PROGRAMARE DECLARATIVĂ ÎN ÎNVĂȚAREA AUTOMATĂ

DECLARATIVE PROGRAMMING IN MACHINE LEARNING

 

Teams code: ckgdp36

 

Lecture 

 Lecture 1-2

 Course requirements. Introduction in declarative programming and ML
 Lecture 3-6 Ethical AI principles from the design to implementation of software solutions
 Lecture 7-8 Open source AI
 Lecture 9-12 Case studies from industry.
 Lecture 13-14    Case studies and projects presentations

 

 

Seminar/Lab

 1  Discuss requirements. Choose project subject for the report and programm implementation. 
 2  Develop design report using ethical AI principles.
 3  Report 1 presentations.
 4  Report 1 presentations. 
 5  Implementation.
 6  Implementation.
 7   Report 2 project presentations.

 

 

Requirements

Report 1:

Content:  A written structured report (2-4 pages).

  1. Description (short) of the application/functionality chosen.
  2. In detail identification of possible issues based on at least 4 ethical AI principles.
  3. Presentation of mitigation strategies to the issues identified that can be used in implementation.

Step to present:

  1. Choose the subject of the report.
  2. Schedule report presentation (in weeks 5-8) by week 4. Delayed presentations (after week 8) are penalised by 3 points per week of delay.
  3. Prepare report and present it in weeks 5-8 (aim at 5-7 mins presentation and 3-5 mins questions).

You are strongly encouraged to use the template report discussed at the lecture.

 

Report 2:

Content: The implemented experiments, features, analysis, comparisons are presented. The implementation must contain some AI appproach or ML model and use some tools/techniques/algorithms to analyse/mitigate at least two of the ethical AI principles issues identified.

You may use a powerpoint, demo, workbook etc.

Steps to present:

  1. Implement functionalities/conduct experiments/perform analysis
  2. Schedule presentation in weeks 11-14 (at the lecture or seminar/lab).
  3. Upload code on git or similar and present it (5 mins).

 

Medical motivations must be presented in physical form only, at the first activity you are attending, at the beginning of the class. 

 

Evaluation

 

30%  Report 1: (Week 5-8)  This is a written report that focuses on the design of an application focusing on ethical AI principles. Must be presented.
30%  Report 2: (Weeks 11-14) This is the presentation of the actual implemented application, experiments, models, considering the ethical AI principles discussed. Must be presented.
40%  Written exam in the exam session.

 

The final grade must be at least 5. The average between Report 1 and Report 2 must be at least 5. No projects can be presented after week 14.

In the retake session, only the written exam can be repeated.