- 2069 Bhadra · 6 marks
Learning framework with example
Short answer
A learning system consists of an environment providing examples, a learning element that updates a knowledge base using feedback from a critic, and a performance element that uses the knowledge to act. For a spam filter, the environment supplies labeled emails, the performance element classifies new mail, the critic compares to user corrections, and the learning element adjusts word weights to reduce errors.
Explain the learning framework with suitable example.
A learning system improves its performance at some task with experience. Rich & Knight and Russell & Norvig describe a general framework made of a few basic components.
Block diagram
Environment
| (examples / experience)
v
+-----------+ knowledge +-------------+
| Learning |-------------->| Knowledge |
| element |<--------------| base |
+-----------+ feedback +------+------+
^ |
| v
+-----+------+ evaluation +-------------+
| Critic |<--------------|Performance |--> action
+------------+ | element |
+-------------+
Components
- Environment: supplies the training examples or situations.
- Learning element: improves the knowledge base using the information from the environment and the critic's feedback.
- Knowledge base: stores what has been learned (rules, weights, decision trees).
- Performance element: uses the knowledge to carry out the task and choose actions.
- Critic: compares performance with a standard and sends feedback to the learning element.
- Problem generator (optional): suggests new experiments to explore.
Design issues
- Which component is to be improved, and how it is represented (rules, networks, trees).
- What feedback is available (labelled examples, rewards, none).
- Type of learning: rote, by advice, from examples (induction), analogy, discovery.
Example: spam filter
- Environment: incoming emails marked "spam" or "not spam" by users.
- Performance element: classifier that labels a new email.
- Critic: compares the label with the user's correction.
- Learning element: updates word weights/rules to reduce errors.
- Knowledge base: the word weights.
After many emails the filter's accuracy improves, which shows learning.
✓ Answer checked by Claude Opus 5.5 (high effort)