Following the classification of Pedro Domingos, we consider five main classes of AI algorithms. For each tribe, we introduce the area, analyze some of its main topics, indicate applications and provide an overview. At the end of the course, the student/s will have a comprehensive overview of the AI area, including key approaches, possibilities, limitations and challenges. You will also be able to determine which approaches are appropriate for specific real-world problems.

Main skills to develop: Instrumental - analysis and synthesis, problem solving; Personal - critical thinking Systemic - practical application of theoretical knowledge; research.  Secondary skills: Instrumental - organization and planning; Personal - teamwork Systemic - autonomous learning; creativity.

Program:

1. Introduction to Artificial Intelligence

2. IA Symbolic

3. IA Conexionist

4. IA Bayesian

5 Overview of the Area

6. IA Evolutionary

UC aims to teach the student an overview of the area, its methodological principles, its challenges and its main applications. Introduce the basic algorithms of a data analysis pipeline: data preparation, attribute extraction and dimensionality reduction, models based on computer learning techniques and their validation.

Program:

1| Objectives of computer learning. Data analysis pipeline. Type of data, types of learning, typical problems and applications.

2| Data preparation. Preliminary data analysis. Data problems: treatment of missing values, outliers, unbalance, normalization. Attribute types and attribute conversion Binning. Univariate and multivariate outliers, continuous and categorical data imputation, duplicate detection and similarity measures.

3| Dimensionality reduction: ranking methods.  Selection of attributes.

4| Classification and regression. KNN Minimum Distance Estimation.

The discipline is a continuation of the Computational Learning - I discipline with the intention of deepening the topics already covered, exploring new concepts and introducing advanced concepts of classification and regression. 

Program:

1- Introduction

i. Pipeline review

ii. Concepts about supervised and unsupervised models

2- Unsupervised learning

i. Clustering based on combined Gaussian models

ii. Evaluation methods: Intrinsic and Extrinsic

3- Supervised models, not connectionists

i. Linear discriminants (Euclidian and Mahalanobis) and Fisher

ii. Bayes Classification; Bayes and Risk Estimation; Maximum A Posteriori (MAP).

iii. Support Vector Machines iv. Combination of sorters

4- Supervised models connectionists

i. Non-recurrent neural networks: MLP and RBF

ii. Recurrent neural networks

5- Models based on rules

Provide students with advanced concepts, principles and theories for developing real applications.

Program:

1. Agents

1.1 Rationality and Intelligent Agents

1.2 Agents, tasks and environments

1.3 PEAS 

1.4 Properties of Environments

1.5 Structure/Architecture and Taxonomy of the Agents

2. Multi-agent system

2.1 Multi-agent architectures

2.2 Cooperation

2.3 Collaboration

2.4 Negotiation

2.5 Communication

3. Uncertain Knowledge and Reasoning

3.1 Approaches to environment/world representation

3.2 Quantification of uncertainty

3.3 Probabilistic reasoning

4. Uncertain Knowledge and Reasoning with Temporal Dimension

4.1 Time and uncertainty

4.2 Transition models and sensors

4.3 Hidden Markov Models

4.4 Dynamic Bayesian Networks

5. Individual and Complex/Sequential Decision-making

5.1 Theory of Utility

5.2 Sequential decision-making in Atomic Representations

5.3 Sequential decision-making in factored representations 

Provide the student with theoretical knowledge and tools that allow to extract, select, merge and transform information so that it can be used efficiently by algorithms of analysis and computer learning.

The student must be able to:-analyze the quality of information;

- extract attributes from different types and domains;

- develop strategies for data annotation;

- convert data;

- merge information from different sources at three different levels: data, attributes and decision.

Program:

Chapter 1: Introduction to EA and FI.

Chapter 2: Raw data processing

Chapter 3: Attribute extraction-Attributes-Domain specific Attribute extraction:

Chapter 4: Data fusion-At the data level: weighted average, Kalman Filter, particle filtering.

Advanced data visualization tools and techniques; 

Exploratory data analysis techniques;

State-of-the-art interaction techniques;

Knowledge and experience in developing visualizations for AI.

Program:

 The principles of effective visualization;

 Representation of different types of data;

 Perception and cognition Interactive visualization;

 Libraries and tools for visualization;

 Interactive applications and dashboards;

 Principles for human-centered design;

 Exploratory data analysis;

 Data analysis and statistical visualization;

 Visualization for AI algorithms and models;

 Interpretation of the models;

 Visualization to understand ML models;

 Visualization of training processes;

 Visualization in LN Processing;

 Data visualization;

 Techniques to visualize high-dimensional data and reduction of dimensionality;

 Visualization of the selection and extraction of characteristics;

 Ethics in data visualization.

Provide students with methodological and application skills in the area of optimization models and methods and decision-making support, especially in the context of planning or operational problems where environmental impacts are relevant, including identifying types of problems, building mathematical models that include their fundamental characteristics, especially with multiple criteria, applying algorithms that produce solutions for the models, and carrying out a critical analysis of the solutions obtained.

Program:

1. Introduction to linear programming. Mathematical model building. The simplex method.

2. Whole programming. Mathematical model building. The branch and bound method.

3. Linear and integer multiobjective programming.

4. Multicriteria decision analysis: Problems of choice, seriation and classification; Notion of preferences; Prevalence relations (outranking); Value functions.

Present, discuss and develop computational methods for natural-inspired (i.e., biological, social) artificial intelligence solutions to high complexity problems that either have no analytical solution, or are computationally untreatable. Acquisition of skills to rigorously evaluate alternative solutions to problems.

Acquisition of skills in analysis and synthesis, oral and written communication, computer skills and statistical analysis, problem solving, knowledge of a foreign language, critical reasoning, group work, autonomous learning, creativity, practical application of knowledge, research.

Program:

1. Introduction

2. Classic Meta-Heuristics 

3. Evolutionary Algorithms 

4. Genetic programming 

5. Evolutionary Strategies 

6. Differential Evolution

7. Collective intelligence 

8. Coevolution

9. Multi-objective optimization

10. Evolutionary Machine Learning

11. Design of Experiments and Evaluation

This UC seeks that students acquire knowledge on advanced topics of computer learning and skills for developing solutions involving: deep computer learning networks, generative models and reinforcement learning models. In the end they should have the ability to analyze, model, implement, train and execute: - fully connected networks, convolutional, sequential, recursive, of graphs and "transformers" - autoencoders, variational autoencoders, generative adversarial networks and diffusion models model free q-learning and deep q-networks.

Program:

1. Model training

2. Deep Learning

3. Generative computer learning

4. Learning by Reinforcement

Study and development of generative models of AI following approaches, connectionistic, evolutionary and biological.

Analysis of the main challenges and opportunities. Learning by practice following a Project Based Learning approach.

The student/s will have a vision of the area of generative AI and will be able to develop and/ or adapt generative systems in order to respond to real needs.

Key competencies:

Instrumental - analysis and synthesis, problem-solving

Personal - critical thinking Systemic - practical application of theoretical knowledge; research.

Secondary skills:

Instrumental - organization and planning;

Personal - Systemic teamwork.

Program:

1. Introduction to Generative Artificial Intelligence 

2. Classical methods 

3. Evolutionary Generative Models

4. Techniques of Exploitation 

5. State of the Art in Generative IA 

6. Computational Creativity 

7. Applications 

Ethical Considerations, Challenges and Opportunities.

Knowledge of high-performance computing infrastructure and service management for support of massive data processing in AI applications.

Knowledge of computing at the edge for distributed learning techniques.

Planning and administration of infrastructures for support of massive data processing in AI applications.

Edge computing models in low-power hardware and System-On-Chip.

Planning and administration of centralized and edge computing infrastructures.

Program:

1. Infrastructures supporting Artificial Intelligence;

2. Management of centralized infrastructures/data centers;

3. Container orchestration systems for cloud and edge;

4. Real-time big data architectures: Kappa and Lambda;

5. Scalable and reliable transport in distributed environments; 

6. Big Data solutions;

7. Edge computing platforms on state-of-the-art hardware and software;

8. Computing on GPUs and System-on-Chips at the edge.

Fundamental concepts relating to the theory, design and implementation of time series analysis methods;

Linear and non-linear time series analysis techniques;

Understand, identify, select and implement time-series analysis and forecasting methods appropriate to the concrete problems to be solved.

Acquisition of skills:

1. Instrumental: analytical and synthesis skills in complex problems; competence to solve concrete problems in the field of time series forecasting;

2. Personal: group work; critical thinking;

3. Systemic: self-learning; research.

Program:

1. Basic concepts about time series and forecasting

2. Stochastic processes

3. Basic description techniques:

4. Linear models for stationary data:

5. Linear models for non-stationary data

6. Forecasting:

7. Classic multivariate models

8. Non-linear models

Intelligent robots are autonomous robotic systems capable of processing sensory information in order to perceive the environment in which they operate.  Building representations of a knowledge base that allow reasoning, planning and decision-making. Use of AI techniques.

Knowledge to design and implement algorithms, based on intelligent robots.

Program:

1. Perception-reasoning-action cycle ROS, mobile robot simulators.

2. Extraction of information from sensory data.

3. Representation of environment, location and SLAM Probabilistic Maps.

4. Planning, reasoning and decision-making under uncertainty. Movement planning with sampling. Planning and exploration based on information theory. Reasoning using formal logic. Partially observable Markov processes.

5. Human-robot interaction. Interaction modalities. Principles and theories of man-robot interaction. 

6. Application case studies service robots.

The objectives include the acquisition of knowledge related to the execution of Computer Vision applications using "Deep Learning" tools.

Following the completion of this course, the student will acquire skills that will allow him to apply "machine learning" techniques and, in particular, "deep learning", in computer vision applications that require detection and recognition of entities/ objects, semantic segmentation, optical flow and motion analysis, pose estimation and "SfM-Structure from Motion".

Program:

Introduction to Computer Vision. Review of the fundamentals of "Deep Learning".

Classification of entities and images.

Detection and recognition of objects.

Optical flux.

Visual Tracking of Objects.

Pet pose.

Structure Recovery by Movement (SFM).

The following skills are expected to be acquired:

- Methods for the representation and computational manipulation of natural language;

- Application of various techniques for interaction in natural language, namely in the search of documents, in obtaining answers to questions, and in dialogue.

It is envisaged to acquire the following concepts:

- Natural Language Processing

- Word and Sentence Embedding

- Information Retrieval

- Acts of Dialogue

- Flows of Dialogue

- Modeling of Language

- Large Language Model

- Prompt Engineering

Program:

1. Introduction to Natural Language Processing;

2. Vector semantics;

3. Automatic Answer to Questions;

4. Dialogue Systems;

5. Language Models;

6. Evaluation.

The course unit aims to provide students with a comprehensive understanding of reliable and responsible AI, equipping them with critical thinking skills and enabling them to design and implement AI systems that are aligned with ethical principles, Fairness, transparency, accountability, privacy, security and sustainability.

1. Introduction to Responsible and Trusted AI;

2. Agency and Human Supervision;

3. Transparency and Interpretability/Explainability in AI Systems;

4. Diversity, Non-Discrimination and Justice in AI Systems;

5. Responsibility in AI Systems;

6. Robustness and Security in AI Systems;

7. Privacy and Data Governance in AI Systems;

8. Societal and Environmental Welfare in AI Systems;

9. Ethics and Morality in AI Systems;

10. Structures and Guidelines.

Understanding of the concepts of computational creativity, as well as its multidisciplinary framework.

To know and understand the main models and techniques that support current research in computational creativity, namely those from Artificial Intelligence.

Understand development techniques of computational creativity systems in some relevant application domains such as Design, Art, Image, Music and Sound, Poetry and Text in general.

1. Theories and models of creativity;

2. Formal characterization of Creativity;

3. Theories and models of computational creativity;

4. Evaluation of computational creativity: metrics and models;

5. Current research topics in computational creativity (e.g., concept creation, co-creativity, social aspects);

6. Computational techniques (e.g., evolutionary algorithms, neural networks, constraint programming);

7. Applications (e.g., in Design, Art and Image, Music and Sound, Poetry and Text).

The main objectives of the internship are:

- Design techniques and software development and systems for artificial intelligence;

- Realization of technological development projects;

- Contact with the preparation of projects in the business environment;

- Initiation of basic and applied research activities;

- Integration in the labor market;

- Preparation of a document with the internship proposal including the following aspects:

- Analysis of the state of the art;

- Justified choice of tools and methodologies to be used;

- Analysis of requirements regarding the theme to be developed;

- High level specification of the system to be developed/ work to be carried out including concrete objectives and schedule for the second half.

The internship/dissertation course aims to be a vehicle for consolidating, applying and integrating the knowledge acquired throughout the course.

This course corresponds to 42 ECTS, divided into two semesters of 12 ECTS + 30 ECTS.

The work may be developed at FCTUC or in some external entity, under the guidance of a teacher from FCTUC.

The final report can be written in Portuguese or English.

a) Have an overview of the communication cycle in science and the scientific and technological objects it produces;

b) Understanding changes in the way science is communicated;

c) Understand how research can benefit from Open Science;

d) Be able to determine the most appropriate way for publication in Open Access;

e) Understand the importance of research data and the issues of sharing and re-using;

f) Understand the contribution of open data to Citizen Science

g) Understand the main issues related to intellectual property;

h) Be able to design strategies to increase the visibility of research

i) Understand the limitations of conventional metrics and the role of new-generation metrics.

Program:

1. Science as a communication system;

2. Openness in the communication of science;

3. Research data;

4. The evaluation of research;

5. The dissemination of research.

Understand the concepts involved in entrepreneurship; develop analytical skills and ideas;

Apply the concepts involved in structuring a Business Plan;

Competence in organization and planning;

Ability to decide;

Understand the language of other experts;

Initiative and entrepreneurial spirit;

Ability to negotiate.

Program:

The Entrepreneurial Challenge and the Entrepreneur’s Profile

Cognitive Techniques: Creativity Innovation and Entrepreneurship: the "Knowledge Economy";

 "Create to Innovate": the First Steps for the Development of a Business Idea, SWOT analysis; Intellectual Property Protection; Business Opportunity Analysis; Value Proposition; Legal Aspects with the Creation of Companies; Sources of Financing for  "Creation of Own Employment",

Preparation of a business plan:

(1) Analysis and definition of the needs of the client and the market;

(2) Specification of goals;

(3) Feasibility studies and defence of the Business Plan.