02. Meeting

17 Nov 2023

Meeting on 17 Nov 2023

1. Learner Profiles Blueprint
2. AI in Learning
3. Modeling the interaction between resilience and ability in assessments with allowances for multiple attempts

24 Nov 2023

Meeting on 24 Nov 2023

Task 1
1. SATURN of githbu official website
2. Focus on the Sequences: How to apply to education
3. Responsible: Kevin

Task 2
Try to find potential useful stuff
1. http://snap.stanford.edu/
2. https://github.com/snap-stanford

Read the papers below

1. Motifs in Temporal Networks
2. Motifs in Temporal Networks
3. Motifs in Temporal Networks
4. Using online learner trace data to understand the cohesion ofteams in higher education

01 Dec 2023

Meeting on 01 Dec 2023

Alt text

08 Dec 2023

Meeting on 08 Dec 2023

Algorithm Improvement & Monitoring

Reading Resources

1. Paul LeBlanc to Leave Presidency of Southern New Hampshire

2. A summary presentation on the Act by the European Commission

3. California Consumer Privacy Act (CCPA)

4. ARTIFICIAL INTELLIGENCE

5. American Council on Education

6. https://app.mural.co/t/bigideassignalsandtrends7760/m/bigideassignalsandtrends7760/1701781872622/be7ccf9b8888d6739b1d66e1adc1d17b5a518c69?sender=uf5dfef54353a241f80343848

Tasks

• What kind of algorithms can be used for Feedback Seeking Behivour.
  • Responsible: Haiyue
  • Direction: Processing mining/ sequence analysis/ edm/aim
• Emotinal Response to Feedback (positive vs negative)
  • Responsible: Kevin
  • Direction: Machine learning, etc.
  • Data Source: Video/Camera/Text etc.

Start Point:

1. Soft Skills: Free Measurement Tools
2. Another Resource

19 Dec 2023

Meeting on 19 Dec 2023

Aims for the meeting

A pretty clear list for what should we do.

Agenda

1. Review the Profile Blueprint [0:00 ~ 20:00]
2. Blueprint Overview on Mural from Srecko [20:00 ~ ]
3. Algorithm
4. Details of data

Basic Concepts Definitions

1. Variable: data item/characteristic/attribute (variables build indicators)
2. Indicator: How to measure/create the construct (indicators build constructs)
3. Construct: An aggregate/cluster of behaviors/attributes
4. Domains: In our case, profile domains/dimensions are combinations of various constructs: cog, metacog, affect, social, etc.
5. Model: the weighting of variables to create constructs.

Domains

1. Cognitive
2. Metacognitive
3. Affective
4. Social
5. Wellness
6. Skills

Construct

1. Mood
2. Connect
3. Motivation
4. Metacognitive Strategies
5. Cognitive Engagement

Q&A

1. Construct consists of multiple domains.
2. How to measure construct that could following multiple indicators. Intervention could be triggered by Construct level or indicator level.
   • Srecko: For MVP, all the interventions would be triggered in the variables level. Triggered in the indicator and construct level require further research.
   • Interventions are triggered by the behivours of individuals
3. Computation is expensive. How could we compute the construct levels (aggregate to a time period or real time)?
   • Defer this to a data engineering problem.

Other key points

1. Mood is a survey or something
2. Intervension is source of data
3. we talk about data build collect from student learn
4. wellness = domain

Tasks

Haiyue: Fill data source table.
Kevin: Fill algorithm table

Resources

1. Data to Domains from George Siemens
2. Interventions from Tanya Gambby
3. Blueprint Overview on Mural from Srecko