PHD COLLOQUIA-3

The teaching is part of the course of study, providing a solid theoretical and practical basis necessary to face the challenges and skills required in the sector.
Through theoretical insights and practical exercises, students will be able to gain an in-depth understanding of fundamental concepts and develop application skills essential for their research career.

At the end of the course, students should be able to:

1. Understand the basic terminologies and concepts of biometric systems, such as biometric-modalities/traits, biometric features, biometric-templates, enrollment/identification/verification, error rates, and so on.
2. Identify the main challenges to adaptation of biometrics, such as privacy, spoofing, liveness detection, etc.
3. Survey, understand and explain different types of behavioral biometrics, such as keystroke, voice, gait, signature, mouse movements, eye tracking, etc.
4. Understand the data acquisition, feature extraction, matching methods for each modality, as well as their strengths and weaknesses and some use cases.
5. Understand and apply machine learning techniques commonly used for behavioral biometric recognition, e.g., supervised/unsupervised learning, dimensionality reduction, classification, deep learning.
6. Formalize the challenges and opportunities of applying machine learning to behavioral biometrics, such as data quality, data scarcity, data imbalance, concept-drift, and adversarial attacks.
7. Understand the designing of and conducting experiments to evaluate the performance, reliability, and usability of behavioral biometric systems.
8. Assess the experimental results and draw conclusions based on the findings.

I prerequisite per questo corso potrebbero includere:

1. Comprensione di base dei concetti di informatica e dei linguaggi di programmazione.
2. Conoscenza dei concetti fondamentali delle strutture dati e degli algoritmi.
3. Conoscenza della matematica, inclusa la probabilità e le statistiche.
4. Comprensione dei principi e delle tecniche di apprendimento automatic.
5. Conoscenza di base dell'elaborazione dei segnali digitali e del riconoscimento dei pattern.
6. Precedente esposizione ai sistemi biometrici o tecnologie correlate è utile ma non richiesta.
7. Accesso a un computer con strumenti software appropriati per l'analisi dei dati e gli esperimenti di apprendimento automatic.

Behavioral biometrics is a branch of biometrics that focuses on the identification and verification of individuals based on their behavioral patterns, such as keystroke dynamics, voice, gait, signature, and mouse movement. Behavioral biometrics can provide continuous and implicit authentication, as well as enhance the security and usability of traditional biometric systems. We aim to cover the following topics:


1. Introduction to Biometric Systems: This module aims at covering the basic terminologies and concepts, such as biometric-modalities/traits, biometric features, biometric-templates, enrollment/identification/verification, error rates, and so on. Further, the module will also introduce the main challenges to adaptation of biometrics, such as privacy, spoofing, liveness detection, etc.

2. Behavioral Biometric modalities: This module aims at surveying and explaining different types of behavioral biometrics, such as keystroke, voice, gait, signature, mouse movements, ey tracking, etc. Further, the module will describe the data acquisition, feature extraction, matching methods for each modality, as well as their strengths and weaknesses and some use cases.

3. Machine Learning for Behavioral Biometrics: This module aims at introducing machine learning techniques commonly used for behavioral biometric recognition, e.g., supervised/unsupervised learning, dimensionality reduction, classification, deep learning. Additionally, this module will discuss the challenges and opportunities of applying machine learning to behavioral biometrics, such as data quality, data scarcity, data imbalance, concept-drift, and adversarial attacks.

4. Evaluation of behavioral biometric systems: This module aims at explaining the designing of and conducting experiments to evaluate the performance, reliability, and usability of behavioral biometric systems. Additionally, it aims at covering the metrics, and methods of measuring the accuracy, robustness, stability, and user acceptance of behavioral biometrics along with the statistical tools and techniques for analyzing and comparing the obtained results.


5. Behavioral biometrics in mobile and online environments: This module aims at exploring the applications and challenges of behavioral biometrics in mobile and online settings, such as mobile banking, e-commerce, e-learning, and social media. It will discuss the factors, e.g., device variability, behavior variability, device/network latency, and user feedback, etc., that affect the feasibility and effectiveness of behavioral biometrics in these contexts. It also aims at presenting some case studies and best practices of applying behavioral biometrics in mobile and online environments.


6. Ethical and legal issues of behavioral biometrics: This module aims at examining the ethical and legal implications of using behavioral biometrics for authentication and identification purposes. Further, it aims at addressing the questions and concerns that arise from the collection, processing, storage, and sharing of behavioral biometric data, such as informed consent, data protection, data ownership, data access, data breach, data misuse, and data deletion. Furthermore, the module will review the emerging regulations and standards such as General Data Protection Regulation (GDPR), Biometric Information Privacy Act (BIPA) and the ISO/IEC 30137 series.

1. Behavioral Biometrics for Human Identification: Intelligent Applications: Intelligent Applications. Wang, Liang, and Xin Geng, eds. IGI Global, 2009.
2 Advances in biometrics: sensors, algorithms and systems. Ratha, Nalini Kanta, and Venu Govindaraju, eds. Springer Science & Business Media, 2007.

Assessment: Writing a Research Article in Behavioral Biometrics

In this assignment, students will demonstrate their understanding of behavioral biometrics by writing a research article. The task is to explore a specific topic within behavioral biometrics (such as gait recognition, keystroke dynamics, voice biometrics, or mouse movement analysis) and present the findings in a scholarly format.

Assignment Guidelines:

Topic Selection: Students should choose a relevant and interesting topic related to behavioral biometrics.
Literature Review: Students need to conduct thorough research by reviewing existing studies, scholarly articles, and research papers. They should understand the current state of the field, challenges, and recent advancements.
Article Structure:
Abstract: Summarize the study’s objectives, methods, and key findings.
Introduction: Provide context, research questions, and objectives.
Literature Review: Discuss relevant studies and theories.
Methodology: Describe the research design, data collection, and analysis methods.
Results: Present their findings, statistical analyses, and visual representations.
Discussion: Interpret results, compare with existing research, and propose implications.
Conclusion: Summarize key points and suggest future directions.
References: Properly cite all sources used.
Original Research or Review Article:
Students can choose between:
Original Research: Conduct their own experiments, surveys, or data analysis related to behavioral biometrics.
Review Article: Synthesize existing research, critically analyze it, and propose new insights.
Assessment Criteria:
submissions will be evaluated based on content, structure, critical thinking, writing style, and proper referencing.
Submission: The research article needs to be submitted by the specified deadline.

The teaching approach will be frontal, conducted face-to-face in the classroom setting. This means that the instructor will directly engage with the students during lectures, discussions, and other learning activities within the physical classroom environment.

English

written