We keep finding new ways to bring AI into education

AX

Education AX: from management to experience

We help learning platforms shift from management to experience. All learning activity is collected as learning data built on global standards, and AI analyzes it to suggest what administrators, instructors, and learners do next.

1. 01

   Collect

   Lecognizer

   Collects and stores learning activity as learning data using global standards (xAPI · 1EdTech Caliper) and links it to 1EdTech CASE frameworks.

2. 02

   Analyze

   Lecognizer AI
   • AI finds anomaly signals in learning data,(Patent No. 10-2883249)
   • and profiles learners with an LLM.(Patent No. 10-2767110)
3. 03

   Suggest

   LearnHubble AI

   On a teaching and learning platform where people interact with AI, it suggests next steps along with the evidence.

4. 04

   Next semester

   Administrators · instructors · learners

   People judge and apply. The results build up again as learning data and inform next semester’s learning design.

AI Model

AI learning analytics: models we built, used in the field

We build our own AI models that diagnose learning levels, find unusual signals, and read learner tendencies from learning data, and have applied them to our products and public education projects. You can check the certificate or patent behind each model.

• Learning diagnosis and recommendation model

  Diagnoses learning levels from learning behavior and experience data, and recommends items and learning content for weak areas. Applied to KERIS’s ‘Toc Toc! Math Expedition’.

  (KOLAS AI model certification)

• Anomaly learning detection

  Finds unusual learning signals in xAPI learning behavior data and flags where a closer look is needed first. Used in Lecognizer AI’s anomaly detection.

  (Patent No. 10-2883249)

• LLM-based learner profiling

  Uses learning process data and an LLM to summarize learners’ tendencies and states, as a basis for customized diagnosis and item recommendations.

  (Patent No. 10-2767110)

A basis for teaching and learningAI diagnoses and suggests the next learning; instructors review it and apply it in class, and learners continue with the suggested learning on their own.

AI Agent

AI agents: working alongside learners and instructors

In LearnHubble AI, AI agents give learners explanations and hints when they need them, and suggest to instructors what needs to be supplemented. Learners write their own answers, and instructors confirm interventions and what is applied.

LLM Wiki

LLM Wiki: scattered knowledge in one place

We gather in-house documents, discussions, and answers to recurring questions into our LLM Wiki, and build AI into our workflows to refine the way we work.

1. 01

   Gather

   Documents, discussions, and answers to recurring questions that used to be scattered are kept in one place.

2. 02

   Connect

   Documents and discussions are woven into concepts and relationships (an ontology), so anyone can ask for and find the knowledge they need.

3. 03

   Draft, then decide

   AI finds and gathers relevant materials and creates a draft; the person in charge then reviews, edits, and finalizes it.

The same principle as our productsAI finds, organizes, and drafts. People decide what to use and how.

Read the LLM Wiki article