Model-based Learning Task

A Model-based Learning Task is a learning task that requires a learned model in a representation language that is more general than the representation language used to describe the training data.

• AKA: Model Induction/Training.
• Context:
  • Input: Learning Records.
    • optional: a Model Family.
  • output: Learned Model.
  • It can range from being a Supervised Model-based Learning Task (such as Semi-Supervised Model-based Learning) to being an Unsupervised Model-based Learning Task/Model-based Unsupervised Learning Task.
  • It can range from being a Non-Parametric Model Training Task to being a Parametric Model Training Task.
  • It can range from being a Parametric Model-based Learning Task/Model-based Parametric Learning Task to being a Non-Parametric Model-based Learning Task/Model-based Non-Parametric Learning Task.
  • It can be solved by a Model-based Learning System (that implements a Model-based Learning algorithm).
  • …
• Example(s):
  • a Statistical Modeling Task.
  • a Decision Tree Learning Task.
  • a Language Model Training Task.
  • a Vector Space Model Learning Task.
  • a Probabilistic Graphical Model Learning Task.
  • …
• Counter-Example(s):
  • a Model-Free Learning Task, such as a lazy learning task.
  • a Model Deployment Task.
  • an Approximate Search Task.
• See: Supervised Learning Task.

---

---

References

2003

• (Perlich & Provost, 2003) ⇒ Claudia Perlich, and Foster Provost. (2003). “Aggregation-based Feature Invention and Relational Concept Classes.” In: Proceedings of the ninth ACM SIGKDD International Conference on Knowledge discovery and data mining. ISBN:1-58113-737-0 doi:10.1145/956750.956772
  • QUOTE: Model induction from relational data requires aggregation of the values of attributes of related entities.

---