cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Andri Putra Kesmawan
Contact Email
journal@idpublishing.org
Phone
+6281990251989
Journal Mail Official
journal@idpublishing.org
Editorial Address
Perumahan Sidorejo, Jl. Sidorejo Gg. Sadewa No.D3, Sonopakis Kidul, Ngestiharjo, Kapanewon Kasihan, Kabupaten Bantul, Daerah Istimewa Yogyakarta 55182
Location
Kab. bantul,
Daerah istimewa yogyakarta
INDONESIA
Journal of Technology and System Information
Published by Indonesian Journal Publisher
ISSN : -     EISSN : 30322081     DOI : https://doi.org/10.47134/jtsi
Core Subject : Science,
Computer Science & IT
The Journal of Technology and System Information is dedicated to publishing cutting-edge research and advancements in the broad and dynamic intersection of technology and information systems. The focus of the journal is to facilitate the exchange of knowledge and ideas in these interconnected domains, fostering a deeper understanding of the role of technology in shaping information systems and vice versa. The journal welcomes contributions that span theoretical, empirical, and practical aspects, with an emphasis on the transformative impact of technology on information systems and vice versa. The scope of JTSI is a Information Technology and Systems, Data Management and Analytics, Emerging Technologies, System Design and Optimization, Cybersecurity and Privacy, Networks and Communication Systems, Artificial Intelligence and Machine Learning, Human-Computer Interaction.
Arjuna Subject : Ilmu Komputer (semua kategori) - Ilmu Komputer (Lain-Lain)
Articles 76 Documents
 4   5   6   7   8 
AdverShield-LLM: Adversarial Robustness Certification for IoT-Integrated Retrieval-Augmented Generation via Randomized Smoothing Yasser Samir Hadi
Journal of Technology and System Information Vol. 3 No. 2 (2026): April
Publisher : Indonesian Journal Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47134/jtsi.v3i2.6047

Abstract

The emergence of the Internet of Things (IoT) ecosystems, Retrieval-Augmented Generation (RAG) systems have become commonplace and provide a means for embedding dynamically retrieved external knowledge in the response from a Large Language Model (LLM). While potentially helpful, IoT-enabled RAG pipelines present significant adversarial threats such as poisoning passages into the IoT knowledge base, altering dense retrieval embeddings, and conducting indirect prompt injection attacks via the inputs through the IoT sensors, all of which can impact the fidelity of generated responses and compromise the trustworthiness of the system. Current defenses are based mostly on heuristic filtering or empirical adversarial training, and are not known to be robustly certified or are fragile under adaptive adversaries. In response to these challenges, this article introduces a new certified defense framework named AdverShield-LLM to combine the randomized smoothing technique with a multi-granular noise injection mechanism well-suited to the distributed and low latency requirements of RAG systems in the IoT domain. AdverShield-LLM consists of three synergistic modules: (i) Passage-Level Smoothed Aggregation (PLSA) module which certifies the robustness of RAG retrieval against bounded corpus poisoning under an isolate-then-smooth paradigm, (ii) Token-Adaptive Gaussian Defense (TAGD) layer that certifies LLM generation against indirect prompt injection by propagating l_2-norm perturbation bounds through the transformer attention stack, and (iii) IoT-Aware Certified Radius Scheduler (IACRS) that dynamically schedules noise budgets among constrained edge nodes while preserving the certified radius. AdverShield-LLM is evaluated on three IoT security benchmarks—MS-RAG-IoT, NQ-Adversarial and IoTQA-Poison—with extensive experiments showing its certified accuracy is 81.4% under l_2 perturbation radius σ=0.50 compared to the strongest baseline RobustRAG which reported +9.3% accuracy, and reduced the attack success rate from 74.2% to 8.6% against PoisonedRAG. Moreover, AdverShield-LLM ensures the accuracy of clean answers within 2.1% of the undefended RAG accuracy, proving that certified robustness does not compromise the utility of RAGs in resource-limited IoT environments.
Performance Improvement of f-OFDM Systems Using a Concatenated RS/LDPC Coding Scheme Ghasan Ali Hussain
Journal of Technology and System Information Vol. 3 No. 2 (2026): April
Publisher : Indonesian Journal Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47134/jtsi.v3i2.6091

Abstract

In wireless communication systems, transmission errors are inevitably generated due to interference and fading. Therefore, various approaches have been developed to enhance system reliability, including increasing the transmitted signal power and utilizing error detection and correction schemes. Among these approaches, channel coding is an important technique for reducing the Bit Error Rate (BER). Although LDPC codes are widely adopted in 5G wireless communication systems; achieving ultra-low BER values remains challenging in LDPC codes due to the error-floor phenomenon. To address this limitation, a modified concatenated RS/LDPC code is proposed for an f-OFDM system in this paper. It employs RS codes with LDPC codes followed by an interleaver. Unlike the conventional concatenated RS/LDPC codes that separated both codes by an interleaver. Simulation results demonstrate that the suggested system achieves superior BER performance compared with the conventional concatenated scheme and standalone RS and LDPC codes under both BPSK and QPSK. Furthermore, the suggested f-OFDM system provides superior Out-of-Band Emission (OOBE) suppression compared with the conventional OFDM system, while maintaining a Peak-to-Average Power Ratio (PAPR) performance comparable to that of the conventional OFDM system. Based on the results, the suggested f-OFDM system is considered a promising candidate for 5G and beyond systems.
SAW-ROC Feeder Route Prioritization with Thermal Criteria: Cimahi City Rasoki Daulay; Dhanny Karewur; Okyza Prabowo
Journal of Technology and System Information Vol. 3 No. 2 (2026): April
Publisher : Indonesian Journal Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47134/jtsi.v3i2.6335

Abstract

This study addresses the prioritization of angkutan kota (angkot) routes to be integrated as feeder services for the Bus Rapid Transit (BRT) system in Cimahi City, West Java, extending a previous GeoAI-based spatial model of urban surface temperature developed for the same city. Four registered feeder-corridor routes were evaluated using seven operational and environmental criteria: load factor, operation rate, service evenness, route overlap, schedule deviation, headway, and mean land surface temperature (LST) extracted through zonal statistics within a 150 m buffer of each route, based on a random-forest-predicted LST raster. Criteria weights were derived objectively using the Rank Order Centroid (ROC) method based on an importance ranking, then aggregated with Simple Additive Weighting (SAW) to produce a composite priority score. Results show that route 01.02.05 (Ps. Antri-Cibeber via Contong) ranked highest (SAW score 0.827), driven mainly by its superior load factor, while the route previously rated highest in the official performance report ranked second once the thermal criterion was included. The findings indicate that incorporating a spatial thermal criterion shifts prioritization outcomes relative to conventional operational-only assessment, offering transport authorities a more environmentally responsive basis for feeder network restructuring decisions.
Artificial Intelligence Adoption in Government Institutions: Ethical Governance, Public Trust, and Risk Management Ghazwan Hani Hussein; Faiza Mohamed; Abdelsalam Abuzreda
Journal of Technology and System Information Vol. 3 No. 3 (2026): July
Publisher : Indonesian Journal Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47134/jtsi.v3i3.6339

Abstract

The application of artificial intelligence (AI) technologies in the public sector has led to improved public services, enhanced administrative performance, and strengthened automated decision-making. However, the increasing reliance on AI systems has raised concerns regarding accountability, ethical compliance, privacy protection, transparency, human oversight, and risk management. This study, employing both conceptual and qualitative research methodologies, examines the governance factors and requirements for responsible AI implementation in the public sector. The research methodology includes a comparative analysis of international AI governance frameworks and regulations. The study identifies key dimensions influencing responsible AI implementation, such as accountability, human oversight, ethical governance, legal compliance, risk management, and transparency. The findings demonstrate that the adoption of responsible AI cannot be achieved through technological means alone but also requires a commitment to comprehensive governance mechanisms. Furthermore, the sequential interaction and interdependence of governance factors reduce operational and societal risks, increase transparency and explainability, and foster public trust in the systems. This study contributes to enriching the culture and knowledge of AI governance, and the proposed framework helps government sector leaders develop responsible AI governance in accordance with international standards and regulations.
A Time-Domain Power Flow and Genetic Algorithm Approach to Optimal PV and Battery Storage Siting for Grid Resilience Muhammed F. Alwaeli
Journal of Technology and System Information Vol. 3 No. 3 (2026): July
Publisher : Indonesian Journal Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47134/jtsi.v3i3.6341

Abstract

Higher ambient temperature can adversely affect conductor operating conditions and distribution-network losses, while comparatively less attention has been given to their mitigation through coordinated PV/BESS optimization. This study develops a climate-aware framework for the IEEE 33-bus distribution system using a temperature-dependent conductor resistance model under climate-stress conditions with 24-hour time-domain power flow and Genetic Algorithm optimization of PV/BESS sizing and placement. The reconstructed network reproduced the reported base-case loss of 202.7 kW. Under a severe climate-stress scenario of 50°C and 55% RH, daily losses increased from 2845.4 to 3158.0 kWh (+11.0%), while daily efficiency decreased from 95.76% to 95.30%. The optimized configuration, consisting of 1794.1 kW PV and a 448 kW/1760.4 kWh battery at bus 14, reduced daily losses to 2431.5 kWh, corresponding to reductions of 23.0% relative to the climate-stress case and 14.6% relative to the reference case. The minimum voltage during the critical operating period increased from 0.9063 to 0.9240 pu, and the number of hours within the specified voltage range increased by two. A Particle Swarm Optimization benchmark and sensitivity analysis further evaluated the robustness of the proposed configuration across the tested optimization methods and climate scenarios, with the optimal installation location varying under different capacity constraints.
Comparative Evaluation of ChatGPT, Gemini, and Claude Response Quality on Programming Questions Based on Bloom's Taxonomy Kahfi Bintang; Median Zikri; Rivaldo Nugraha; Dhanny Karewur
Journal of Technology and System Information Vol. 3 No. 3 (2026): July
Publisher : Indonesian Journal Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47134/jtsi.v3i3.6354

Abstract

The rapid advancement of Large Language Models (LLMs) has significantly influenced programming education and software development by providing automated code generation and problem-solving assistance. However, differences in the quality of responses generated by ChatGPT, Gemini, and Claude require a comprehensive evaluation using multidimensional assessment criteria. This study aims to compare the quality of responses produced by these three LLMs on programming questions based on Bloom's Revised Taxonomy and the ACM/IEEE Computing Curricula assessment rubric. A quantitative comparative approach was employed using 40 programming questions distributed across six Bloom cognitive levels (C1–C6). Each response was evaluated using five indicators: accuracy, completeness, clarity, logical reasoning, and efficiency. The collected data were analyzed using descriptive statistical methods and mean comparison. The results indicate that ChatGPT achieved the highest overall mean score of 21, followed by Claude (20) and Gemini (19). ChatGPT also obtained the highest scores in accuracy (4.3), clarity (4.1), logical reasoning (4.4), and efficiency (4.5), while ChatGPT and Claude shared the highest completeness score (4.1). Furthermore, performance across all models declined as the cognitive level increased from Remember (C1) to Create (C6), although ChatGPT demonstrated the most consistent performance across all Bloom levels. These findings indicate that ChatGPT provides superior and more consistent programming solutions and that integrating Bloom's Taxonomy with multidimensional evaluation criteria offers a comprehensive framework for assessing the quality of LLM-generated programming responses.

Page 8 of 8 | Total Record : 76

 4   5   6   7   8 

Filter by Year

2024 2026


Reset
Filter By Issues
All Issue Vol. 3 No. 3 (2026): July Vol. 3 No. 2 (2026): April Vol. 3 No. 1 (2026): January Vol. 2 No. 4 (2025): October Vol. 2 No. 3 (2025): July Vol. 2 No. 2 (2025): April Vol. 2 No. 1 (2025): January Vol. 1 No. 4 (2024): October Vol. 1 No. 3 (2024): July Vol. 1 No. 2 (2024): April Vol. 1 No. 1 (2024): January More Issue