Summer School: How to Model Human Behaviour Using Artificial Intelligence?

The challenges associated with the twin digital and green transition require not only technological development but also social development. Up to now, in many sectors, this social development has not received enough attention, with the result that some products or policies have failed. One way to tackle this issue is by developing behavior models to study the social development.

This course is aligned with the United Nations' 9th Sustainable Development Goal “Build resilient infrastructure, promote sustainable industrialization and foster innovation”. The course allows students to get to know the key points that explain behavior and decision-making and include them in mathematical models.

Key Information

Who can attend?

• This course is mainly addressed to master or PhD students that are interested in learning different ways of modeling human behaviour.
• Other potential interested participants are junior modellers interested in learning how to process high frequency time series or causal models in order to improve their modeling toolboxes.
• Finally, technical staff from different public authorities or utilities could also be interested in learning these techniques.
• Requirements: Previous knowledge of programming language.
• B2 level of English is recommended.

What will you learn?

The objective of this course is to train students in different behavior modeling methodologies so that they can autonomously develop such models.

In addition, the course aims to develop generic and specific competences:

Generic

• Participation in multidisciplinary activities
• Development of group activities in general and especially in online group activities
• Oral and written communication of scientific results

Specific

• Using quantitative modeling techniques to segment behaviors
• Use of qualitative modeling techniques to explain behaviors
• Development of analysis skills that allow evidence-based decision making

Programme

Methods to extract behaviour patents from high frequency time series

Introduction to machine learning

• Data visualization: get familiar with the information contained in the datasets
• Data cleaning: imputation, extension, outlier removal 33Feature extraction: one set fits for all vs tailored features
• Data segmentation: clusterization following different algorithms
• Quantitative assessment (validation) of the model
• Qualitative results assessment: taxonomy creation
• Knowledge extraction: household personae

Causal models for behaviour modelling

• Introduction to Causal Models
• Writing a causal diagram
• Adjusting a model using Do-Why (selecting a potential outcome)
• Measuring the causal effect
• Refuting an estimand
• Qualitative methods
• How to set up the intrinsic and extrinsic variables in your model?
• How to build from scratch a causal diagram?

Lecturers

Armando Aguayo Mendoza. Researcher at DeustoTech at the University of Deusto and research professor of TESCo;
Cruz Enrique Borges Hernández. Coordinator of the course, researcher at DeustoTech, at the University of Deusto;
Diego Casado Mansilla. Researcher at MOREIlab and DeustoTech and professor at the University of Deusto;
Carlos Quesada Granja. Researcher at DeustoTech at University of Deusto.

Your Learning Path

The methodology of the Deusto Summer School courses is based on University of Deusto Learning Model (MAUD), adapted to distance learning. You will participate in an opening and a closing session via the University of Deusto's live video conferencing tool. During the course, you will have the opportunity to do personal learning with the use of various learning materials.

This model integrates the stages of contextualization, reflective observation,  conceptualization, active experimentation and assessment. A methodology that enables autonomous learning and encourages the development of knowledge, skills, attitudes, capacities and values. This methodology guarantees a complete learning experience, so that every student can apply the acquired knowledge in the future.

To implement this methodology the following teaching resources are used:

• Combination of live video conference classes with online learning, through resources and activities on the university’s learning platform;
• Use of learning resources in different formats for the presentation of theoretical content;
• Different learning activities to verify the acquisition of both theoretical and practical knowledge and the acquisition of skills;
• Follow up and assessment.

For an overview on the course, download the brochure below! | Do you want to learn more? Visit the University of Deusto Summer School Page!