Hello, I'm Gaston Gautreneau, a cybersecurity and data protection engineer and I made this page to share a discover made during august 2022.
A revolution at the crossroads of art and machine learning. A technology that raises many questions, especially ethical ones. A wonderful discovery for me at the creative level and whose only limit is the imagination.
Welcome on my journey around Artificial Intelligence and Art + Some real photographs at the bend of a museum or an exhibition.
More specifically on A.I. using text-to-image diffusion model to generate pictures based on interpretation of the "prompt" given to the A.I.. (exemple of a prompt in Midjourney : Japanese landscape ::650, japanese plants ::120, Fujiyama in the background ::420, at dawn ::380, mist ::100, photorealistic, 4k, God rays, Highly detailed, Octane Render --ar 16:9 --s 1500)
This website concentrate on DALL·E 2, Midjourney v3/v4/V5/V6 & Disco Diffusion v5.6.
Please find a few links below to help you discover this world too :
- DALL·E 2 by OpenAI is a new AI system that can create realistic images and art from a description in natural language. Here is an article by OpenAI on DALL·E: Creating Images from Text.
- GPT-3.5 is "a powerful artificial intelligence system that can generate text. In this paper, we explore GPT-3's ability to write about itself. We find that GPT-3 can generate clear and concise descriptions of its own capabilities and features. This is a significant advance over previous systems, which have often struggled to produce coherent text about themselves".
- OpenAI trained a model called ChatGPT which interacts in a conversational way. The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests. ChatGPT is a sibling model to InstructGPT, which is trained to follow an instruction in a prompt and provide a detailed response.
- An article on ethical questions presented by artificially intelligent first author, "We Asked GPT-3 to Write an Academic Paper about Itself—Then We Tried to Get It Published", by Almira Osmanovic Thunström for the Scientific American.
- CLIP (Contrastive Language-Image Pre-Training) is a neural network trained on a variety of (image, text) pairs. The GitHub page of the project.
- Midjourney is "an independent research lab exploring new mediums of thought and expanding the imaginative powers of the human species".
‘An engine for the imagination’: the rise of AI image generators - An interview with Midjourney founder
Midjourney Founder David Holz On The Impact Of AI On Art, Imagination And The Creative Economy
- Imagen is "a text-to-image diffusion model with an unprecedented degree of photorealism and a deep level of language understanding", as described by the Brain Team of Google Research.- Disco Diffusion and the GitHub of Katherine Crowson. Here is a great tutorial to
"Get Started With Disco Diffusion to Create AI Generated Art" written by EdXD, also with very useful resources.
- For further information about A.I. in general and text-to-image diffusion models, please visit my Youtube playlist.
N.B.: Some pictures have been enhanced, for demonstration purposes only, using the A.I. based Animation Enhancement tool available on the website of the Tencent ARC (Applied Research Center) Lab.
All images can be used for non-commercial purposes only.
All "prompts" can be used to create new images.
DISCLAIMER: All creators, painters, sculptors, photographers mentioned in the prompts are used to guide the A.I..
Text-to-image artificial intelligence (AI) is a technology that enables a computer system to generate an image based on a given text description. This technology is often used in the field of natural language processing (NLP), which involves the analysis and understanding of human language by computers.
Text-to-image AI systems use machine learning algorithms to learn the relationships between words and visual concepts, and to generate an image that reflects the content described in a given text. These systems are trained on large datasets of images and their corresponding text descriptions, which allows them to learn how to generate images that are coherent with the text descriptions.
There are several applications for text-to-image AI, including generating images for social media posts, creating illustrations for documents and presentations, and providing visual aids for the visually impaired.
Text-to-image AI technology is an example of how machine learning can be used to generate visual content based on text input, and it is an active area of research in the field of NLP. As the technology continues to improve, it has the potential to revolutionize the way we communicate and interact with visual information.
Text-to-image artificial intelligence (AI) and machine learning have the potential to revolutionize the way we create, interpret, and interact with visual information and art. These technologies allow for the automatic generation of images based on text descriptions, which can save time and effort in the creative process and enable the creation of visual content that would be difficult or impossible to produce manually.
However, it's important to note that text-to-image AI systems are not yet able to generate high-quality images that are indistinguishable from those created by humans. While the technology is improving rapidly, it is still in the early stages of development and has a long way to go before it can fully replace human creativity and artistic skill.
In the future, text-to-image AI and machine learning may change the way we think about art and visual information, but it is unlikely to fully replace the role of human artists and designers. Instead, these technologies may augment and enhance the creative process, allowing artists and designers to experiment with new styles and techniques and to produce unique and innovative visual content.