The lab members meet once a week at a lab seminar. Starting from spring 2016 this seminar is officially registered as the IV127 course. The language of the seminar is typically Czech (unless we have some visitor), the instructions are nevertheless in English (to fit with the rest of the lab website and to provide orientation for visitors).

The seminar has two parts: "presentation" and "free discussion". For those enrolled in the IV127 course the first part is compulsory. The seminar is open to anybody interested.

##### Presentations

Presentations are of the following types:

Presenters are expected to have slides and prepare the presentation thoroughly. Presentations are nevertheless interactive (discussion during the presentation). Slides should contain mainly "pictures" (illustrations, graphs) or equations (minimize text, bullet points). Presentations should try to be stand alone (minimal prior knowledge necessary).

Typically there are two "smaller" presentations (approx. 20 minutes + discussion each) or one "larger" presentation (approx. 40 minutes + discussion).

##### Free discussion

Discussion of hot issues, brief progress reports, results of data analysis, ... This part may depend on "prior knowledge" about the work done in the laboratory, discussion thus may be hard to follow and understand. Although it is often the most interesting part of the seminar, it is thus not compulsory.

### References

These references are intended mainly for students starting to work within the laboratory. The referenced books are often available in FI library.

###### High-level overviews, motivation

###### Student modeling, educational data mining

Overviews of notions and techniques directly relevant to our research.

- Student modeling techniques
- Good overview paper: A review of recent advances in learner and skill modeling in intelligent learning environments (Desmarais, Baker)
- The Elo rating system (which we heavily use): Applications of the Elo Rating System in Adaptive Educational Systems,
- Bayesian knowledge tracing (which we do not use, but is used very often by others): Properties of the bayesian knowledge tracing model (van De Sande)
- Metrics for Evaluation of Student Models

- Wider context
- Big Data and Education (Baker, "MOOT")
- Educational data mining: a review of the state of the art
- Practical Learning Analytics (Coursera)

- For current research directions take a look at:
- Most relevant conferences: Educational Data Mining, Intelligent Tutoring Systems, Artificial Intelligence in Education, Learning Analytics and Knowledge, User Modeling, Adaptation and Personalization
- Most relevant journals: Journal of EDM Journal of AIED Journal of User Modeling

###### Machine learning, statistics

At least basic knowledge of machine learning and statistics is necessary for most of our research and development.

- Highly recommended sources
- Machine learning (Ng, Coursera)
- Top 10 algorithms in data mining

- Advanced sources (not necessary for most of the work done within our group, but definitely useful)
- Probabilistic graphical models (Koller): coursera + book
- Learning from data: MOOC + book
- Probabilistic Programming and Bayesian Methods for Hackers (available online, with Python code)
- Pattern recognition and machine learning (Bishop)
- Doing Bayesian Data Analysis (Krusche)

###### Other related areas

It is useful to have understanding of at least the basic notions in the following domains:

- Item response
theory:
- A visual guide to item response theory
- book "The theory and practice of Item response theory"

- Collaborative filtering:
- Matrix factorization techniques for recommender systems (Koren, Bell, Volinsky)
- Advances in Collaborative filtering (Koren, Bell)

- Cognitive psychology:
- Cognitive psychology (Eisenck, Keane)

- Education, learning sciences:
- Applying the Science of Learning (Mayer)