Singapore – There may be no stupid question in schools, but there are questions with obvious answers. These questions, students, make assistant professor Leonard to shake his head.
Each day, the educators of the Nanyang Technological University (NTU) as the teacher receive a series of students of students who seek comments on homework or ask questions about future tests.
“We get questions like:” When is the exam? ” Or “should we arrive next week?” The answers are generally available online or in emails, but sometimes students simply do not check, “he said.
These information requests tend to set up just before a test – often late at night, because students surround it during the last -minute revision sessions.
This is a problem for artificial intelligence (AI), thought the professor, which prompted him as well as a team of teachers from the NTU’s School of Materials Science and Engineering (MSE) to develop Leodar, a chatbot d ‘IA which serves as a teaching and administration assistant. Leodar is among the first personalized generative chatbots traveling from NTU students.
Since his Launch in January, Student emails have decreased to one handle per week for each tutor, from several A day. Most requests are now processed directly by Leodar, who answered more than 12,000 questions of more than 150 users in 2024.
This released the teaching staff to manage more complex queries, the professor told Times Straits.
Ntu is inspired by Leodar’s success and will apply his approach to AI has future chatbots used In the rest of the university, through the NTU AI Learning Assistants Project. The project standard means the development of the chatbot in the faculties to minimize duplication.
Leodar is managed by a team of tutors, who update the AI model database with the latest course information, course notes and the study program. The system is built in collaboration with Amazon Web Services, based on Claude, a large language model built by AI Start-up Anthropic.
The school takes notes from the team approach to build chatbots. This includes a coding technique called increased generation of recovery, which limits AI’s responses to a predefined knowledge base, reducing the probability that AI constitutes its own facts. Teacherwho teaches data sciences and AI, Functional members must be familiar with programming and know how to update the assistant’s knowledge database so that a chatbot like Leodar is reliably.
The three tutors managing Leodar are able to see students’ questions and can answer them individually if they believe that AI has not provided good answers.
The MSE student, you Yan Yan, 21, said Leodar has been the cat of choice for many students in the past year, because his answers are specifically aligned with student learning equipment, y including coding and mathematics lessons.
“Other models of AI such as Chatgpt will give information from other sources, unlike Leodar, which uses the materials taught in class,” she said.
As the best practice, the first cycle MSE WANG TZU Wei checks the answers, in particular for assignments, by checking the responses or by executing the code generated by the AI to identify errors.
In addition Helping undergraduate students in their studies, the AI assistant allows users access to the school website and to open several files to find the details they need, said Wang , 21 years old.
“Now, I just type a question and get the answers I want,” he said, adding that the service allows students to ask last minute questions, because the tutors take the time to answer.
AI assistants are also introduced into other educational establishments. National University of Singapore Launched a series of AI tools in July called A-Know, which can recover information on courses, research documents and school policies for staff.
He plans to launch A-Know for students in 2025 for university issues and as study help.
The Institute of Technical Education (ITE) tests an AI assistant who recommends subjects for courses based on the latest employment and market trends, and plans to extend its use to more teaching staff ‘Here the beginning of 2025, said ite senior Director of the Eric Cheung study and educational development program.
The assistant aims to accelerate criticism of course programs To ensure that the learning equipment is relevant, he said. For example, he analyzed online sources on the aging population of Singapore and recommended to incorporate geriatric care skills into the program of studies For health care courses and suggested specific points in the teaching agenda for these subjects.
Dr Cheung said technology is also tested in a quiz generator for studentsPulling data from existing course equipment To create self-assessment modules For them.
- Osmond Chia is a technological journalist with Times Straits, covering cybersecurity and artificial intelligence subjects to the last consumer gadgets.
Join Channel Télégramme de St And get the latest news provided to you.