ERTCThe Hong Kong Polytechnic University
← ERTC Hub
R&D Framework

RIDE-I: from theory to deployed practice, with evidence at every gate.

RIDE-I is the Centre's operating model — a disciplined pipeline that carries an idea from theoretical claim to classroom deployment, insisting on independent evaluation before implementation. It is why ERTC ships working systems, not slideware.

RIDE-I is not a line but a lattice. Every stage is interrogated along two orthogonal axes — four commitments it must honor, and four components it acts upon — so a single learning system is resolved into a grid of questions that must each be answered with evidence. The stages are how the work moves; the axes are what the work is accountable to.

Axis I · four commitments every stage must honor — the cube's rows
Θ
Theory
Advanced theories of learning — adaptive, personalized, engaging. A standing commitment to pedagogy that keeps evolving with the evidence, not a fixed method.
Technology
State-of-the-art enabling technologies — treated not as tools but as enablers that make learning more interactive, accessible, and efficient.
Application
Strategic, appropriate application — the right method in the right context, so technology integration is purposeful and beneficial rather than merely novel.
Efficacy
Proven efficacy of learning — continuous, evidence-based validation. Novel is not enough; only what is measurably impactful is allowed to survive.
Axis II · four components every stage acts upon — the cube's columns
L
Learners
Human learners — the agents whose growth the whole system exists to serve.
E
Environments
Learning environments — physical, digital, and hybrid settings in which learning takes place.
P
Processes
Learning processes — the interactions and trajectories that carry a learner forward over time.
R
Resources
Learning resources — human, physical, and digital; static content, and dynamic content (games, ITS).
Axis III · five stages, from claim to classroom — the cube's depth
STAGE 01

Research

R

Theory first. AA/HAA/SAA, psychometrics (GPT models, SATA, LCC, BKT extensions), and the learning-science evidence base define what a system must respect before a line of code exists.

STAGE 02

Innovation

I

New mechanisms derived from the theory: dual O×A scoring, learner-as-questioner (Q2L), authorship-by-defensibility, self-critiquing AI markers.

STAGE 03

Development

D

Cloud-native builds — Cloud Run, GCS, xAPI/LRS instrumentation — so every learner interaction becomes analyzable evidence from day one.

STAGE 04

Evaluation

E

Independent, rubric-governed review of pedagogical effectiveness — an FDA-analogous stance toward AI-enabled educational products (PERP).

STAGE 05

Implementation

I

Deployment into real courses, grants, and institutional practice — with longitudinal analytics feeding back into Stage 01.

The loop closes. Implementation data — xAPI streams from live courses — re-enters Research, so the framework improves the systems and the systems test the framework.

The formal core

The Symmetry Theorem: a learner and a dynamic resource are the same mathematical object.

Human LearnerDynamic Learning Resource
Assumption
For every human learner there exist individualized dynamic resources, environments, and processes that optimize that learner's learning. Adaptivity is not a feature added on top; it is the framework's founding premise.
Symmetry
Human learners and dynamic learning resources are mathematically symmetric — the same duality psychometrics draws between persons and items. A resource has estimable "behavior"; it can be modeled, measured, and located on a scale exactly as a learner can. This is where the Centre's measurement heritage — the GPT models, item-response theory — becomes structural, not decorative.
Consequence
Every method for studying learners therefore applies to studying resources — and resources can be tutored, refined, and made to improve themselves. This is SIAIS: Self-Improving Adaptive Instructional Systems, and it is the formal warrant for the loop above: deployment data does not merely inform the framework, it is the same kind of evidence the framework was built to read.
← ERTC Hub