> News > Fasilkom-TI Successfully Held an International Symposium on Artificial Intelligence and the Internet of Things in Perspective from Indonesia and Malaysia
Fasilkom-TI Successfully Held an International Symposium on Artificial Intelligence and the Internet of Things in Perspective from Indonesia and Malaysia
Published At
07 May 2024
Published By
Anonymous Writer
Thumbnail Fasilkom-TI Successfully Held an International Symposium on Artificial Intelligence and the Internet of Things in Perspective from Indonesia and Malaysia
The Faculty of Computer Science and Information Technology (Fasilkom-TI) of the Universitas Sumatera Utara (USU) has successfully held an International Symposium on Artificial Intelligence and Internet of Things in Perspective from Indonesia and Malaysia on Tuesday, May 7, 2023, at the Digital Learning Center Building, 8th Floor, Universitas Sumatera Utara.
This activity was opened with remarks from the Vice Rector IV of USU Prof. Dr. Opim Salim Sitompul, M.Sc., remarks from the Dean of Fasilkom-TI USU Dr. Maya Silvi Lydia B.Sc., M.Sc., and remarks from the Dean of Fasilkom-TI Universiti Putra Malaysia (UPM) Dato' Prof. Dr. Shamala K. Subramaniam.
The activity continued with signing the agreement and giving souvenirs, group photos, and coffee breaks. Then entered the first keynote speech session presented by Assoc. Prof. Dr. Fatimah Khalid (UPM) moderated by Amer Sharif, S.Si, M.Kom. In her presentation, she explained her “Revolutionizing Livestock Management for Food Security” research. He described the importance of livestock management in the context of global food safety and how revolutionary technological developments have led to fundamental changes in livestock management practices. In addition, it explains the risks associated with traditional livestock management and how it affects food safety. Advanced solutions are needed to address challenges in traditional livestock management, including Artificial Intelligence (AI) and Precision Livestock Breeding. With AI, we can revolutionize animal husbandry by enabling predictive modeling, disease detection, and precision feeding. This research uses the application of AI and Smart Herd Management. This research has three projects: buffalo identification through snout image, surti goat weight estimation, and goat face recognition.
The second keynote speech was delivered by Prof. Dr. Imas Sukaesih Sitanggang S.Si, M.Kom., from Computer Science Bogor Agricultural University (IPB) and moderated by Dr. Mohammad Andri Budiman, S.T., M.Comp.Sc., M.E.M. He explained his research entitled “Patrol System for Forest and Land Fire Prevention in Indonesia”. He created the Forest and Land Fire Prevention Patrol Information System (SIPP Karhutla) - SMART Patrol Information System in his research. This system is web-based and mobile. The user for the mobile application is the fire and land prevention patrol team. The mobile application is easy to use, manages 88 patrol parameters collected from the field, automatically retrieves location and weather data, works with or without internet connection, real-time hotspot data from Sipongi+, and a dictionary of patrol terms. The web-based system is used by administrators, regional operational heads, heads of MoEF's climate change and forest and land fire control, regional coordinators, and administrators of the directorate of forest and land fire control.
The third keynote speech was delivered by Assoc. Prof. Dr. Amalia S.T., M.T., and Dr. T Henny Harumi moderated. In this keynote speech, she presented her research entitled “Reducing Assessment Bias and Inconsistency with BERT-Based Automatic Short Answer Assessment for the Indonesian Language”. In her presentation, she explained that essay assessment encourages deeper learning, critical thinking, and the development of important skills that are valuable outside the classroom. However, manual essay assessment poses challenges due to diverse student responses, weaknesses of subjectivity, inconsistency, efficiency, and scalability. In addition, manual scoring is also time-consuming and prone to human error, especially when multiple raters are involved. Therefore, short answer scoring using BERT and the SQuAD dataset was created. BERT is a language model with Transformer architecture focusing on Attention Mechanism. As a result, the model achieved a scoring accuracy of 91%. This research highlights the efficacy of BERT-based systems in improving equivalence and precision.