The Electronic Curator

The Electronic Curator

A generative adversarial network creates vegetable faces artworks and curates them

Artificial Intelligence

  • 1 Collaborators

  • 0 Followers

    Follow

Description

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

Gallery

Video

11146596 10152972153311461 7756296601164880218 n

Eyal G. added photos to project The Electronic Curator

Medium 21873e0b 966e 48ab af68 beac49d4b5e0

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

11146596 10152972153311461 7756296601164880218 n

Eyal G. added photos to project The Electronic Curator

Medium f9e7700a eb41 4513 bcb3 c72892622982

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

11146596 10152972153311461 7756296601164880218 n

Eyal G. added photos to project The Electronic Curator

Medium 6d090060 c46f 445a a229 47e0a4cfa975

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

11146596 10152972153311461 7756296601164880218 n

Eyal G. added photos to project The Electronic Curator

Medium 01f82edf 45c3 4564 9d99 a470bd2fea54

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

11146596 10152972153311461 7756296601164880218 n

Eyal G. added photos to project The Electronic Curator

Medium 548e91f7 61c2 40f3 8b93 7537e64867f5

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

11146596 10152972153311461 7756296601164880218 n

Eyal G. added photos to project The Electronic Curator

Medium 7db5ba34 765d 45ff bdaf 6593803f44e3

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

11146596 10152972153311461 7756296601164880218 n

Eyal G. added photos to project The Electronic Curator

Medium f2e36666 583e 4f1a 831e 225cac77260b

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

11146596 10152972153311461 7756296601164880218 n

Eyal G. added photos to project The Electronic Curator

Medium e3e6bcfc c678 44cb 8dc1 55a5c0e6f1a9

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

Medium 11146596 10152972153311461 7756296601164880218 n

Eyal G. created project The Electronic Curator

Medium 346f8fd4 a966 4291 b1f0 28cb3f4fe334

The Electronic Curator

The Electronic Curator examines whether a computer can not only generate art, but also evaluate its quality. It is a dialog between two competing neural networks, representing a painter and a curator. The dialog between the competing networks represents the artistic process. Using cycle-consistent generative adversarial networks (CycleGAN), the networks are trained together, each getting better in its own task. The painter-network learns to create vegetable-face portraits from face images, while the curator-network learns to evaluate the painter's creation. Training is unsupervised and requires only a set of face images and an unrelated set of vegetable-faces collected from the Internet. In order to avoid mode collapse and get diverse and interesting results, we use a modified loss function inspired by DistanceGAN (arxiv.org/abs/1706.00826). In exhibition mode, the painter observes the spectator's face and turns it in real time into a vegetable face. The curator then grades the outcome, and a curatic text is generated based on the grade, as well as on the foods found in the artwork by object detection. In a world of creative machines and computer generated art, the act of curation is one of the last strongholds of the Human creator. The Electronic Curator discusses whether with the advent of GANs, this may soon be lost to the machines.

No users to show at the moment.