Phd in Physics

Big Data Science and Machine Learning

Big Data Science and Machine Learning

Academic year 2022/2023

In particular I suggest to get familiar with Jupyter notebooks, numpy and pandas before the course starts. No expert knowledge is required, but doing a couple of tutorials on these topics (easily found on the web) is highly recommended.

Sommario del corso

• Course objectives
• Results of learning outcomes
• Program
• Course delivery
• Learning assessment methods
• Suggested readings and bibliography
• Notes
• Class schedule
• Teaching tools
• Course card

Course objectives

Data science is one of the fastest growing fields of information technology, with wide applications in key sectors such as research, industry, public administration. The course will cover the definition of big data and the basic techniques to store, handle and process them. Machine Learning (ML) and Deep Learning (DL) algorithms will be briefly introduced. We will focus on the technical implementation of different ML algorithms, focusing on the parallelisation aspects and the deployment on distributed resources and different architectures (CPUs, FPGAs, GPUs). A basic introduction to the current computer architecture will be given, with a focus on parallel computing paradigms aimed at the exploitation of the full potential of parallel architectures. An overview of the fundamental OpenMP and MPI coding patterns is covered during hands-on sessions.

Results of learning outcomes

KNOWLEDGE AND UNDERSTANDING

Fundamental concepts of:

- big data science

- Machine Learning and Deep Learning

- computer architectures and distributed systems

- parallel programming with OpenMp

APPLYING KNOWLEDGE AND UNDERSTANDING
Ability to:

- implement various machine learning model architectures and metrics

- use set of machine learning libraries

- faster code execution by parallelization of tasks, avoiding race conditions

Program

- Introduction to big data science
- The big data pipeline: state-of-the-art tools and technologies
- ML and DL methods: supervised and unsupervised training, neural network models
- Introduction to computer architecture and parallel computing patterns
- Initiation to OpenMP and MPI
- Parallelisation of ML algorithms on distributed resources
- Beyond CPUs: ML applications on distributed architectures, GPUs, FPGAs

Course delivery

The course will be held in person, and it will not be possible to attend remotely. To pass the course you need to follow 80% of the lessons and pass the final test.

Learning assessment methods

Practical test

The exam will be on June 22nd (but can be turned in up to June 30th)

Chen, M., Mao, S. & Liu, Y. Mobile Netw Appl (2014) 19: 171. https://doi.org/10.1007/s11036-013-0489-0

Yao, Yuanshun & Xiao, Zhujun & Wang, Bolun & Viswanath, Bimal & Zheng, Haitao & Y. Zhao, Ben. (2017). Complexity vs. performance: empirical analysis of machine learning as a service. 384-397. 10.1145/3131365.3131372

Notes

Students wishing to take this course must register!

Class schedule

• Teaching material
• Exams and finals
• Class schedule
• Go to Moodle
• Register for the course
• Enrolled students
• Invia email agli studenti registrati
• Export table