“I didn’t realize that AI is already entrenched in Dentistry!” remarked a student midway through our workshop on Machine Learning (ML). The workshop is part of the new Data Science and Analytics (DSA) course in IMU‘s Bachelor of Dental Surgery (BDS) programme which was specially designed for dental students.
The DSA is a new interactive course specially designed for the revised BDS 2021 curriculum in IMU. The course is primarily delivered through hands-on active learning workshops by DSA experts in the field of dentistry. The primary goal of the course was to explore, identify and practice the enormous potential of AI in dental practice. The idea came from the stakeholders, while effective modeling by software/coding experts brings the ideas to fruition.
The human effort for effective machine learning is an important initial step in computers interpreting the available data for the benefit of humans. The open-source web based CVAT (app.cvat.ai) was used to label and annotate the orthopantomography (OPG) radiographic images. The students, who had done some prior learning through videos, labelled and annotated various anatomical structures in the OPGs and was amazed at how humans can teach machines to do mundane repetitive tasks to perfection.
The hands-on learning with peers as facilitators brought out a lot of interesting queries. “How do I mark the amount of bone loss, put a line or a block?” Maybe we need to annotate the tooth, CEJ, bone and other anatomical features; the algorithm which the coders write, can instruct the computer to measure the distance from CEJ to upper crest of bone at different tooth locations.
One discussion which fascinated us – how do humans and machines learn? Is it same or different? We found it was quite different. Human learning was very complex and least predictable. ML was simple, repetitive, and straight forward. “Will AI take our jobs?” Even the first Industrial Revolution (IR) had taken a lot of human jobs. AI is in the 4th IR.
The workshop was facilitated by few student champions and an IMU lecturer, Dr Shaju; who guided the students in groups on labelling and annotating. This helped the workshop to be interesting and engaging.
|Few Take-home Messages from the Workshop
|Group learning is engaging as peer support and learning happens effortlessly. (Do machines peer learn?)
|The more alien the topic, the less engaging it is. Prior learning materials must be provided to make the content familiar.
|Thinking out of box and creativity thinking needs practice and reinforcement.
|Boldness to make mistakes facilitates learning.