Machine Learning for Practical Problem Solving in Agrifood: An Introduction via Online Micro-Course

Dr. Samsuzana Abd Aziz, Associate Professor; Dr. Hazreen Haizi Harith, Lecturer; Dr. Norulhuda Mohamed Ramli, PhD student; Ts. Dr. Siti Khairunniza Bejo, Professor; Dr. Nurulhuda Khairudin, Tutor; Dr. Diyana Jamaludin, Lecturer; Dr. Anas Mohd Mustafah, Lecturer at Universiti Putra Malaysia

Efficient, sustainable and Innovative Agricultural Practices

The intersection of agriculture and technology has never been more promising. With the advent of artificial intelligence (AI) and machine learning (ML), the agrifood sector stands on the brink of a significant transformation. This revolution promises enhanced efficiency, sustainability, and innovation in agricultural practices. ML is a subset of AI that focuses on building systems that can learn from and make decisions based on data. In recent years, ML has become a transformative force across various industries, including agriculture and food production. In the agrifood sector, ML applications range from crop monitoring and yield prediction to pest detection and resource management. These technologies enable farmers to optimize their practices, resulting in increased efficiency, reduced costs, and enhanced sustainability. Traditionally, learning ML involves a deep dive into advanced statistics, mathematics and computer science. This approach, while thorough, can be daunting and inaccessible to many, particularly those without a strong technical background. The good news is that with the advent of user-friendly tools and platforms, it is now possible to practice ML without an extensive background in advanced statistics and math. Online micro-courses and practical workshops have made ML more accessible, allowing a broader audience to harness its power for problem-solving in the agrifood sector.

ML for Problem-Solving and Micro-Course

To use ML effectively, one must understand the basic principles of data handling, model selection, and algorithm application. Additionally, practical problem-solving skills and the ability to interpret ML results are crucial. Micro-courses provide a focused, bite-sized learning experience that is ideal for busy professionals and students. They offer flexibility and allow participants to learn at their own pace. This format is particularly effective for introducing complex subjects like ML in an accessible and manageable way.

“Online micro-courses and practical workshops have made ML more accessible, allowing a broader audience to harness its power for problem-solving in the Agrifood sector.”

Circuit and Systems Society Micro-learning Environment (CASSMiLe) hosted by the International Society of Electrical and Electronics Engineers (IEEE) is an innovative online micro-courses platform that exemplifies this shift towards accessible and efficient education (Figure 1). Designed to simplify learning, CASSMiLe makes education easy and accessible for everyone, providing high-quality content in a concise and engaging manner.

Key Features of CASSMiLe:

1.  Micro-Learning Modules:

  ●  Short, focused lessons that cover specific topics, making it easier for learners to grasp complex concepts in a short amount of time.

  ●  Modules are designed to be completed in minutes, fitting seamlessly into busy schedules.

2.  Expert-Lead Course

  ●  Expertly curated content from leading educators and professionals in the field.

  ●  Regular updates to keep the material current and relevant.

3.  Personalized Learning Paths

  ●  Learners can choose their own path and pace, allowing them to focus on areas of interest or need.

  ●  Track progress and set personal learning goals to stay motivated and on target.

4.  Comprehensive Learning

●  A supportive learning community where users can connect, share ideas, and collaborate on projects.

●  Access to mentors and experts for guidance and personalized feedback.

●  Engaging multimedia elements such as videos, quizzes, and interactive simulations to enhance understanding and retention.

●  Opportunities for learners to apply their knowledge through practical exercises and real-world examples.

One real-world application of utilizing this platform is the recent IPB CSAgri summer course held in conjunction with the Seasonal School on Artificial Intelligence (AI), Integrated Circuits and Systems for Medical and Surgery Technologies (SSAICAS_MST 2023), and the IEEE CASMAKER Outreach Program coordinated by Universiti Putra Malaysia and IPB University.

The intensive summer course, held from 8^th to 13^th October 2023 at the IPB University campus in Bogor, brought together 38 participants from Malaysia, Indonesia, and Thailand. They delved into the fascinating intersection of sustainable agriculture, medical and surgical technologies, AI, and optimization techniques. This course aimed to enhance good practices and maximize efficiency by leveraging innovative technologies to address the challenges faced by the agricultural and biomedical industries.

The course commenced with keynote sessions delivered to around 100 students by prominent figures such as Prof. Imas Sukaesih Sitanggang, Dr. Yeni Herdiyeni, and Ts. Dr. Luthffi Idzhar Ismail. These sessions set the stage for a deep dive into AI principles and their applications in agrifood and medical surgery technologies. Following the keynotes, participants engaged in a technical excursion to the university’s lab and the Agribusiness and Technology Park (ATP) at IPB University.

A significant highlight of the course was the hands-on exercises facilitated through the CASSMiLe platform (Figure 2). Participants learned to apply AI algorithms and optimization models to solve real-world problems. A team of seven instructors who provided practical insights and guidance led these exercises. The micro-course format allowed participants to grasp the fundamental principles of AI in an accessible and structured manner.

After the on-site classes, participants were grouped into teams to work on a virtual project. They were given a week to discuss and develop their innovations in agricultural and biomedical fields using AI. These projects were virtually supervised by 16 lecturers from IPB University and Universiti Putra Malaysia, culminating in the presentation of nine innovative projects. These projects showcased practical solutions to problems in agriculture and biomedical fields, demonstrating the participants' newfound skills and knowledge.

The IPB CSAgri summer course exemplifies how online micro-courses can effectively introduce and teach complex subjects like AI and ML in the context of agrifood. By focusing on practical problem-solving, these courses can equip participants with the tools and knowledge necessary to drive innovation in their respective fields. Feedback from 22 of 38 participants was overwhelmingly positive (Figure 3). They appreciated the practical focus and the opportunity to apply ML to real-world problems. The virtual project component, where teams developed AI-driven innovations in agriculture and biomedical fields, was particularly well-received. The success of this summer course underscores the potential of micro-courses to make advanced technologies more accessible and applicable to real-world challenges.

As the agrifood sector continues to evolve, initiatives like the IPB CSAgri summer course will play a crucial role in bridging the gap between traditional agricultural practices and modern technological advancements. By fostering a new generation of tech-savvy agricultural professionals, we can look forward to a future where AI and ML drive sustainable and efficient agricultural practices worldwide.professionals, we can look forward to a future where AI and ML drive sustainable and efficient agricultural practices worldwide.

Figure 1: The micro-courses on the CASSMiLe platform. Two series were prepared, namely “Machine Learning in Agrifood – Series 1” covering the fundamentals of model development, and “Machine Learning in Agrifood – Series 2” covering the practical aspect of model development using agrifood-related examples.

Figure 2: Intensive workshops were conducted to 38 participants utilizing micro-course in Machine Learning for Agrifood via CASS Microlearning (CASSMiLe) platform.

Figure 3: Participants feedback on the micro-course (n = 22). 5-point Likert scale was adopted to capture participants’ feedback, 1: strongly disagree, 3: neutral and 5: strongly agree. Data presented as average and standard deviation.

Machine Learning

Artificial Intelligence