Τraining modules for SMEs - Module 1 "Introduction to Artificial Intelligence and High-Performance Computing"

Thursday 13 Mar 2025, 11:00 → 13:00 Europe/Athens

GRNET announced, in the context of SmartAttica EDIH (European Digital Innovation Hub), the 1st Module of Τraining modules for SMEs with the subject "Introduction to Artificial Intelligence and High-Performance Computing", that took place online on March 13th, 2025.   

Date: March 13th, 2025, at 11:00 EET  

Location: Online via Zoom

Presentation Languages: Greek

Audience:

• SME owners and managers who want to learn how AI and HPC can aid the digital transformation of their company
• Professionals in software development, data science, engineering, and related fields.
• Industry practitioners interested in AI and HPC applications.

Description: This seminar provided a comprehensive introduction to Artificial Intelligence (AI) and High-Performance Computing (HPC). Participants explored the fundamentals of AI, including its overarching goals, historical milestones, and societal impacts, before examining the integral role of HPC in accelerating AI innovations. The course also covered ways to access HPC resources, emphasizing practical applications, parallel processing, and the synergy between AI and HPC in solving complex real-world problems.

Learning Objectives:

• Understand the historical evolution and current scope of AI and its subfields.
• Grasp the core principles of HPC and its technological drivers.
• Analyze the role of HPC in enhancing AI capabilities and applications.
• Learn about the process and requirements for accessing HPC resources.
• Explore real-world applications and future advancements at the intersection of AI and HPC.

Prerequisites:

• Basic familiarity with technology concepts, such as using computers and software applications.
• Understanding of general business processes and challenges.
• Interest in how new technologies can improve efficiency and innovation.

Note: Please enter your institutional/corporate email when registering.

 

The project is co-funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Commission. Neither the European Union nor the granting authority can be held responsible for them.

 

Read the Privacy Statement here.pdf

Recorded Material

11:00 → 11:10 Welcome - The Smart Attica EDIH and the training events

Speaker: Mr Ilias Hatzakis (GRNET)

11:10 → 11:50 Introduction to Artificial Intelligence

1 Introduction
1.1 Artificial Intelligence (AI)
1.2 Data Mining and Big Data
1.3 History of Artificial Intelligence
1.3.1 Turing’s Test (1950)
1.3.2 The Formal Beginning
1.3.3 Early Success
1.3.4 1st AI Winter
1.3.5 Revival
1.3.6 2nd AI Winter
1.3.7 The Rise of Machine Learning
1.3.8 AI in the 21st Century
1.3.9 Current AI Technology
1.4 Supercomputers & AI

2 Introduction to Machine Learning
2.1 Traditional Programming
2.2 Machine Learning Algorithms
2.3 Input - Output
2.4 Fields of ML/AI
2.5 Tabular datasets
2.6 Computer Vision

3 Mathematical Models
3.1 A fundamental problem
3.2 ML Models
3.3 Artificial Neural Networks
3.4 Training a Model
3.5 Newton’s Method for Root Finding
3.6 Newton’s Method for Optimization
3.7 In Multiple Dimensions
3.8 Fundamental inefficiencies

4 Introduction to Large Language Models
4.1 Gödel’s incompleteness theorem & Turing’s Test
4.2 2024 LLMs
4.3 Language issues
4.4 Semiotics
4.5 Word representation
4.6 One-hot encoding
4.7 Representation of sentences
4.8 Vocabulary with N words
4.9 Semantic Limitations of One-hot Encoding
4.10 Word Embeddings and Distributional Semantics
4.11 Training Transformers with Embedded Word Representations
4.12 Training and Fine-tuning
4.13 Sequences of words
4.14 Train Dataset
4.15 The Transformer model architecture
4.16 Big Data
4.17 Model and Token Size
4.18 EuroHPC JU Access Call for AI and Data-Intensive Applications

5 Applications
5.1 Generative AI
5.2 Large Language Models
5.3 Computer Vision
5.4 Recommender Systems
5.4.1 Timeseries Analysis and Predictions
5.5 Black-Box Optimization
5.6 Analysis of Scientific Literature
5.7 More

6 Introduction to Prompt Engineering
6.1 Definition and significance
6.2 Overview of language models
6.3 Fundamentals of Crafting Effective Prompts
6.4 Basic principles of prompt design
6.5 The role of prompts in steering the model’s responses
6.6 Common pitfalls
6.7 Ethical Considerations and Best Practices
6.8 Technical Aspects of Prompt Engineering

Speaker: Dr Nikolaos Bakas (GRNET)

11:50 → 12:05 Break

12:05 → 13:00 Introduction to High-Performance Computing

1 History of HPC
1.1 The Early Days: 1960s
1.2 Vector Processors and the Rise of Cray: 1970s-1980s
1.3 Parallel Processing: 1990s
1.4 The Advent of Clustering: Late 1990s and 2000s
1.5 Petaflops and Beyond: 2010s
1.6 Exascale Computing: The Next Frontier
1.7 Moore’s Law
1.8 AI and HPC

2 General Concepts of High Performance Computing (HPC)
2.1 Definition of HPC
2.2 Importance of HPC
2.3 Components of an HPC System
2.4 Parallel Computing
2.4.1 Instructions
2.4.2 CPUs, Cores and Threads
2.4.3 Threads: Software vs Hardware
2.4.4 Tasks, Threads, and Cores
2.5 CPUs, GPUs and Nodes
2.5.1 Sample Slurm Script
2.6 Types of Parallel Computing

3 Scaling
3.1 Why Parallelization Matters
3.2 Scalability in HPC systems
3.3 Speedup
3.4 Parallelization Efficiency
3.5 Scaling Tests
3.6 Strong Scaling
3.7 Amdahl’s Law
3.8 Weak Scaling
3.9 Gustafson’s Law
3.10 Performance Metrics in HPC

4 Programming Models in HPC
4.1 OpenMP (Open Multi-Processing)
4.2 MPI (Message Passing Interface)
4.3 GPUs (Graphics Processing Units)
4.4 CUDA (Compute Unified Device Architecture)

5 State of the art machines
5.1 The Top 500 list
5.1.1 Exponential Growth
5.2 Top 8 European Supercomputers
5.3 ARIS – HPC Infrastructure in Greece
5.4 Daedalus - EuroHPC supercomputer in Greece

6 Apply for Access at EuroHPC JU
6.1 EuroHPC JU Benchmark Access
6.2 EuroHPC JU Development Access
6.3 EuroHPC JU Regular Access
6.4 EuroHPC JU Extreme Access
6.5 EuroHPC JU Access Call for AI and Data-Intensive Applications
6.6 Frequently Asked Questions (FAQ)
6.7 Indicative Application
6.7.1 The project
6.7.2 Research Fields
6.7.3 Societal impact
6.7.4 CPU Partition
6.7.5 Input / Ouput
6.7.6 GPU Partition
6.7.7 Code Details
6.7.8 Scalability & Performance
6.7.9 Optimization
6.7.10 Performance
6.7.11 Data Consent
6.8 Resources

Speaker: Dr Nikolaos Bakas (GRNET)