Leveraging (Generative) AI in Science and Mathematics Education

(Upcoming) 8 Jan 2026 Hang Seng University of Hong Kong Center for Teaching and Learning( CTL) Workshop
(Upcoming) 1-3 Dec 2025 QS Reimagine Education Awards and Conference, Shortlisted in the "Best Use of Generative AI" Category for "YoChatGPT!: From AI Monologue to Collaborative Educational Dialogue", London, UK
(Upcoming) 14-16 Nov 2025 YoChatGPT! in Top 30 Shortlisted for the T4 EdTech Global Award at the World School Summit in Abu Dhabi, UAE - see https://t4.education/world-schools-summit/
(Upcoming) 7-8 Nov 2025 Workshop in The Future of Learning Conference, Hong Kong. Organized by Apple Distinguished Educators - see https://folhk.com/ for updates.
(Upcoming) 5-6 Nov 2025 Case study presentation: Supercharging classrooms with AI-enhanced collaborative learning with EdUHK - EdUTech Asia under the “Higher Ed - Digital Transformation” track, Singapore. See Nov. 5 at 17:10pm - see agenda here: https://www.terrapinn.com/exhibition/edutech-asia/agenda.stm. Also, Rountable at 13:55 on Nov. 6 on Collorative AI implementation in classrooms.
(Upcoming) 21 Oct 2025 International Webinar Series 2025 Humanizing Education in the Age of AI: Emotional Intelligence, Equity, and Innovation, Talent Quest and UNIVERSITAS SEBELAS MARET, Indonesia
(Upcoming) 18-26 October 2025 Hong Kong Science and Technology Parks InnoCarnival; YoChatGPT! at EdUHK's showcase booth highlighting EdTech and CultureTech.
(Upcoming) 17 Oct 2025 YoChatGPT! Collaborative AI talk at TED webinar at CityU.
(Upcoming) 16-17 Oct 2025 YoChatGPT! will have a booth at the Hong Kong Productivity Council: Fund & Tech Sourcing Fair 2025 - https://smereachout.hkpc.org/
(Upcoming) 15 Oct 2025, 11:50am-12:30pm "Exploring Collaborative Generative AI Pedagogies and their Impact on Student Learning", Symposium for Scholarship of Teaching and Learning, Banff, Alberta, Canada (Collaborative work with Speaker: Dr. Christopher Chlebovec)
12-14th Sept. 2025 "YoChatGPT!: Unleashing the Power of Collaborative AI" Digital Entertainment Leadership Forum (DELF) organised by Cyberport.
30 August 2025 YoChatGPT! entry into The 10th International Invention Innovation Competition in Canada, iCAN 2025 https://www.tisias.org/ican-2025.html
27 August 2025, 6pm-8pm, TEC@Tech Zoom Talk at the Teaching Excellence Center (TEC) at Indiana Tech University, USA
20-21 August 2025 "Integrating Generative AI into Inquiry-Based Learning for Personalized Mathematics Education", 2nd EdUHK x HKUST Joint International Conference on AI and Education (AIEdu2025)
14-15 August 2025 "From My AI to Our AI: Transforming Education with Collaborative Generative AI Pedagogies" Shortlisted for eLEARNING FORUM ASIA (eLFA) 2025 Awards in the "Exemplary Teaching and Learning" Category
8-10 August 2025 - YoChatGPT! entry into the Silicon Valley International Inventions Festival 2025 (SVIIF), USA, California, Santa Clara, Missio City Ballroom - Result: Won the Silver Award
7 July, 2025. Empowering Global Learning: Exploring the Role of Generative AI in Collaborative Online International Learning (COIL), Collaborative Online International Learning (COIL) GenAI workshop at EdUHK
4 July 2025 "Supercharging Classrooms with Collaborative AI: Unleashing YoChatGPT! for Next-Gen Learning", HK Learning and Teaching Expo (2-4 July 2025), https://ltexpo.org/en/, "Edtech Classroom" Theater; 15:00-15:40. Direct Link talk on LTE website: https://s.jemex.me/wFnYeYiSI
30 June 2025 Guest Lecture on YoChatGPT! in "Applied GenAI for Interdisciplinary Projects" course at HKUST.
13-14 June 2025, "Leveraging Generative AI for Personalized Inquiry-Based Mathematics Education: Impacts on Critical Thinking and Academic Performance", EdUHK-Tsinghua University Education Forum: Future Education and Learning
9-11 June 2025 "Increasing problem solving and critical thinking skills in mathematics students in the age of AI via prompt engineering", The Mathematical Cognition and Learning Society (MCLS) 2025 - accepted poster but decided not to register and present.
10 June 2025 "Powering PolyU's Strategic Plan : A Transformative Teaching and Learning Experience with Collaborative Generative AI Pedagogies", PolyU Educational Development Center Internal Training, 11am-1:30pm, QR404.
(Organizer) Thurs. 5 June 2025, "Global Perspectives on Collaborative Generative AI Pedagogies: Real-World Use Cases from K12 to University" - co-organized with AIREA EdUHK (speakers from the Philippines, China, Turkey and Canada)
May 30, '25 "Generative AI for Practical Applications for English Language Teaching and Learning", Teaching Brown Bag Series, EdUHK ELE Dept. talk. 12:30-1:30pm
21-23 May 2025 "Personalized Learning with Generative AI in Teacher Education via the Concept Prompt", International Conference on Learning and Teaching, EdUHK campus (joint work with Chan C.L. Lawrence, EdUHK)
May 16-18, '25, "AI facilitated learning of data analysis: Student engagement with YoChatGPT!" (collaborative work led by Jay San Pedro, iAcademy, Phillipines), at the 8th International Congress on Action Research, Action Learning, ARAL 2025.
May 8 ‘25 “Leveraging Collaborative Generative AI for Design and Prototyping in Innovation and Entrepreneurship Education” at the Harvard CES-EdUHK Symposium
23 Apr. 2025 "AI for Education Innovation" Guest Lecture invited by Dr. Lucas Kohnke (Ed.D. Programme Coordinator, GS, Graduate Studies, EdUHK), Zoom.
9-13 Apr. 2025 "YoChatGPT! Unleash the Power of Human-Team-AI Collaboration" at the “The 50th International Exhibition of Inventions Geneva” - won Silver Award at the 50th International Geneva Inventions in the "W" Category (Hardware, Software, Cybersecurity, Blockchain, IoT)
26 March 2025 "Teacher-AI-Student Dynamic" (Invited) Zoom Workshop; 2025 Teaching Commons Summer Institute, Lakehead University, Canada. See Website here: https://libcal.lakeheadu.ca/event/3843157?hs=a
19 March 2025 HKU talk in their AI Literacy for U project: “Unleashing Superminds: Boosting Creativity with Human-AI Collaboration in Context-Aware Chatrooms” - see https://lib-instruction-events.hku.hk/ai-literacy-for-u/unleashing-superminds/
17 March 2025 “Exploring Multidisciplinary Applications of Collaborative Generative AI Pedagogies in Education”, in Zoom workshop titled “Disciplinary Use of GenAI to Maximize Student Learning: Sharing Frontline Experiences” for UGC’s FITE/IICA project led by PolyU on “GenAI in learning, teaching and assessment”
11 March 2025 "YoChatGPT! Igniting Collaborative Learning, Teaching and Work with Generative AI", Emerging Tech Research Salon, Global Institute for Emerging Technologies, EdUHK.
Dec. '24 “YoChatGPT! Unleashing the Power of Collaborative AI learning”, EdUHK KTSO Booth at the Learning and Teaching Expo in Hong Kong, 11-13 December 2024
Dec. '24 "Implementing generative AI integrated gamified learning and peer instruction through YoChatGPT's context-aware chatrooms", Poster presentation at Teaching and Learning Innovation Expo, 11 to 12 December 2024, Yasumoto International Academic Park, Chinese University of Hong Kong
Dec. '24 “YoChatGPT!: A Multi-LLM integrated Chatroom for Collaborative learning and teaching with GenAI” Reimagine Education Awards and Conference, 9-11 December 2024, QE II Center, London, UK (YoChatGPT! among 8 shortlisted in the "AI in Education" category)
Dec. '24 “Implementing personalized and active learning using a chatroom integrated with generative AI”, e-Learning Forum Asia (ELFA), The Hong Kong Baptist University, 4-5 December 2024
Nov '24 Collaborative STEM Learning with Generative AI using YoChatGPT! State University of Malang, Indonesia. Zoom workshop, Tues. Nov. 26, 3pm
Oct '24 "Personalized and collaborative learning with generative AI", CUHK AI for Education Project seminar. (Oct 31, 2024)
Oct ' 24 "YoChatGPT! Next generation LMS+SRS" CEAR Open Day at HKSTP, 21 Oct. 2025 - see press release 1 here and 2 here
Aug. '24 "AI for Science and Math Education", Zoom Workshop, Alliance Primary School, Sheng Shui, HK.
Aug '24 "Unlocking Personalized and Active Learning Using a Collaborative Chatroom Integrated with Generative AI", EdUHK LTTC Hybrid Seminar, 28 Aug 2024, 3-4pm, E-P-12
Aug'24 "Caution: Construction Ahead! Guided use of generative AI in Education using an online collaborative chatroom", Gratia Christian College Staff Development Seminar for their Staff Orientaiton Day on Friday, 23 August, 2024, 11:05 to 12:00pm (invited by Prof. Horace HO)
Mar ‘24 “Utilizing AI for Teaching and Learning”, Capability Enhancement Training for Faculty Members and Staff, Cavite State University, Philippines (Zoom)
Feb ‘24 “Leveraging Prompt Engineering in Large Language Models with Active Learning Strategies to Boost Mathematics Outcomes”, CUHK Project on AI for Education
Aug. '23 HKTEA CoP in STEM workshop on Generative AI: Opportunities and Challenges in STEM Education - 31 August 2023 (Organized and presented)
Dec. '22HKTEA CoP in STEM workshop on "STEM Across the Curriculum in AI Education" - 16 Dec 2022 (Organized)

Introductory Remarks

The guiding principle for the work below, is that, the way academics, researchers and industry people discover methods on how to improve the performance of various task of large language models (or multi modal large language models) on science and math related (question answering related) tasks, is PARALLEL to how we oursevles learn and how we (as teachers) try to teach our STEM students, i.e., how to learn concepts and how to solve problems on their on their course assessments (such as assignments, quizzes, midterm tests and final exams). Simple examples for which academic researchers and professional industrial LLM researchers are implementing to improve "quantitative reasoning" in math and science question answering are "zero shot" or "chain of thought "(CoT) to answer a question.

Let me explain further. As learners and teachers, we have actually used "zero shot" and "CoT" methods in our own learning and teaching strategy when lecturing to our students on how to solve science and math questions, for centuries! For example, the general strategy for learning how to do a problem is to first understand how to do a problem first, step by step; and then applying it to solve other problems. In teaching, to teach our students how learn about a topic/idea/concept (e.g., limits in Calculus) and how to solve a math and science problem, is to first introduce the idea, give the definition, and give a "simple example" of how to apply the idea to solve a given problem. To build their understanding of the concept and idea, we show them how to solve more complicated problems, by showing them the step by step solutions. Students, then, will most likely mimic and understand how to solve similar problems, by following our step by step solutions.

Therefore, the aim of this work is to (1) find out and list all the researchers work (in academia or industry) on improving task performance on LLMs on quantitative reasoning, and (2) explore whether it is feasible to integrate or transform the prompt engineering technique (for improving task performance of LLMs on solving science and math problems) into an active teaching pedagogy in science and math education to help improve STEM students' learning outcomes.

Finally, I believe the grand goal here is to help students eventually learn how to use AI to solve real world problems (e.g., in science, engineering, etc....). Why: there is already evidence that AI helps us solve some of the most difficult scientific problems: Google Deepmind has shown us, using AI and deep learning, we maybe able to solve many scientific problems in the world, starting with AlphaGo to beat professional Go players; AlphaFold to fold proteins (now up to AlphaFold3), AlphaTensor to speed up a calculation at the heart of many different kinds of code; AlphaDev to make key algorithms used trillions of times a day run faster; and finally, FunSearch (function search) to solve an old mathematics problem. In fact, as a classically trained mathematician, I also believe that AI is set to revolutionize mathematics. Albeit, Google Deepmind's approach is more fundamental and foundational, students these days are already using LLMs now (whether in their homework or spare time (just curious about the technology, etc...); when students graduate and get out in the real world, they'll also encounter generative AI when trying to streamline their work and make themselves more efficient and productive (using tools like Microsoft Co-Pilot, Google AI Overviews, etc...); and when they are working in a scientific or engineering type enterprise, they'll inevitably try to use generative AI to help them (initially) getting started with some technical (quantitative) projects which require them to ask LLMs some technical questions. In any case, students should be aware and have access to AI (in this case, generative AI), which will help them get an intuitive feel of what AI is, how they work, what they are able to do (and not do), and hopefully, begin to use them to solve some hard problems.

A method called "Zero-shot" prompting technique has been found to improve LLM performance on these tasks. One effective strategy within this method is the "Let's think step by step" technique.

Instead of simply presenting the question, we follow our prompt with "A: Let's think step by step", before the question.

Large Language Models are Zero-Shot Reasoners https://arxiv.org/pdf/2205.11916.pdf

1.Zero Shot Prompting

Example of a "Zero Shot" prompt

Chain of Thought (CoT) is a technique we can use to improve the LLMs' performance on certain tasks. Instead of asking the LLM a question directly, we break down the problem into smaller parts, and guide the model through solving each part step by step. This method can help the model generate more accurate answers, and it allows us to understand the problem-solving process more clearly.

2. Chain of Thought Prompting

Normal prompt or original question (O-Q) ChatGPT:

Find the solution to.....

Chain of Thought prompt engineering template:

Q: Problem similar to original problem to be solved

A: Full step of solution

Q: Find the solution to...

Chain-of-Thought Prompting Elicits Reasoning in Large Language Models https://arxiv.org/pdf/2201.11903.pdf - "enables" reasoning capabilities of the LLM through intermediate steps of reasoning

3. Few Shot Prompting

Few shot Prompting is where you would string together two or more CoTs, before asking the original question. In fact, when companies (OpenAI, Google, Meta, Anthropic, etc...) evaluate their (open sourced or proprietary) models, they always use few-shot prompting to test their models against on bench marked test against competing models.

Language Models are Few-Shot Learners https://arxiv.org/abs/2005.14165

How to use LLMs for conceptual understanding

Start by having students select a concept they are struggling with and want to better understand (e.g. limits in calculus, wave-particle duality in physics, etc...)
Then have them choose a hobby, interest, or passion that is personally meaningful to them (e.g. sports, music, gaming)
Next, prompt the LLM (like ChatGPT) to explain the concept using examples related to the student's interest
The LLM can make the abstract concept more concrete by tying it to the student's own experiences and passions
Students can then try quantifying or modeling parts of the explanation mathematically to deepen understanding
This makes the learning active, contextualized, and tailored to the individual

Studies show relating concepts to personal interests improves engagement and comprehension

Click here for examples on relating, e.g., the concept of "limits" in Calculus, voltage and quantum superposition in Physics, and cellular respiration in Biolology to Blackpink!, and more!.

How to use LLMs for more personalized examples

In problem sets, have students identify or choose any question in their lecture, textbook or online.
Prompt the LLM to modify the question context to the student's personal interests. This creates more relatable, real-world worked examples. LLMs can generate a breadth of personalized examples on demand. Tailoring questions to students' lives improves motivation and connection to the material.
After working through the personalized example, have students complete the original generic question
This transfers the learning to abstract contexts while still leveraging the engagement from personalized examples
Variations on this idea include:
- Students submitting their interests and LLMs generating personalized homework
- Students picking examples for peers based on shared interests
- Teachers assigning personalized questions aligned to students' profiles

Click here for on how to personalize your examples.

How to use LLMs for increasing problem-solving and critical thinking skills

Prompt engineering with "Chain of Thought" - Have students break down complex problems into logical step-by-step reasoning chains. Use these chains of thought to prompt the LLM to walk through solving the problem. This scaffolds the problem and builds problem-solving skills.

Collaborative problem solving - In groups, have students discuss and solve problems together using the LLM as a reference. Encourage them to identify any inconsistencies in the LLM's logic through peer discussion.

Evaluating LLM solutions - Present an LLM's solution to a problem and ask students to critique its validity, identify assumptions, assess the argument, and suggest improvements. hones critical analysis abilities.

Click here for more details.

Research in Generative AI in Science and Mathematics Education

Please click in this Section to find out research on generative AI in science and mathematics education; including mathematics, computer science, physics, chemistry and biology education!

GenAI Educational Technology for Science and Mathematics

Improving performance of LLMs on Science and Math Problems

Generative AI with diagrams

Examples Generative AI integrated Active learning Lesson Plans for Secondary/High School Mathematics

Examples Generative AI integrated Active learning Lesson Plans for University/Tertiary Mathematics

Project on the Teachable Machine and Green Sustainability using Plastics Classification

This is joint work with Prof. Stephen CHOW and his Ed.D. student Mr. Lim. In this work, we explore how using a teachable machine for plastics classification education influences students' knowledge, attittude and behavior towards green sustainability measures.

ICAP Model for Active Learning with Generative AI

From the seminal work of Professor Michelene CHI on The ICAP Framework: Linking Cognitive Engagement to Active Learning Outcomes , we aim to see how generative AI integrated teaching strategies affects the ICAP model. Here, we aim to classify how does generative AI integrated into a learning activity affects its ICAP classification of the learning activity (if at all)?

Educational Policy on Generative AI in Education

UNESCO's Guidance for generative AI in education and research: https://www.unesco.org/en/articles/guidance-generative-ai-education-and-research
European Union Publication on Ethical guidelines on the use of artificial intelligence (AI) and data in teaching and learning for educators https://op.europa.eu/en/publication-detail/-/publication/d81a0d54-5348-11ed-92ed-01aa75ed71a1/language-en Here is the official summary of this report
Study Budy or Influencer - Inquiry into the use of generative artificial intelligence in the Australian education system, by House of Representatives, Standing Committee on Employment, Education and Training, Australian Parliament https://www.aph.gov.au/Parliamentary_Business/Committees/House/Employment_Education_and_Training/AIineducation/Report
Policy implications of artificial intelligence (AI), Research Briefing POSTnote, UK Parliament https://post.parliament.uk/research-briefings/post-pn-0708/
ISTE's Artificial Intelligence in Education website: https://iste.org/ai
UNESCO's AI competency frameworks for teachers and students https://www.unesco.org/en/articles/ai-competency-framework-teachers
US Dept. of Education's AI Guidance in response to US Government's M-24-10 Bill on Governance, Innovation and Risk Management for Agency Use of AI https://www.ed.gov/about/ed-overview/artificial-intelligence-ai-guidance
Canadian Governement's "Guide on the general use of AI" https://www.canada.ca/en/government/system/digital-government/digital-government-innovations/responsible-use-ai/guide-use-generative-ai.html (but not specifcally geared towards education)

Page updated

Google Sites

Report abuse