 | Indonesian Imagined Speech Recognition from EEG Signals Using Conformer-Transducer |
| 11:30 – 11:45 AM | 83 | SleepWell: Designing and Evaluating a Low-Interaction Mobile Sleep-Tracking Interface with User-Centered Design |
| 11:45 AM – 12:00 PM | 43 | Design of an AI-Based Clinical Decision Support System for Drug Interaction Prediction Based on Indonesian-Language Patient Complaints |
| 1:15 – 1:30 PM | 26 | SOS FL: Secure One Shot Federated Learning using BBS+ signature and Threshold Traceability |
| 1:30 – 1:45 PM | 29 | Generative Artificial Intelligence Adoption in Human Resource Management: Organizational Drivers in Malaysian Workplaces |
| 1:45 – 2:00 PM | 37 | Frequentist Model Averaging for Autoregressive Distributed Lag Models under Specification Uncertainty |
| 2:00 – 2:15 PM | 113 | SignalAgent: A Confidence-Guided Dynamic Agentic Signal Graph for Autonomous Industrial IoT Signal Analysis |
| 2:15 – 2:30 PM | 94 | Predicting Extraversion from EEG Signals using Higuchi Fractal Dimension Features and Optimized Boosting Models |
| 2:30 – 2:45 PM | 21 | Graphical Password Authentication Systems: A Review of Capture Attacks, Experimental Methods, and Countermeasures |
| 2:45 – 3:00 PM | 131 | One Size Fits All? A Focused Diagnostic Audit of Audience Evidence and Design Traceability in Gamified Cybersecurity Awareness Interventions |
| 1:15 – 1:30 PM | 167 | Speaker-Independent Speech Disorder Classification Using an Attention-Conditioned Dynamic GRU |
| 1:30 – 1:45 PM | 136 | Dynamic Memory and Compute Partitioning for SLO-Aware Colocated LLM Inference on GPUs |
| 1:45 – 2:00 PM | 104 | Forecasting Profitability of Indonesian Energy and Basic Material Sector Issuers through Deep Learning-Based ESG Disclosure Scoring of Sustainability Reports |
| 2:00 – 2:15 PM | 47 | VISTA: Efficient Cross-Lingual Visual Integration via Small Transfer Adapter |
| 2:15 – 2:30 PM | 54 | Design of an Artificial Intelligence Agent for Indonesian Criminal Law Research based on Retrieval-Augmented Generation and Knowledge Graph |
| 2:30 – 2:45 PM | 156 | Multi-Label Classification of Depressive Emotions in Indonesian Social Media Using XLM-RoBERTa-based Classifier Chains |
| 2:45 – 3:00 PM | 67 | An LLM-Based Structured Guided Reflection Chatbot Using the 5R Framework and FSM-Based Dialogue Control |
| 1:15 – 1:30 PM | 154 | Bridging Sensor Data and Clinical Language: LLM-Generated Explainable Alerts for IoT Cyber-Physical Health Monitoring Systems |
| 1:30 – 1:45 PM | 134 | Equilibria: A Collaborative Adaptive Assessment Framework for Addressing Academic Overpersonalization |
| 1:45 – 2:00 PM | 159 | Adaptive Agentic Retrieval-Augmented Generation for Question Answering in Higher Education |
| 2:00 – 2:15 PM | 133 | AURA: An Explainable Multi-Agent LLM-Based Geospatial Decision Support System for Land Feasibility Assessment |
| 2:15 – 2:30 PM | 79 | NirSisa: Development of an AI-Based Adaptive Recipe Recommendation System for Mitigating Household Food Waste |
| 2:30 – 2:45 PM | 85 | Question Class Prediction in Gastrointestinal Visual Question Answering Using Multimodal Feature Fusion |
| 2:45 – 3:00 PM | 114 | VibeTesting: An Agentic AI Framework for Automated Test Case Generation |
| 3:20 – 3:35 PM | 173 | GMIAS: Quantifying Security Overhead in Multi-Agent Intelligence Workflows |
| 3:35 – 3:50 PM | 75 | Transfer Learning-Based Convolutional Neural Networks (CNNs) for Pneumonia Classification with Enhanced Preprocessing and Sampling Strategies |
| 3:50 – 4:05 PM | 138 | A Systematic Literature Review on Customer Analytics and Personalized Banking using Big Data and AI |
| 4:05 – 4:20 PM | 122 | Dataset Quality Evaluation for ML Surrogate Modeling: A Comparative Study of Sampling Strategies for Compact Antennas |
| 4:20 – 4:35 PM | 88 | Cross-LLM Generalization in AI-Generated Phishing Detection: A Multi-Classifier Benchmark |
| 3:20 – 3:35 PM | 110 | Evaluating Cross-Lingual Transfer of Pretrained Speech Separation Models to Bengali Using the BanglaMix-v1 Dataset |
| 3:35 – 3:50 PM | 46 | When Does Contrastive Distillation Help? Contrastive Learning Strategy and Student Architecture in Dialogue Summarization |
| 3:50 – 4:05 PM | 105 | Retrieval-Augmented Generation for Reducing Hallucination in an Indonesian History Question Answering System |
| 4:05 – 4:20 PM | 78 | Implementation of Hijaiyah Reading Detection Using a Contrastive Learning-Based Pre-Trained Speech Recognition Model in the Tilawati Learning Application |
| 4:20 – 4:35 PM | 59 | Enhanced Genetic Algorithm with Single Parent Mating for Hair Dryer System Identification |
| 4:35 – 4:50 PM | 169 | User-Centered Interaction Design for LLM-Assisted Gamified Public Speaking Practice |
All times are Malaysia Time (GMT+8). Rooms are in Block A, Heritage Hall, Universiti Tunku Abdul Rahman (UTAR), Kampar. Refreshments and lunch are served in A009.