U19967 Artificial Intelligence Level 6 Portfolio Assessment 2025/26 | CCCU

ASSESSMENT REQUIREMENTS

Assessment Feedback

Feedback to support the work will be given in the classes
Formal feedback will be within 15 working days after the submission deadline.

Detailed Assessment Guidance

Activity 1:

In predicate logic, there are many different rules of inference. Being universally true, they may be used either to validate complete arguments or to generate conclusions. Moreover, individual rules of inference may either be used on their own or may be applied in conjunction with others.

Given the following list of basic inference rules:

1.Modus ponendo ponens (MPP): A  B, A |-- B
(e.g. IF my program is correct THEN it will run, my program is correct, THEREFORE it will run)

2.Modus tollendo tolens (MTT): AB, ~ B |-- ~A
(e.g. IF my program is correct THEN it will run, my program will NOT run, THEREFORE it is not correct)

3.Double negation (DN): A |-- ~(~A)
(e.g. My program has run THEREFORE My program has not not run)

4.&introduction (&INT) A,B |--(A & B)
(e.g. My program has run; it is correct THEREFOR my program has run AND is correct)

5.Reductio ad absurdum (RAA): AB, A~B |-- ~A
(e.g. IF my program is correct THEN it will run, IF my program is correct THEN it will NOT run THEREFORE My program is not correct)

6.Universal specialisation (US): Ɐ(X) W(X), A |-- W(A)
(e.g. All things which are computers are unreliable, a ‘TIPTOP’ is a computer THEREFORE a
‘TIPTOP’ is unreliable).

A.Given the rule:
‘ IF artificial intelligence is a growing subject THEN there is no shortage of applicants’ and the fact ‘there is a shortage of applicants’

• Can you use the list of inference rules above to prove that ‘artificial intelligence is not a growing subject’.
  Explain step by step your reasoning, the rules used and provide a summary of the formal structure of your proof.
• Is there an alternative way to prove the same fact ‘artificial intelligence is not a growing subject’? Explain step by step your reasoning, the rules used and provide a summary of the formal structure of your proof.

Activity 2:

The following is the rule set of a simple expert system for diagnosing plant issues:

1. IF leaves are yellow AND soil is dry → THEN watering_status = underwatering
2. IF leaves are yellow AND soil is wet → THEN watering_status = overwatering
3. IF watering_status = underwatering → THEN problem = plant is dehydrated
4. IF watering_status = overwatering → THEN problem = root rot risk
5. IF problem = plant is dehydrated AND temperature is high → THEN diagnosis = water more frequently
6. .IF problem = root rot risk AND drainage = poor → THEN diagnosis = improve soil drainage

A.Use forward chaining to reason about the weather if the working memory contains the facts: leaves are yellow, soil is dry, temperature is high. Show your answer in a table naming the rules matching the current working memory (conflict set), which rule you apply, and how the working memory contents changes on the next cycle after a rule has fired.

B.Use backward chaining to reason about the plant issues diagnosis if the working memory contains the fact: problem = root rot risk. Show your answer in a similar table.

C.Provide your own example demonstrating backward chaining to reason about the plant issue diagnosis. Show your answer in a similar table.

D.Suppose that the user interface of our expert system allows the system to ask a user about the facts whether they are true or false. What question (or questions) should the system ask the user in order to conclude that the diagnosis is improve the soil drainage? What will the user answer? Which rule will require clarification from the user?

Activity 3:

Can you design then implement a mini medical diagnosis expert system in prolog:

1.Define a set of symptoms and corresponding facts that represent medical conditions: Example:
Fact: "Fever above 38°C"
Fact: "Persistent cough for more than two weeks" Fact: "Severe headache accompanied by nausea" Fact: "Shortness of breath"
Fact: "Joint pain and swelling"

2.Create rules based on medical knowledge to infer possible diagnoses from the symptoms provided. For example:
Rule: If the patient has a fever above 38°C and persistent cough for more than two weeks, consider tuberculosis.
Rule: If the patient has shortness of breath and wheezing, consider asthma.
Rule: If the patient has severe headache accompanied by nausea and vomiting, consider migraine or meningitis.

Activity 4:

Let a Finite State Machine (FSM) A be defined by A = (Q, Σ, q0, δ, F) with: Q

= {0, 1, 2, 3}    Σ = {a, b}    q0 = 0    F = {3}

And the transition function δ:

• Q = set of states
• Σ = input alphabet
• q0 = initial state
• δ = transition function
• F = set of accepting (final) states.

1.Draw this FSM (hand-drawn).

2.Give the shortest word recognized by the automate.

3.Give an example of a word non recognized by the FSM.

B.Design a finite state machine (FSM) for a non-player character (NPC) that can navigate a maze. Provide the formal definition of your FSM, including:

• The set of states
• Input alphabet
• The initial state
• The transition function
• The set of final (goal) states
  Provide the diagram to illustrate your FSM (hand-drawn).

Activity 5:

Choose a local, real-world domain that you are personally familiar with - such as:

• A traditional local market or trade system
• A student club or campus activity
• A local transportation system.
• A family-owned business or digital service in your area.

Task:

1.Identify the domain and describe it in 3–5 sentences.

2.Design a simple ontology for this domain, including at minimum:

• 4–6 classes
• At least 2 subclass relationships
• At least 2 object properties
• At least 2 data properties
• 2–3 individuals (instances)

3.Design the ontology as a diagram (using Protege).

4.Explain your modeling choices, especially:

• Why you chose these classes and properties
• How the ontology supports reasoning or querying in that domain (provide examples using SWRL or SQWRL or another ontology querying language of your choice)
• Any assumptions or challenges in modeling.

Instructions for submission:

You are expected to submit two elements:

• Portfolio submission: a document in word or pdf that contains the detailed solution for the activities including the necessary explanation for each question, a photo of the hand-drawn solution where necessary.
• Video Submission: Record and upload a short video (less than 15 minutes) to the ‘Video Recording
  Submission’ bucket on Blackboard, covering the following activities:
  -Activity 2: Demonstrate, using cards and paper, how forward chaining and backward chaining work in the diagnostic system for plants. How would you do it to explain your answers to a 6-year-old child?
  -Activity 3: Demonstrate your medical diagnosis system on SWI-Prolog with examples of queries. Use the ‘?- trace’ function to show how facts and rules are evaluated to reach the results. Explain the reasoning steps clearly.
  -    Activity 5: Present your ontology implemented in Protégé and discuss your modeling choices as well as the execution of some queries.

Marking Criteria

In the appendix pages 5 and 6.

Further Information

-In class recordings.
-Use the last 30 minutes of the practical sessions to seek feedback from the module leader.