AI Innovation

Driving Malaysia's leadership in AI research and innovation initiatives.

i-SISDA

The project presents MAICoE’s award-winning initiative, i-SISDA (Intelligent Surveillance and Intervention on Supply and Demand of Agrofood), which leverages advanced AI forecasting models to enhance Malaysia’s agri-food supply chain management. The project integrates micro and macroeconomic indicators to predict agrofood supply and demand accurately. It employs model optimization techniques to minimize forecasting error, specifically, reducing the Mean Squared Error (MSE), ensuring reliability in national food planning. Additionally, the system includes scenario analysis capabilities, empowering stakeholders to test various policy interventions and their impacts. Recognized with international accolades from IFIA and MIIA at the 9th iCAN Awards 2024, this smart agriculture solution reinforces MAICoE’s expertise in data-driven governance. With collaboration from Universiti Malaysia Sabah (UMS) and FAMA, the project exemplifies MAICoE’s commitment to impactful, AI-driven public sector innovation aligned with Malaysia’s food security agenda.

Smart Healthcare

The Smart Healthcare project also highlights MAICoE’s strategic initiative in developing an AI-powered personalized healthcare delivery platform through a robust academia-industry collaboration. This project integrates data from hospitals, laboratories, researchers, and doctors into a centralized e-health web portal, enabling seamless coordination of patient care. Leveraging AI tools such as facial recognition, medical imaging diagnostics, handwritten transcription, and disease outbreak prediction, the system enhances diagnostic accuracy and healthcare efficiency. Conducted in partnership with institutions like UMS, MQ, UTP, UNIMAS, and local hospitals such as Hospital Tuaran, the initiative showcases how AI can be operationalized to deliver personalized and data-driven healthcare solutions. This collaborative ecosystem also reflects MAICoE’s leadership in aligning innovation with national healthcare goals, optimizing resource use, and improving patient outcomes across Malaysia’s medical infrastructure.

AquaSense

AquaSense highlights MAICoE’s core strengths in delivering practical AI solutions through the integration of Artificial Intelligence, IoT, and Edge Computing. It showcases how MAICoE implements various machine learning architectures, from local servers to enterprise-level systems, to support real-time sensing and decision-making. A key application is in the Blue Economy, where MAICoE has developed smart aquaculture shipping tanks equipped with pH, temperature, and dissolved oxygen sensors, alongside AI-based monitoring for predictive control. This reflects their expertise in sensor integration and edge AI for sustainable aquaculture operations. The project also demonstrates strong multi-stakeholder collaboration, involving Universiti Malaysia Sabah (UMS), and NCT Logistics, reinforcing MAICoE’s strategic role in the Triple Helix Model by uniting academia, industry, and government for impactful deployment.

Smart Agriculture

The Smart Agriculture project led by MAICoE, which won the Canadian Gold & Special Awards at the 6th Annual International Invention Innovation Competition (iCAN 2021) in Toronto, exemplifies Malaysia’s growing leadership in applying AI to enhance food security and agricultural sustainability. This initiative focused on developing a deep learning-based cyber-physical system to monitor and improve paddy production. It was driven by two critical challenges: the lack of modern technological solutions for assessing paddy health, and significant crop losses, estimated between 37% to 50%, due to pests and diseases. The system integrates AI-powered aerial imaging via drones and sensor technologies to detect anomalies such as disease symptoms and pest invasions at an early stage. Collaborating with partners such as the Federal Agricultural Marketing Authority (FAMA) and academic experts from Universiti Malaysia Sabah (UMS), this project illustrates MAICoE’s ability to deliver vertical AI solutions that meet real-world challenges. It also underscores MAICoE’s alignment with national agendas like MyDIGITAL and RMK13, by promoting smart farming technologies that enhance productivity, reduce losses, and strengthen Malaysia’s digital and green economy goals.

Smart Cybersecurity

RTFaceDeep is an award-winning cybersecurity innovation developed by MAICoE that leverages deep learning architecture to perform real-time multi-face recognition for video surveillance systems. This advanced system is designed to enhance public safety and smart city infrastructure by accurately identifying multiple individuals simultaneously across video streams. The system integrates AI with computer vision to support law enforcement and security agencies in proactive threat detection, identity verification, and real-time monitoring. Its strength lies in high-speed processing, adaptability to varying lighting conditions, and robustness against occlusions or dynamic backgrounds. The project was internationally recognized with the Gold Medal at the iENA Germany International Trade Fair 2019, under the “New Inventions & Products” category. This global achievement reinforces MAICoE’s excellence in AI-driven cybersecurity solutions and highlights Malaysia’s innovation on the world stage. A collaboration between Universiti Malaysia Sabah (UMS) and CyberSecurity Malaysia, the project exemplifies the power of the Triple Helix Model to deliver secure, intelligent, and scalable AI solutions for national and global impact.

Smart Farming

The Smart Farming: A Novel Two-Tier Convolutional Neural Network to Predict the Health of Beef Cattle for Sustainable Beef Production is an internationally recognized AI-powered initiative led by Professor Ts. Dr. Rayner Alfred and his research team from Universiti Malaysia Sabah (UMS), in collaboration with Hafiz Farm Sdn Bhd. Awarded the Canadian Gold Medal and Inventors Circle Special Award at the 7th Annual International Invention Innovation Competition (iCAN 2022) in Toronto, this project introduces an innovative two-tier CNN model optimized to non-invasively assess and predict the health of beef cattle. Unlike conventional methods that are costly and time-consuming, the proposed ensemble-based AI approach offers a practical, fast, and efficient alternative for monitoring cattle health, ultimately supporting more sustainable and profitable beef production in Malaysia. This solution exemplifies the fusion of deep learning with real-world agricultural challenges to empower smart livestock management.

FusionCare: Smart Patient Behavior Analytics through UWB & IMU Integration

The FusionCare project showcases MAICoE’s commitment to real-world AI innovation in healthcare, combining Ultra-Wideband (UWB) technology and Inertial Measurement Units (IMU) to create a precise, real-time patient tracking and behavior monitoring system. This system enables highly accurate indoor localization and motion analytics—vital for fall detection, rehabilitation tracking, and behavioral pattern analysis in clinical or assisted-living environments. Using a network of UWB anchors and IMU sensors, the system fuses multi-modal data through advanced AI algorithms to determine 3D positioning, gait phase, and activity status. The project also addresses critical concerns such as cybersecurity, data ethics, and the necessity of continuous patient monitoring, illustrated by its holistic design approach. Developed in collaboration with MIMOS and MQ, FusionCare underscores MAICoE’s technical leadership and collaborative strength within the Triple Helix framework. It further reinforces Malaysia’s national digital health ambitions aligned with the National AI Roadmap and MyDIGITAL blueprint.

An Integrated RNN-CNN Approach to Classifying Time-Series Forests

The project represents a groundbreaking AI-driven innovation aimed at enhancing environmental sustainability. Spearheaded by A/Prof. Ts. Dr. Rayner Alfred and his team at Universiti Malaysia Sabah (UMS), this initiative earned the Silver & Tummiad Special Awards 2022 at the 7th Annual International Invention Innovation Competition in Canada (iCAN 2022). The core of the project lies in an optimized conjoined neural network that combines recurrent and convolutional architectures to accurately classify time-series hyper-spectral forest imagery. This classification enables precise estimation of carbon stock in forested areas, aiding in biodiversity conservation and carbon offset initiatives. The model has been meticulously fine-tuned to enhance its classification accuracy, offering a powerful tool for automated, data-driven environmental monitoring and forest management.