Dr. Wisnu Ananta Kusuma Shares Potential Utilization of Machine Learning in Medicine

Machine learning is part of artificial intelligence (Artificial Intelligence). This technologically advanced method is related to the design and development of algorithms that allow computers to learn from empirical data.

The result is a model that can generalize. This method has the potential to be applied in the medical field, especially to help doctors diagnose diseases.

Dr. Wisnu Ananta Kusuma, IPB University Lecturer from the Department of Computer Science, Faculty of Mathematics and Natural Sciences (FMIPA) said that to build an intelligent system, what must be done is the acquisition of data or knowledge.

He explained this in a Guest Lecture at the Department of Lung, Faculty of Medicine, University of Indonesia, last week.

After being acquired, he explained, it is necessary to separate the training data and test data. He described that learning in machine learning is like teaching to students.

“Students are given examples of questions during lectures (training phase) so that they gain an understanding that enables students to answer questions during exams (testing phase). The algorithm used plays a role in finding patterns and structures of knowledge and building generalizations," he explained.

According to him, one of the machine learning methods that are suitable for the characteristics of problems in the medical world is the Decision Tree. The decision tree generated by this method can be transformed into rules that are used to assist doctors in making diagnoses.

“The rules resulting from this Decision Tree are not final. But instead, it becomes the basis for helping doctors or experts to carry out assessments and refinements according to the experience and knowledge of doctors to produce rules for the appropriate diagnosis," he said.

The researcher from the Tropical Biopharmaca Research Center (TropBRC), Institute for Research and Community Service (LPPM) IPB University revealed that machine learning methods have been applied to screening herbal compounds for COVID-19.

“This research is one of the efforts to build a scientific basis for herbal medicine or herbal medicine. Usually, herbal medicines are consumed starting with local people's beliefs. In this approach, exploration is carried out by first selecting the target protein related to COVID-19," he added.

Furthermore, he said, a model of the interaction relationship between the target protein and conventional drugs that had been used for the treatment of COVID-19 was built. This model is used to test herbal compounds that have the same characteristics as conventional drugs.

"Thus, the effects and functions of these predicted herbal compounds have the potential to have the same characteristics," he said. He added that other research related to machine learning includes the determination of biomarkers and precision drug candidates with a multi-omics approach on cancer and the prediction of adverse drug reactions to estimate the side effects of drugs. (MW/Zul) (IAAS/YHY)

Published Date : 25-Nov-2021

Resource Person : Dr Wisnu Ananta Kusuma

Keyword : IPB University Lecturer, Tropical Biopharmaca Research Center (TropBRC) IPB, machine learning, Covid-19, medicine