Generative models
📰 OpenAI News
Generative models are a promising approach to develop algorithms that endow computers with an understanding of the world by training them to generate data like the data they are trained on
Action Steps
- Collect a large amount of data in a domain
- Train a generative model to generate data like the training data
- Use the model to discover and internalize the essence of the data
- Apply the model to various applications such as generating images or text
Who Needs to Know This
Data scientists and AI engineers on a team can benefit from understanding generative models as they can be used to analyze and understand large amounts of data, and product managers can use them to develop new applications
Key Insight
💡 Generative models can automatically learn the natural features of a dataset, whether categories or dimensions or something else entirely
Share This
💡 Generative models can help computers understand the world by generating data like the data they're trained on #AI #MachineLearning
Key Takeaways
Generative models are a promising approach to develop algorithms that endow computers with an understanding of the world by training them to generate data like the data they are trained on
Full Article
# Generative models | OpenAI
[Skip to main content](https://openai.com/index/generative-models#main)
[](https://openai.com/)
* [Research](https://openai.com/research/index/)
* Products
* [Business](https://openai.com/business/)
* [Developers](https://openai.com/api/)
* [Company](https://openai.com/about/)
* [Foundation(opens in a new window)](https://openaifoundation.org/)
Log in[Try ChatGPT(opens in a new window)](https://chatgpt.com/)
* Research
* Products
* Business
* Developers
* Company
* [Foundation(opens in a new window)](https://openaifoundation.org/)
[Try ChatGPT(opens in a new window)](https://chatgpt.com/)Login
OpenAI
Table of contents
* [Generating images](https://openai.com/index/generative-models#generating-images)
* [DCGAN](https://openai.com/index/generative-models#dcgan)
* [Training a generative model](https://openai.com/index/generative-models#training-a-generative-model)
* [More general formulation](https://openai.com/index/generative-models#more-general-formulation)
* [Three approaches to generative models](https://openai.com/index/generative-models#three-approaches-to-generative-models)
* [Our recent contributions](https://openai.com/index/generative-models#our-recent-contributions)
* [Going forward](https://openai.com/index/generative-models#going-forward)
June 16, 2016
[Publication](https://openai.com/research/index/publication/)
# Generative models
Illustration:Ludwig Pettersson
Loading…
Share
This post describes four projects that share a common theme of enhancing or using generative models, a branch of unsupervised learning techniques in machine learning. In addition to describing our work, this post will tell you a bit more about generative models: what they are, why they are important, and where they might be going.
One of our core aspirations at OpenAI is to develop algorithms and techniques that endow computers with an understanding of our world.
It’s easy to forget just how much you know about the world: you understand that it is made up of 3D environments, objects that move, collide, interact; people who walk, talk, and think; animals who graze, fly, run, or bark; monitors that display information encoded in language about the weather, who won a basketball game, or what happened in 1970.
This tremendous amount of information is out there and to a large extent easily accessible—either in the physical world of atoms or the digital world of bits. The only tricky part is to develop models and algorithms that can analyze and understand this treasure trove of data.
**Generative models are one of the most promising approaches towards this goal**. To train a generative model we first collect a large amount of data in some domain (e.g., think millions of images, sentences, or sounds, etc.) and then train a model to generate data like it. The intuition behind this approach follows a famous quote from[Richard Feynman(opens in a new window)](https://en.wikipedia.org/wiki/Richard_Feynman):
> “What I cannot create, I do not understand.”
Richard Feynman
The trick is that the neural networks we use as generative models have a number of parameters significantly smaller than the amount of data we train them on, so the models are forced to discover and efficiently internalize the essence of the data in order to generate it.
Generative models have many short-term[applications](https://openai.com/index/generative-models/#going-forward). But in the long run, they hold the potential to automatically learn the natural features of a dataset, whether categories or dimensions or something else entirely.
## Generating images
Let’s make this more concrete with an example. Suppose we have some large collection of images, such as the 1.2 million images in the[ImageNet(opens in a new
[Skip to main content](https://openai.com/index/generative-models#main)
[](https://openai.com/)
* [Research](https://openai.com/research/index/)
* Products
* [Business](https://openai.com/business/)
* [Developers](https://openai.com/api/)
* [Company](https://openai.com/about/)
* [Foundation(opens in a new window)](https://openaifoundation.org/)
Log in[Try ChatGPT(opens in a new window)](https://chatgpt.com/)
* Research
* Products
* Business
* Developers
* Company
* [Foundation(opens in a new window)](https://openaifoundation.org/)
[Try ChatGPT(opens in a new window)](https://chatgpt.com/)Login
OpenAI
Table of contents
* [Generating images](https://openai.com/index/generative-models#generating-images)
* [DCGAN](https://openai.com/index/generative-models#dcgan)
* [Training a generative model](https://openai.com/index/generative-models#training-a-generative-model)
* [More general formulation](https://openai.com/index/generative-models#more-general-formulation)
* [Three approaches to generative models](https://openai.com/index/generative-models#three-approaches-to-generative-models)
* [Our recent contributions](https://openai.com/index/generative-models#our-recent-contributions)
* [Going forward](https://openai.com/index/generative-models#going-forward)
June 16, 2016
[Publication](https://openai.com/research/index/publication/)
# Generative models
Illustration:Ludwig Pettersson
Loading…
Share
This post describes four projects that share a common theme of enhancing or using generative models, a branch of unsupervised learning techniques in machine learning. In addition to describing our work, this post will tell you a bit more about generative models: what they are, why they are important, and where they might be going.
One of our core aspirations at OpenAI is to develop algorithms and techniques that endow computers with an understanding of our world.
It’s easy to forget just how much you know about the world: you understand that it is made up of 3D environments, objects that move, collide, interact; people who walk, talk, and think; animals who graze, fly, run, or bark; monitors that display information encoded in language about the weather, who won a basketball game, or what happened in 1970.
This tremendous amount of information is out there and to a large extent easily accessible—either in the physical world of atoms or the digital world of bits. The only tricky part is to develop models and algorithms that can analyze and understand this treasure trove of data.
**Generative models are one of the most promising approaches towards this goal**. To train a generative model we first collect a large amount of data in some domain (e.g., think millions of images, sentences, or sounds, etc.) and then train a model to generate data like it. The intuition behind this approach follows a famous quote from[Richard Feynman(opens in a new window)](https://en.wikipedia.org/wiki/Richard_Feynman):
> “What I cannot create, I do not understand.”
Richard Feynman
The trick is that the neural networks we use as generative models have a number of parameters significantly smaller than the amount of data we train them on, so the models are forced to discover and efficiently internalize the essence of the data in order to generate it.
Generative models have many short-term[applications](https://openai.com/index/generative-models/#going-forward). But in the long run, they hold the potential to automatically learn the natural features of a dataset, whether categories or dimensions or something else entirely.
## Generating images
Let’s make this more concrete with an example. Suppose we have some large collection of images, such as the 1.2 million images in the[ImageNet(opens in a new
DeepCamp AI