Resources
CSAI Resources for Creatives
This collection of sites, tools, coding libraries, reference sources and additional material have been selected by CSAI to help you get started or find a solution to your current need. An indication of the skill level required and the type of resource is provided for each entry. We’re interested in seeing what results you achieve! Enjoy!
Coding Tools and Libraries for Creatives
ML5 provides an approachable browser-based tool that makes machine learning approachable for a broad audience of artists, creative coders, and students. ML5 allows access to the popular TensorFlow library without needing to be a data scientist. Languages: Javascript. Coding Skills: Minimal coding skills required. User Level: Beginners-Intermediate.
Processing
Processing is a flexible software sketchbook and a coding language that emphasizes usage in the visual arts. Used by thousands of students, artists, designers, researchers, and hobbyists for learning and prototyping. Languages: Python and Javascript. Coding Skills: Coding skills required. User Level: Intermediate-Advanced
Magenta
Open source project that allows users to explore machine learning for artwork and music generation. Languages: Python and Javascript. Coding Skills: Minimal coding skills required. User Level: Beginners-Intermediate
TensorFlow JS
JavaScript based version of the TensorFlow library, that allows creatives to create new AI and ML models in an easier way than the full TensorFlow library. Languages: Javascript. Coding Skills: Coding skills required. User Level: Intermediate-Advanced.
PyTorch
Python based AI and ML development library, that can be used to create new AI and ML models. Languages: Python. Coding Skills: Coding skills required. User Level: Intermediate-Advanced.
SKLearn
Python based AI and ML library that adds a lot of creative functionality, especially when needing to do anything that has to do with statistics. Languages: Python. Coding Skills: Coding skills required. User Level: Intermediate-Advanced.
Photo Editing with AI
Deep Dream Based for AI Dreamlike Effects
Google Deep Dream
GitHub repository for Google Deep Dream. Languages: Python. Coding Skills: Intermediate Coding skills required. User Level: Intermediate-Advanced
Deep Dream Generator
Create stylized images and Deep Dream generated images using this pre-packaged and online accessible tool based on Google Deep Dream. Languages: None. Coding Skills: None required. User Level: Beginners-Intermediate.
Deep Dream Video
Extends the Google Deep Dream library to create videos. Languages: Python. Coding Skills: Intermediate Coding skills required. User Level: Intermediate-Advanced
Neural Style Transfer for Artistic Photo Interpretations
DeepArt.io
Apply different art styles easily or create artistic portraits of yourself from a photo. Uses neural style transfer. Languages: None. Coding Skills: None required. User Level: Beginners-Intermediate.
Apply artistic styles to photos using this iOS app, including a wide variety of filters. Uses neural style transfer. Languages: None. Coding Skills: None required. User Level: Beginners-Intermediate.
GoArt
Turn photos into artistic painting interpretations using AI for desktop, Android and iOS. Uses neural style transfer. Languages: None. Coding Skills: None required. User Level: Beginners-Intermediate.
Deep Fakes
Deep Angel
Apply artistic styles to photos using this iOS app, including a wide variety of filters. Uses neural style transfer. Languages: None. Coding Skills: None required. User Level: Beginners-Intermediate.
Generative Art
Art Breeder
Image generation tools that uses GANs to blend images together to form entirely new ones and collaborate with others to explore the capabilities of GANs. Previously known as GANBreeder. Uses StyleGAN and BigGAN. Languages: None. Coding Skills: None required. User Level: Beginners-Intermediate.
Image generation tools that uses GANs to blend images together to form entirely new ones and collaborate with others to explore the capabilities of GANs. Previously known as GANBreeder. Uses StyleGAN and BigGAN. Languages: None. Coding Skills: None required. User Level: Beginners-Intermediate.
Video Editing with AI
Runway ML
Image generation tools that uses GANs to blend images together to form entirely new ones and collaborate with others to explore the capabilities of GANs. Previously known as GANBreeder. Uses StyleGAN and BigGAN. Languages: None. Coding Skills: None required. User Level: Beginners-Intermediate.
AI Painting
AI Painter
AI painting creation tool from photos. Uses neural style transfer. Languages: None. Coding Skills: None required. User Level: Beginners.
Quick, Draw!
Draw on your browser and play a game with AI, where a neural net tries to guess what you’re drawing as you’re sketching. Languages: None. Coding Skills: None required. User Level: Beginners.
Neural Net Sketching
Draw sketches with a neural network in your browser, trained on millions of examples from the Quick, Draw! game. Uses RNNs. Languages: None. Coding Skills: None required. User Level: Beginners.
Use AI to create a computer generated cartoon drawing from your photo. Languages: None. Coding Skills: None required. User Level: Beginners.
Handwriting, Calligraphy and Typography with AI
Handwriting with a neural net
Play around with a neural net that analyses and generates handwriting based on your writing style. Languages: JavaScript. Coding Skills: Minimal coding skills. User Level: Beginners-Intermediate.
Use deep learning to generate new font combinations that match. Languages: JavaScript. Coding Skills: Minimal coding skills. User Level: Beginners-Intermediate.
Explore and find graphical relationships between different font types. Languages: None. Coding Skills: None required. User Level: Beginners.
Voice Speech
Speech Scribbler
Speak a description of a scene and have a neural net sketch it out on your browser. Languages: None. Coding Skills: None required. User Level: Beginners.
Real Time Voice Cloning
Use deep learning to generate new font combinations that match. Languages: JavaScript. Coding Skills: Minimal coding skills. User Level: Beginners-Intermediate.
Music and Audio
Speech Scribbler
Speak a description of a scene and have a neural net sketch it out on your browser. Languages: None. Coding Skills: None required. User Level: Beginners.
Magenta Studio
AI powered music plugins built using Magenta’s open source tools and models. Languages: Python and JavaScript. Coding Skills: Beginner-Intermediate. User Level: Beginner-Intermediate.
Make music with a neural network trained on MIDI sound samples. Languages: Python and JavaScript. Coding Skills: Beginner-Intermediate. User Level: Beginner-Intermediate.
NSynth Super
Use AI to create new sounds and instruments. Comes with custom hardware designed for AI music. Languages: C++ and Python. Coding Skills: Intermediate-Advanced. User Level: Intermediate-Advanced.
MuseNet
Generate 4-minute musical compositions with ten different instruments, combining styles from country to Mozart to the Beatles. Uses a GPT-2 variant. GitHub code available [https://github.com/AbhilashPal/MuseNet]. Languages: Python. Coding Skills: Intermediate. User Level: Intermediate-Advanced.
Dance and Performance Art
Body, Movement and Language
A collection of experiments between Google and renowned choreographer Bill T. Jones using AI to explore the creative possibilities of using AI within performance art. Videos and experiments do not require coding skills. Languages: Javascript and Tensorflow. Coding Skills: Intermediate. User Level: Intermediate-Advanced
AI Choreographer
Generate dance motion in 3D together with a dance movement dataset, with AI doing pose estimation and generation. Languages: Python. Coding Skills: Intermediate-Advanced. User Level: Intermediate-Advanced.
Up Coming Events
Explained Terms
The world of science and AI and art comes with its unique terminology and jargon that may be difficult for people in different areas to understand. We have asked our Scientific Director, Angelo Dalli and our Artistic Director, Selina Scerri, to provide short explanations of common terms. This list will be constantly updated over time based on community feedback and as more material is added.
Deep Fake
Synthetically produced or manually doctored images that replace people or objects with other people or objects. AI has produced new techniques for such fake images, video and audio that surpass the quality of old manual techniques (sometimes referred to as Photoshopping) by so much that they have become increasingly indistinguishable from reality.
General Adversarial Network (GAN)
GANs are a type of neural network architecture where two networks are used: the generator, which creates candidate content based on some training data (for example, paintings from a particular artist, or photos of real human faces), and the discriminator (or critic) that tries to determine whether the generated candidates are real content or not. GANs have proven highly successful in creative applications due to their ability to generate realistic looking content relatively easily.
https://en.wikipedia.org/wiki/Generative_adversarial_network
Paradeiol
The tendency to see patterns or objects where none exist, for example, seeing animals or faces in clouds. AI can also exhibit pareidolia due to a phenomenon called overfitting, in which the AI tends to generalise patterns excessively, leading it to raise false alarms too quickly and misclassify content by being too eager to match patterns that do not actually exist.
Singularity
The Singularity is a hypothetical future turning point in human civilization when technological growth becomes uncontrollable and irreversible, heralded by the rise of artificial sentient beings, also referred to as Artificial General Intelligence (AGI). The term has its origins with John von Neumann, who invented modern computing architectures, and has been popularised by various science fiction authors, futurists and philosophers.
Terrain Relative Navigation (TRN)
TRN is a vision based system for navigation used in autonomous systems. Originally developed for use in spacecraft navigation and adopted by some automated car manufacturers, most notably Tesla, TRN allows robots to move around in unknown environments without needing a precise map of their surroundings.
CSAI Recommended Newsletters
AI is constantly changing with new publications coming out all the time and new advances made so rapidly that it can be difficult to keep up with new creations. The newsletters below are recommended sources of information and we also invite you to follow our Twitter for timely updates: @csai_institute
AI Newsletter
Description
Link
The Algorithm
MIT Technology Review update with an in-depth section.
KDnuggets News
KDnuggets is an established blog that focuses on business analytics, big data, data mining, data science, and machine learning.
Recent Posts
Emerging AI Trends for 2025
3 March 2025
AI-POWERED KINETIC ART
17 February 2025
