AI030: AI Implementation Advanced

Agenda:

• Module 1: Introduction to Data Visualization
  • Importance of visualization for data analysis
  • Visualization libraries in Python: matplotlib, seaborn, plotly
  • Different visualization types: Scatter Plots, Bar Charts, Violin Plots, etc.
  • Fundamentals of visualization: Dos and Don'ts
  • Creating compelling visualizations
  • Exercises and examples for applying learned concepts
  • Introduction to interactive visualizations
  • Best practices and tips for selecting the right visualization type
  • Introduction to data visualization
  • Importance of visualization for data analysis
  • Visualization libraries in Python: matplotlib, seaborn, plotly
  • Different visualization types: Scatter Plots, Bar Charts, Violin Plots, etc.
  • Fundamentals of visualization: Dos and Don'ts
  • Creating compelling visualizations
  • Exercises and examples for applying learned concepts
  • Introduction to interactive visualizations
  • Best practices and tips for selecting the right visualization type

• Module 2: Introduction to Recommendation Systems
  • Introduction to recommender systems and their importance
  • Comparison of different recommender systems: Collaborative Filtering, Content-based Filtering, Hybrid Recommender Systems, Semantic Recommender Systems
  • Data sources, data collection and processing: Insights into common data sources and techniques for data collection, data preparation and analysis
  • Evaluation metrics for recommender systems: Presentation of different metrics for assessing recommendation quality
  • Practical exercises: Building a simple recommender system with Python
  • Summary and final discussion: How can the learned concepts be transferred to one's own context?

• Module 3: Introduction to Prompt Engineering for Developers
• Introduction to prompt engineering and terminology: LLMs, prompts, context, etc.
• Fundamentals of different prompting techniques: Best practices, examples and exercises
• Application scenarios: Discussion of different use cases and their implementation
• Possibilities and limitations of prompts: Identifying opportunities and boundaries through examples
• Integration of LLMs: Basic techniques for integrating LLMs into own projects
• Final discussion: Experiences, challenges and future developments

• Module 4: Introduction to Large Language Models and their Integration
  • Introduction to Large Language Models (LLMs)
  • Functionality of LLMs: sequential generation, tokens, context, implicit and explicit knowledge
  • Fundamentals of prompt engineering
  • Introduction to embeddings as central building block
  • Techniques of knowledge injection and introduction to vector databases
  • Presentation of different LLM providers and integration via API
  • Many practical exercises for applying learned concepts
  • Final discussion: How can the learned concepts be applied in one's own context?

• Module 5: Extraction and Structuring of Data with AI
  • Introduction to the problem statement: Necessity of an interface between unstructured data and structured databases and processes
  • Introduction to tools and libraries: Python, NLP libraries, LLMs
  • Techniques for information extraction: Named Entity Recognition, Deep Learning, LLMs
  • Output parsing, function calling and tool usage: Application of learned techniques for extracting and integrating information
  • Many practical exercises: Application of learned techniques to realistic examples and data
  • Final discussion: Experiences and challenges / How can the learned concepts be transferred to one's own context

• Module 6: Extraction and Transformation of Information with AI
  • Introduction and basic architecture
  • Usage of large language models (completion, summarization etc.) and Python
  • Reading data from different formats and chunking strategies
  • Text compression techniques
  • Text compilation techniques: Transforming information into a format like a report
  • Export to desired format like PDF or docx
  • Discussion about use cases and strategies for effective deployment of AI-based information transformation in practice

• Module 7: From Automatic Transcription to Structured Information with AI
  • Introduction to automatic audio transcription and underlying concepts
  • Usage of API solutions for transcription of audio files
  • Processing of transcripts and exploration of different use cases, such as
    • Classification of customer feedback
    • Data extraction from transcribed text
    • Translation of text into other languages
  • Practical exercises for implementing learned concepts in own projects
  • Discussion about possible future developments and challenges in the field of automatic transcription

• Certificate Examination

Objectives:

• Learn fundamentals of data visualization, explore various visualization types, utilize visualization libraries in Python, present complex concepts clearly.
• Understand fundamentals of recommender systems, explore different types of recommender systems, master data sources and processing for recommender systems, apply evaluation metrics for recommender systems.
• Understanding prompt engineering, learning effective prompting techniques, recognizing potential and limitations, integration of LLMs.
• Information extraction, Named Entity Recognition, Deep Learning, LLMs, Output Parsing, Function Calling, Tool Usage.
• Learn fundamental architecture and requirements, utilize large language models and Python, read data from different formats, prompting techniques for text compression and text compilation, export to desired formats.
• Convert audio-based speech to text, process transcripts, explore use cases, utilize APIs.

Furthermore, the course provides a solid foundation for additional advanced courses, e.g.:
AI050 AI Security Specialist

Target audience:

• Developers
• IT Professionals
• AI Officers
• AI Auditors

Prerequisites:

• AI020 AI Implementation Basics or equivalent prerequisites

Description:

The certificate program AI030 AI Implementation Advanced trains you to become an AI expert for data and information processing.

The ability to visualize data is of great importance for many areas of data analysis and machine learning. Often it is difficult to identify correlations in data without displaying them visually. The program teaches the fundamentals of appealing data visualization as well as its significance for data analysis.

Furthermore, the program covers fundamentals of prompt engineering explicitly for technical specialists and developers. The objective is to understand how effective prompts are structured and enriched with data to implement various use cases. Participants experience the potential of the technology and learn basic techniques for integration of LLMs (Large Language Models).

To make data-driven decisions, information must be converted into structured data. The program teaches techniques for data extraction and transformation, presenting both established methods of Natural Language Processing (NLP) as well as new possibilities through LLMs.

Participants also learn how they can capture more relevant information in less time using large language models. This involves not only the generation of summaries, but also the further utilization of the obtained information e.g. for automated creation of documents in different target formats. The course covers techniques for knowledge extraction, text compression and compilation as well as export to desired formats such as PDF and docx.

The program also introduces you to the concepts of automated audio transcription and shows how Artificial Intelligence can be used to convert natural language into structured information.

Exam: