Tekoälyosaaja

You will learn to use wide variety of different AI tools to increase productivity and learning, and stay up-to-date in the latest developments of generative AIs.

During the course you will learn:

- prompt engineering or how to write efficient prompts, which get the job done
- how to use image generation tools to create amazing looking art and photorealistic image
- how to utilize image-to-image generation
- how to use AI to create presentation materials and complete videos for you.
- how to use AI to compose, rewrite and analyze any sort of text
- how to use AI to generate ideas and act as an expert to help you
- how to use AI tools perform programming tasks (code generation, finding & fixing bugs, documentation,
testing, rewrite/refactor code) for you.

You will get to know a lot of different AI tools and become efficient in using them.

Most of the learning comes from numerous examples and practical assignments. We will cover the use of wide variety of different AI tools.

Student groups can choose what assignments they will do as long as they complete the required
minimum amount. The assignments will cover a wide range of topics from generating text, teaching
materials and program code to generating images and video.

There are also some compulsory individual assignments, such as the final review.

There is minimal amount of theory: 99 % practice, 1 % theory. There are no mathematics to learn.

During this course you will learn:

- how to use APIs provided by major AI companies, such as OpenAI (ChatGPT and Dall-E 3) and integrate AI as part of your web site or other applications.
- how to use AI to automate tasks, including running programs on your computer. Learn how to utilize AI to scrape information from the web and social media sites.
- how to create your own custom bots and use AI APIs in a multimodal way e.g. combine text and image in creative ways.
- how to install and use AI tools on your local computer. NOTE: this will require quite powerful PC and preferably at least 8 Gigabytes on video memory (VRAM) and NVIDIA display adapter. The more VRAM, the better.

Training your own AI models can be prohibitely expensive and time consuming. The use of AI APIs and existing open source LLMs and image generators (such as Stable Diffusion) are the most efficient ways to bring AI as part of your own tool chain and integrate as part your own applications.

The main objective is to find how to utilize existing AI APIs and how to install AI tools on your local computer or on the cloud. Lots of practical examples and services.

There will be several programming assignments and some of them might require the student to pay for the use of AI APIs.

Most programming will happen in Python, but other programming languages (such as JavaScript) might be used.

Many APIs will be pay to use, but the cost are likely to be moderate (e.g. with about 10 euros you can by enough credits for short experimentation). The student pays for the use of the APIs.

During the AI project course you will implement and document a project, which utilizes AI heavily. After completing the course, you will be able to identify a real-world problem and propose a solution for it using AI. You know to use tools and methods to design the solution. You implement the solution using a systematic approach. You evaluate the solution in different scenarios, considering possible problems and their mitigation. You present your solution to different stakeholders using appropriate language.

As a group you will implement and document a project, which utilizes AI heavily. The end result can be an entire book, video, game, web site, new mobile, desktop or web application, but it cannot be the result of a single prompt. The project could also be the use of existing AI APIs with robotics, IoT or automation. The project could also be an art installation or a tech prototype for a company.

The project is divided into several phases: forming the group and planning, prototype, beta and final version. The students will receive feedback from each phase.

You will familiarize yourself with Data processing with Python. It is expected that you know the basics of programming with some programming language.

During this course data is analyzed systematically starting with the single variables, then proceeding
into two variable case and finally performing statistical tests for getting insight on possible dependencies between variables.

You will become familiar with many Python libraries that proof to be handy for many different kind of tasks, not just for data analysis.

The following topics will be covered:

• Programming structures characteristic of Python
• Visualizing data with Matplotlib
• Using the Pandas library for data processing
• Principles of visual analytics
• Reading and writing data
• Understanding data structures and their basic properties
• Selecting, indexing, grouping, and transforming data
• Working with missing data and duplicates
• Conducting fundamental exploratory analysis of data
• Performing statistical tests to evaluate dependency between variables
• Completing practice work to integrate the points above

This course introduces machine learning to make predictions based on data. You will get to know several supervised and unsupervised learning algorithms. You will use the Python programming language and its several different libraries.

After completing this course you:

- have a good understanding of key machine learning concepts and methods
- are able to apply different machine learning methods to practical problems
- are able to create and deploy simple machine learning solutions

Topics covered:

• Supervised learning, knn algorithm as an example
• Unsupervised learning, k-means algorithm as an example
• Linear regression and correlation
• Underfitting and Overfitting
• Decision Trees
• Support Vector Machines
• Dimensionality Reduction
• Basic topic detection, regular expressions, Natural Language Processing (NLP)
• Sentiment analysis

The course is completed by doing a few practice projects in small groups. Before each practice project, many things related to data processing and Python libraries are covered in lectures and examples are provided.

The objective of the course is to provide you with deep understanding of neural networks together with hands-on experience. First the topics listed below are studied and then a practice project is done and submitted for evaluation.

Topics covered:

• Python library for computer vision
• Artificial Neural Networks
• Gradient Descent algorithm
• Convolutional Neural Networks
• Model training using own data
• Transfer learning

The topics above are studied through lecture slides and literature but also an extensive code base in Python language is provided. These codes are then to be modified by the students for applying them the given dataset which is part of the practice work of the course.