72796 - Data Intensive Applications

Learning outcomes

At the end of the course the student is able to design and develop intelligent applications of corporate interest by processing structured and unsctructured data with modern data science methods and fundamental machine learning technologies.

Course contents

The course covers the fundamentals of data science and machine learning, ranging from introductory algorithms to neural networks for the development of AI applications in Python, as well as the basic concepts and applications of Large Language Models.

Machine learning algorithms will be used to develop models and applications capable of generating a wide range of predictions. Examples include forecasting stock market trends, predicting which products or services each customer is likely to purchase, estimating company sales, energy consumption, and property values, and assessing whether a bank loan is likely to be repaid.

Machine learning will also be applied to introductory natural language processing tasks, together with an introduction to the use of recent Large Language Models. These techniques will be used to classify user opinions and reviews of products and services published on social media and e-commerce platforms.

Finally, question-answering applications will be developed to provide user support through chatbots, along with semantic search engines and systems of this kind used in healthcare, for example to support the diagnosis of medical conditions or suggest how to manage symptoms described by patients in natural language.

COURSE CONTENTS, LECTURES AND LABORATORIES

course web site (sito web del corso https://virtuale.unibo.it/course/view.php?id=70598 )

The material view does not require the enrolment to the course (in this case access the course web site selecting "spontaneous registration")

Readings/Bibliography

• Materials and bibliographic references supplied by the teacher

Suggested course book

• Data Science from Scratch, Joel Grus, O’Reilly Media, 2019 early edition freely available (previous edition 2015). Italian edition 2021, Data Science con Python. Dai Fondamenti al Machine Learning. Scientific Editor Gianluca Moro. Published by EGEA.

Teaching methods

Lectures are followed by aided laboratory practicals on real case studies of machine learning and data science for the development of artificial intelligence applications in several industrial and social domains.

Assessment methods

The assessment consists of a laboratory project, which may be completed either individually or in groups, followed by an individual oral discussion. Students may independently select an AI topic of interest for their project (examples of completed projects are available on the course’s virtual learning platform).

The final grade is calculated as the arithmetic mean of the project grade and the oral discussion grade, according to the following criteria:

The excellence range, from 30 to 30 with honours, applies to rigorous and well-argued projects that demonstrate elements of originality, supported by a clear and accurate discussion of the topics addressed.

The good range, from 27 to 29, applies to solid and well-structured projects accompanied by a clear and effective oral discussion, demonstrating in-depth knowledge but no particularly original elements.

The satisfactory range, from 18 to 26, includes work that is generally adequate but less thorough or precise, with a straightforward presentation and technical terminology that is not always fully appropriate.

The unsatisfactory range, up to 17, indicates significant shortcomings in content, methodology, and argumentative skills, preventing the achievement of the minimum expected learning outcomes.

Teaching tools

• slides of lectures and laboratory practicals available in advance
• python and jupyter open source technologies fully available in the school laboratories
• aided laboratory activities linked also to open cloud computing such as colab or binder to perform them online remotely
• example of a laboratory practical in python available with google colab

Link to the lessons of 2022/23. The contents of machine learning for the 2023/24 will be updated according to the most recent solution of the academic and industrial communities of AI:

https://virtuale.unibo.it/course/view.php?id=37961

Office hours

See the website of Gianluca Moro