The Learn Lens: Inclusive Adaptive Agentic AI for Measuring Learning as It Happens

Primary supervisor

Charith Jayasekara

Co-supervisors

  • Dr Lakshitha Arachchige

LearnLens is an inclusive adaptive agentic AI learning system designed to support students while they learn complex and unfamiliar concepts. Its central purpose is to measure and support how learning develops, rather than evaluating students mainly through a final answer, program, circuit, assessment submission, or completed product.

The initial implementation of LearnLens will focus on teaching quantum computing concepts to computing students. Quantum computing provides an ideal context for this research because concepts such as superposition, measurement, interference, entanglement, and quantum gates are abstract and often difficult to understand through conventional teaching approaches.

A student may produce a correct quantum circuit without understanding why it works. Another student may produce an incomplete solution while demonstrating strong reasoning, meaningful conceptual growth, effective use of feedback, and an ability to correct earlier misconceptions. LearnLens aims to capture and distinguish between these different learning journeys.

The system will use an Assess As You Learn approach in which learning, practice, formative feedback, reflection, and assessment occur continuously. Adaptive AI agents will create quizzes, generate small progressive learning tasks, identify possible misconceptions, recommend suitable next activities, and provide personalised formative feedback.

LearnLens will respond to differences in prior knowledge, cognition, confidence, pace, communication, and preferred ways of engaging with learning content. This will allow the system to support a broad range of learners, including neurodivergent learners who may benefit from clearer structures, smaller task sequences, alternative representations, flexible pacing, repeated practice, or different forms of feedback.

Rather than diagnosing or labelling learners, LearnLens will adapt to observable learning needs and interactions. It may vary the type of explanation, task difficulty, amount of scaffolding, presentation format, pace, and level of guidance provided to each learner.

The system will collect evidence from learners’ predictions, explanations, attempts, revisions, reflections, responses to feedback, misconception correction, confidence, independence, and ability to transfer knowledge to new problems. This evidence will be used to construct a meaningful representation of how learning develops over time.

Although the initial project will focus on quantum computing education, the LearnLens architecture will be designed for extension into other complex learning areas such as programming, algorithms, artificial intelligence, cybersecurity, mathematics, engineering, and computer architecture.

Aim/outline

The primary aim of this project is to design, develop, and evaluate LearnLens as an inclusive adaptive agentic AI system that supports the learning of complex concepts and measures the process of learning as it occurs.

The project will investigate how AI agents can provide personalised learning experiences while maintaining clear pedagogical structure, educator oversight, technical accuracy, and responsible use of artificial intelligence.

The project will focus on the following objectives:

  • Design an agentic AI architecture that coordinates adaptive quizzes, progressive learning tasks, misconception detection, learner reflection, and formative feedback.
  • Develop an initial LearnLens prototype for teaching selected quantum computing concepts to computing students.
  • Create a learner model that represents changes in conceptual understanding, reasoning, confidence, feedback use, and independence.
  • Investigate how learning activities can be adapted according to observable differences in knowledge, cognition, pace, and support requirements.
  • Design inclusive learning pathways that support diverse and neurodivergent learners without making assumptions based on labels or diagnoses.
  • Integrate quantum computing simulations and practical activities using tools such as Qiskit and quantum simulators.
  • Develop methods for measuring learning progression rather than relying only on final answers or completed products.
  • Provide personalised formative feedback that helps learners understand errors, correct misconceptions, reflect on their reasoning, and determine what to do next.
  • Develop educator facing learning analytics that make individual and cohort learning progression visible.
  • Evaluate the usability, inclusiveness, educational effectiveness, technical accuracy, and perceived value of the system.

The research may address questions such as:

  1. How can an adaptive agentic AI system support students in learning complex quantum computing concepts?
  2. What interaction evidence can be used to measure conceptual development and learning progression?
  3. How can AI generated quizzes, progressive tasks, and formative feedback be adapted to different cognitive and learning needs?
  4. How effectively can LearnLens identify misconceptions and respond with suitable learning support?
  5. How can the system support diverse and neurodivergent learners through flexible and inclusive learning pathways?
  6. How does measuring the learning process provide additional insights beyond evaluating the final product?
  7. How can educators remain meaningfully involved in reviewing, guiding, and improving AI supported learning?

The project may involve the following stages:

  1. Review research on agentic AI, adaptive learning, cognition, neurodiversity, formative feedback, learning analytics, and quantum computing education.
  2. Define the educational and technical requirements of LearnLens.
  3. Design the agent architecture, learner model, learning evidence model, and educator interaction model.
  4. Develop an initial prototype for selected quantum computing topics.
  5. Create adaptive quizzes, progressive tasks, simulations, feedback strategies, and reflection activities.
  6. Conduct a pilot study with computing students.
  7. Analyse learning progression, system interaction data, learner experiences, and educator feedback.
  8. Refine the LearnLens framework and investigate its transferability to other learning domains.

The expected outcome is a validated LearnLens prototype and an evidence informed framework for using inclusive adaptive agentic AI to support and measure learning as it happens.

URLs/references

LearnLens project materials

Project website and repository links will be added as the system is developed.

IBM Quantum and Qiskit

  1. https://www.ibm.com/quantum/qiskit
  2. https://quantum.cloud.ibm.com/learning/en
  3. https://quantum.cloud.ibm.com/docs/en/guides

IBM Quantum learning modules for computer science

  1. https://quantum.cloud.ibm.com/learning/en/modules/computer-science

Required knowledge

Applicants (or group) should have a background in computer science, information technology, software engineering, data science, artificial intelligence, education technology, or a related discipline.

Useful knowledge and skills include:

  • Programming experience, preferably using Python.
  • An understanding of software development and system design.
  • Basic knowledge of artificial intelligence, machine learning, or generative AI.
  • An interest in agentic AI systems, large language models, or intelligent tutoring systems.
  • An interest in learning analytics, adaptive learning, human centred computing, or educational technology.
  • An interest in inclusive learning and supporting diverse learner needs.
  • Basic understanding of research methods and data analysis.

Prior knowledge of quantum computing is helpful but not essential. Students can develop the required quantum computing knowledge during the project. Familiarity with Qiskit, quantum circuits, linear algebra, user experience design, web application development, or educational research would be advantageous.

A technically focused students may concentrate on the agent architecture, learner modelling, system development, and analytics. A student with an education or human centred computing focus may concentrate on inclusive design, learner cognition, formative feedback, usability, and evaluation.