103333 - ARTIFICIAL INTELLIGENCE TECHNIQUES

Subject type: Mandatory

Coordinator: Alfonso Palacios González

Trimester: Third term

Credits: 4

The subject will be taught in Catalan. Students will be able to address the teacher in the language that is most comfortable for them.

Some content, transparencies and bibliography will be in English.

Artificial intelligence is a discipline that studies intelligent agents, understanding as such those devices (software and / or hardware) that perceive the environment, reason and take action to achieve their goals. In recent years artificial intelligence has hit the industry with great force and many analysts believe it will be the main factor in the upcoming industrial revolution.

During the course there is an introduction to the most classic Artificial Intelligence with an in-depth study of the search and logic algorithms that are used today to solve countless problems. For example: google search, google maps, Amazon and Netflix recommending systems, making schedules, autonomous vehicles, video games, and a long etc. The last chapter gives a brief introduction to machine learning, more specifically to the classification and clustering that are the basis of data analysis algorithms.

This subject has methodological and digital resources to make possible its continuity in non-contact mode in the case of being necessary for reasons related to the Covid-19. In this way, the achievement of the same knowledge and skills that are specified in this teaching plan will be ensured.

At a general level, this subject contributes to the following learning outcomes specified for the subject to which it belongs (Algorithmic and Programming) 

• Know the objectives, foundations, history, state of the art and the different paradigms of problem solving of artificial intelligence
• Know the basic techniques and methodologies of artificial intelligence: problem solving through search and techniques of knowledge representation and reasoning
• Appropriately use theories, procedures and tools in the professional development of computer engineering in all its areas (specification, design, implementation, deployment, implementation and evaluation of products) in a
  to demonstrate an understanding of the commitments made in the design decisions.
• Demonstrate knowledge of the ethical dimension in the company: social and corporate responsibility in general and, in particular, the civil and professional responsibilities of the computer engineer.
• Demonstrate knowledge and understanding of essential facts, concepts, principles and theories related to computer science and its reference disciplines
• Collaborate in a unidisciplinary environment. Identify the group's objectives and collaborate in the design of the strategy to be followed and a work plan to achieve them. Identify the responsibilities of each member of the group and assume the personal commitment of the assigned task. Evaluate and present own results. Identify the value of cooperation and exchange information with the other members of the group. Exchange information about the group's progress and propose strategies to improve its operation
• Control versions and configurations of the project.

At a more specific level, at the end of the course the student must be able to:

• LO1: Explain what Artificial Intelligence is and its applications
• LO2: Express a problem as a search
• LO3: Choose the most appropriate search algorithm for each problem
• LO4: Design heuristic functions to make searches more efficient
• LO5: Represent knowledge in a set of rules and facts
• LO6: Use inference algorithms to reason about the knowledge represented
• LO7: Explain the difference between classification and clustering
• LO8: Use classification and clustering algorithms and know how to interpret the results

All the theoretical concepts of the subject will be exposed in theory classes (large groups). In these classes, and at the discretion of the teachers, exercises and problems of a more practical nature will also be solved. Likewise, and always at the discretion of the teachers, students may be asked to solve, individually or in groups, short problems and / or exercises. These activities, brief and optional, will serve the student as a tool for self-assessment of their achievement of the contents of the subject and can be used by the teacher to make decisions about the final grade of the student good and never to the detriment of the numerical grade calculated according to the grading system specified by the subject.

Concepts of a more practical nature and everything that can essentially be considered the practical application of theoretical concepts will be worked on more intensively in small (laboratory) groups. In the sessions scheduled for this purpose, the appropriate tools will be given to solve the scheduled activities well and it is expected that these will be extended from a temporal point of view, beyond the laboratory hours and that, consequently, ¨ence, students must complete them during the time of autonomous learning.

It will be made available to students activities of a completely optional nature that will help them prepare and prepare for those of a compulsory nature.

1 Introduction to Artificial Intelligence    
    1.1 History
    1.2 Applications
    1.3 Ethics and feminism
2 Troubleshooting    
    2.1 Troubleshooting
    2.2 Uninformed search: BFS, DFS
    2.3 Informed search: voracious search, algorithm A
    2.4 Heuristic functions
    2.5 Search in games: minimax, alpha-beta prunning
    2.6 Satisfaction of restrictions
3 Logic    
    3.1 Representation of knowledge: facts and rules
    3.2 Inference or reasoning algorithms
4 Machine learning    
    4.1 Supervised. Classification: N-nearest neighbors, decision trees, Naive Bayes
    4.2 Unsupervised. Clustering: K-means

In order to gather evidence of the achievement of the expected learning outcomes, the following evaluative activities will be carried out:

Practices: there will be a maximum of four internships (related to all skills)

• Practice 1 Problem solving (evidence of learning outcomes RA1 - RA4)  
• Practice 2 Problem solving (evidence of learning outcomes RA1 - RA4) 
• Practice 3 Problem solving (evidence of learning outcomes RA1 - RA4) 
• Practice 4 Presentation example of implementation of one of the algorithms explained in class 

All the common and specific competences will be worked on in the practices: CIN1, CIN3, CIN7, CIN8, CIN15 and EFB3, as well as the basic B2, B3, B4 and the transversal T1 and T2

Written test: individual exam on theory and problem solving seen in class. This test collects evidence of all learning outcomes.

All the common and specific competences will be worked on in the written test: CIN1, CIN3, CIN7, CIN8, CIN15 and EFB3, as well as the basic B2, B3 and B5

The following are the most important aspects of each competence assigned to the subject:

• B2: problem solving within their area of ​​study.
• B3: gather and interpret information relevant to the subject
• B5: development of learning skills necessary for further studies (autonomous access to documentation, effective work habits)
• CIN1: design, develop and evaluate computer systems and applications ensuring their reliability, security and quality
• CIN3: show attitudes of teamwork
• CIN7: design and use of the most appropriate data structures
• CIN8: analyze, design and build applications
• EFB3: master basic concepts of discrete mathematics, logic, algorithms and computational complexity 
• EIS1: develop and evaluate systems that meet user requirements 
• T1: materials are given in English to work the third language
• T2: ability to work in a team developing different roles

In order to pass (pass) the assessment activities, students will have to demonstrate

• That they have acquired the theoretical knowledge related to the contents of the subject and that their understanding allows them to put them into practice [MECES-2 point a, point c]
• That they can develop solutions to problems that, although they are similar to others seen above, present aspects that are new [MECES- 2 point f]

Assessment:

• PR_E: Individual written test. Weighting of the final grade 60% if the grade is> = 5
• PRAC: internships from 1 to 4. Weighting in the final grade 40% (each 10%) if a minimum of two internships have been approved

Final grade calculation (NF):

• Si PR_E> = 4 and 2 or more approved practices : NF = PON = PR_E 0,60 + PRAC 0,40 
• If PR_E <4 or not 2 approved practices: NF = min (PR_E, PON)

Recovery:

The written test (PR_E) can be retrieved. The final grade will be calculated as set out above with the written test recovery grade.

Normative:

• Attendance at internships is mandatory. If a student does not attend an internship session he will be graded with a grade of 0 (zero) in the corresponding internship
• Following UPF regulations, if it is detected that a practice or a written test has been copied from a classmate, the mark will be 0 (zero) for both what has been copied and what has been allowed to be copied.
• In order for the student to be entitled to recovery he / she must have taken the written test

Russel, Stuard and Norvic, Peter (2013), "Artificial Intelligence: a moderate approach". (3rd edition) Prentice Hall.