Research Project · GEM Lab · The University of Hong Kong
An interactive annotation platform that helps qualitative researchers code interviews, perform thematic analysis, and build reliable label sets — through three complementary modes: Manual, LLM-assisted, and Active Machine Learning.
MNotation (short for Machine + annotation) is a research platform developed at the 💎 GEM Lab, The University of Hong Kong. It is designed to support qualitative coding workflows — especially interview transcripts, thematic coding, and other open-text annotation tasks — by combining traditional manual labelling with modern AI assistance and human-in-the-loop active learning.
The system was built for classroom and workshop settings: researchers can import a dataset, configure a label taxonomy, and invite collaborators to annotate in three distinct modes — comparing speed, agreement, and label quality across human, LLM, and active-learning workflows.
Each mode is purpose-built for a different stage of the qualitative coding pipeline.
01
Pure human annotation — the ground truth.
Researchers code each unit one-by-one against a configurable taxonomy. Designed for mobile and desktop, with single-tap label selection, instant progress feedback, undo support, and live synchronisation to the moderator dashboard.
02
AI predicts, human verifies.
Each unit is pre-labelled by a large language model using researcher-defined prompts (zero-shot and few-shot). Annotators can Accept, Override, or Customise the prompt — capturing both the prediction and the human correction for downstream analysis.
03
The system chooses what to label next.
Our ED-AL v1 algorithm combines Shannon entropy (uncertainty) with TF-IDF k-center greedy selection (diversity) to pick the most informative samples for human review — getting more reliable labels with far less annotation effort.
MNotation is primarily used for qualitative coding tasks, including:
Code utterance-level data from semi-structured interviews.
Develop and apply codebooks across qualitative datasets.
Cluster, label, and verify free-text answers at scale.
Annotate tutoring dialogues and learner conversations.
MNotation is developed at the 💎 GEM Lab, The University of Hong Kong.
Researcher
💎 GEM Lab, The University of Hong Kong
Researcher
💎 GEM Lab, The University of Hong Kong
GEM stands for Get Everyone Moving and Guiding AI in Education: Methods and Applications. Established in 2024, the lab brings together researchers from HKU, CMU, and beyond to study how generative AI and learning analytics can support — rather than replace — human educators and researchers.
Visit GEM Lab →An AI-assisted systematic literature review companion to MNotation — extending the same human-in-the-loop philosophy from interview transcripts to academic papers.
While MNotation focuses on coding interviews and other qualitative data, our upcoming SLR System brings the same philosophy — AI suggests, human verifies — to literature annotation. Upload your PDFs, apply your coding scheme, and verify results with full evidence traceability.
Preview the SLR System →Upload up to 50 PDFs or a ZIP archive for systematic analysis in one session.
Automatically apply your pre-defined coding scheme to identify themes across literature.
Surface relevant passages that support coding decisions with precise document navigation.
Review AI suggestions transparently — you retain control of every coding decision.
AI automatically applies your coding scheme to each document. Review the suggested labels side-by-side with the original PDF and adjust as needed.
AI surfaces relevant evidence passages from the document. Click any evidence to jump to its location in the PDF. Rate whether each piece of evidence supports your coding decision.
The SLR System is under active development. To be notified when it launches — or to participate in early testing — please reach out via our Contact page.
We welcome collaborations with researchers, educators, and library staff interested in qualitative coding, learning analytics, or literature review tools.
For demos, collaborations, or library workshops:
Please mention "MNotation" in the subject line so we can route your message to the right team member.
