Skip to content

DA5. Consistency-Based Diagnosis and Abductive Diagnosis

Statement

Suppose, you have a job at a company that is building online teaching tools. Because you have taken an AI course, your boss wants to know your opinion on various options under consideration.

They are planning on building an intelligent tutoring system for teaching elementary physics (e.g., mechanics and electromagnetism). One of the things that the system must do is to diagnose errors that a student may be making.

The boss has heard of consistency-based diagnosis and abductive diagnosis but wants to know what they involve in the context of building an intelligent tutoring system for teaching elementary physics.

For each of the following, answer the explicit questions and use proper English. Answering parts not asked or giving more than one answer when only one is asked for will annoy the boss. The boss also does not like jargon, so please use straightforward English.

  • Explain what knowledge (about physics and about students) is required for consistency-based diagnosis.
  • Explain what knowledge (about physics and about students) is required for abductive diagnosis.
  • What is the main advantage of using abductive diagnosis over consistency-based diagnosis in this domain?
  • What is the main advantage of consistency-based diagnosis over abductive diagnosis in this domain?

Your Discussion should be at least 250 words in length, but not more than 750 words. Use APA citations and references for the textbook and any other sources used.

Answer

Introduction

An intelligent tutoring system for teaching elementary physics is a system that builds teaching plans for students. It should be able to observe students’s performance and reactions and be able to explain it in order to generate rules that dictate its future actions with the end goal of improving the agent’s results, and hence, the student’s learning.

The process is named diagnoses, and the prompt suggested two methods to do so: consistency-based diagnosis and abductive diagnosis. Both methods are used to infer the cause of a problem (error), but they differ in the way they do so. Each has its own advantages and disadvantages, and the following sections will discuss them.

Consistency-Based Diagnosis (CBD)

In a consistency-based diagnosis, the system makes assumptions about how it should work normally; hence, flagging the components that deviate from the expected behavior as faulty. In other words, it checks how consistent the observed behavior is with the expected behavior. By not allowing faults to be included in its assumptions, the system can use any conflict with the knowledge base (expected behavior) as a sign of a fault (Poole & Mackworth, 2017).

The knowledge required for consistency-based diagnosis is the expectations that a student should gain from the system; that is, the correct physics concepts, a set of questions and their answers, and a set of expected student approaches that are considered normal.

Abductive Diagnosis (ABD)

In abductive diagnosis, the system tries to find the best explanation for the observed behavior. It does so by generating a set of hypotheses and then selecting the best one. In other words, the system observes the effect of the behavior and tries to find possible causes based on the rules of its knowledge base. It allows for storing faulty/abnormal behavior which may give the system wider flexibility in diagnosing the problem (Poole & Mackworth, 2017).

The knowledge required for abductive diagnosis is the same as consistency-based diagnosis (physics concepts, questions and answers, and expected student approaches). However, it also requires a set of wrong answers and the reasoning behind them. It may also include students’ profiling data, such as their learning style, previous answers and courses taken, and the time taken to answer a question.

Advantages of ABD over CBD

The main advantage of abductive diagnosis over consistency-based diagnosis is that it allows for the faulty behaviors to be included in the knowledge base resulting in detailed and clearer diagnoses. It also allows for diagnosing problems that have no normal behavior or not explicitly defined in the knowledge base.

Advantages of CBD over ABD

The main advantage of CBD that is is simpler, requires less space and computational power, and less prone to errors. The explanations of ABD are good, but they may lead to false positives or negatives errors which are not present in CBD. CBD is also more suitable when the knowledge about the environment is well defined and if the question is only about the correctness of the behavior, and not the reasoning behind it.

Conclusion

Both ABD and CBD are used to diagnose problems in an intelligent tutoring system. ABD is more flexible and can diagnose problems that are not explicitly defined in the knowledge base, and it provides detailed explanations for the reasoning behind its decisions. On the other hand, CBD is simpler, requires less computational power, and is less prone to errors.

In the context of an intelligent tutoring system for teaching elementary physics, ABD may be more suitable because its reasoning provide great benefit for the users. For example, if a student solves a problem in a new approach that is not included in the knowledge base, an ABD system can still diagnose if the solution is correct or not. In contrast, a CBD may flag the new approach as wrong because it is not included in the knowledge base.

References

  • Poole, D. L., & Mackworth, A. K. (2017). Artificial Intelligence: Foundations of computational agents. Cambridge University Press. <https://artint.info/2e/html/ArtInt2e.html> Chapter 5 - Propositions and Inference.