Marian Mazzone | College of Charleston (original) (raw)
Papers by Marian Mazzone
In July of 1888 Vincent van Gogh produced a series of drawings of the plain the Crau. Two drawing... more In July of 1888 Vincent van Gogh produced a series of drawings of the plain the Crau. Two drawings from this series are particularly shaped by the circumstances surrounding van Gogh at that time, and what he wanted to communicate about the French countryside. Wanting to produce drawings that would sell, van Gogh turned to methods of composition and style based on Dutch seventeenth-century panoramic landscapes, which were themselves shaped by the practices of map making. Van Gogh produced representations of the French countryside that reveal his nostalgic attitude and the biases of his class. What van Gogh saw in France was the old Holland of the seventeenth-century landscape artists, not France of the late nineteenth century. The drawings re-connect the artist to his Dutch visual heritage. They also reveal van Gogh's nostalgic view of the rural landscape, and his particularly Dutch attitude toward changes in this landscape caused by nineteenth-century modernization. (Abstract shortened with permission of author.
UMI Dissertation Services eBooks, 1997
Proceedings of the AAAI Conference on Artificial Intelligence
We present a machine learning system that can quantify fine art paintings with a set of visual el... more We present a machine learning system that can quantify fine art paintings with a set of visual elements and principles of art. The formal analysis is fundamental for understanding art, but developing such a system is challenging. Paintings have high visual complexities, but it is also difficult to collect enough training data with direct labels. To resolve these practical limitations, we introduce a novel mechanism, called proxy learning, which learns visual concepts in paintings through their general relation to styles. This framework does not require any visual annotation, but only uses style labels and a general relationship between visual concepts and style. In this paper, we propose a novel proxy model and reformulate four pre-existing methods in the context of proxy learning. Through quantitative and qualitative comparison, we evaluate these methods and compare their effectiveness in quantifying the artistic visual concepts, where the general relationship is estimated by langu...
arXiv: Artificial Intelligence, Jun 21, 2017
We propose a new system for generating art. The system generates art by looking at art and learni... more We propose a new system for generating art. The system generates art by looking at art and learning about style; and becomes creative by increasing the arousal potential of the generated art by deviating from the learned styles. We build over Generative Adversarial Networks (GAN), which have shown the ability to learn to generate novel images simulating a given distribution. We argue that such networks are limited in their ability to generate creative products in their original design. We propose modifications to its objective to make it capable of generating creative art by maximizing deviation from established styles and minimizing deviation from art distribution. We conducted experiments to compare the response of human subjects to the generated art with their response to art created by artists. The results show that human subjects could not distinguish art generated by the proposed system from art generated by contemporary artists and shown in top art fairs.
ArXiv, 2022
We present a machine learning system that can quantify fine art paintings with a set of visual el... more We present a machine learning system that can quantify fine art paintings with a set of visual elements and principles of art. This formal analysis is fundamental for understanding art, but developing such a system is challenging. Paintings have high visual complexities, but it is also difficult to collect enough training data with direct labels. To resolve these practical limitations, we introduce a novel mechanism, called proxy learning, which learns visual concepts in paintings though their general relation to styles. This framework does not require any visual annotation, but only uses style labels and a general relationship between visual concepts and style. In this paper, we propose a novel proxy model and reformulate four pre-existing methods in the context of proxy learning. Through quantitative and qualitative comparison, we evaluate these methods and compare their effectiveness in quantifying the artistic visual concepts, where the general relationship is estimated by langu...
Artnodes, 2020
Artificial intelligence researchers and artists have trained machines and generative processes to... more Artificial intelligence researchers and artists have trained machines and generative processes to produce visually interesting and novel works, thereby devising machinic means of creativity. At Artrendex, Playform was developed as an easy-to-use program specifically to be used by a broad range of artists, from beginners to those with advanced technical skills. This essay focuses on the motivations behind the development of Playform and the early reception and use of it by some artists. Our aim is to better understand both human and machine-based creativity at their intersection in an art generating system such as Playform.
Leonardo, 2020
This article positions Andy Warhol as a model for computational thinking and art-making, linking ... more This article positions Andy Warhol as a model for computational thinking and art-making, linking him to concepts in new media art. Warhol's work is analyzed for its variability in form generation and output, both in painting and on the early Amiga computer. His work becomes a simulation of the abstraction of process and methods of production familiar to us in electronic computational art of today. Rather than seen as banal mass production on the modern assembly line, Warhol's work can be seen as inspiration for new media arts practitioners.
Arts, 2019
Our essay discusses an AI process developed for making art (AICAN), and the issues AI creativity ... more Our essay discusses an AI process developed for making art (AICAN), and the issues AI creativity raises for understanding art and artists in the 21st century. Backed by our training in computer science (Elgammal) and art history (Mazzone), we argue for the consideration of AICAN’s works as art, relate AICAN works to the contemporary art context, and urge a reconsideration of how we might define human and machine creativity. Our work in developing AI processes for art making, style analysis, and detecting large-scale style patterns in art history has led us to carefully consider the history and dynamics of human art-making and to examine how those patterns can be modeled and taught to the machine. We advocate for a connection between machine creativity and art broadly defined as parallel to but not in conflict with human artists and their emotional and social intentions of art making. Rather, we urge a partnership between human and machine creativity when called for, seeing in this c...
2018 IEEE 12th International Conference on Semantic Computing (ICSC), 2018
We investigate a novel mechanism to interpret each coordinate of Deep Neural Network (DNN) activa... more We investigate a novel mechanism to interpret each coordinate of Deep Neural Network (DNN) activations with concepts aligned with art principles such as light, shape, pattern, line, and textures. After training state-of-the-art DNN architectures as paintings' style classifiers, we collect activations from the second to last layer of DNN and conduct a data analysis. Based on the interpretation results, we can demonstrate which art principles are essential when machines understand styles and also can decipher each coordinates value in the context of relatedness to specific art principles.
The short-lived Prague Spring of 1968 in Czechoslovakia was a demonstration of the impossibility ... more The short-lived Prague Spring of 1968 in Czechoslovakia was a demonstration of the impossibility of reform from within the Socialist system, and a lesson to the Czechoslovak population on the lack of constructive possibilities within that order. In the work of the Czech Milan Kní ák and the Slovak Július Koller, I explore how the artists internalised that lesson of failed revolution. Both artists launched new phases in their work in 1968–1969, prompting my inquiry into how and why they changed practices and imagery in this critical period. I approach their work as an example of how to communicate within a political and social culture undergoing ‘normalisation’. Within this context, I argue, the artists began employing techniques that the contemporary artworld would come to know as Conceptualist. Throughout the 1960s, culminating in the Prague Spring, many had believed that the Socialist system could be reformed from within; that it had the flexibility to be shaped to suit the needs ...
We propose a new system for generating art. The system generates art by looking at art and learni... more We propose a new system for generating art. The system generates art by looking at art and learning about style; and becomes creative by increasing the arousal potential of the generated art by deviating from the learned styles. We build over Generative Adversarial Networks (GAN), which have shown the ability to learn to generate novel images simulating a given distribution. We argue that such networks are limited in their ability to generate creative products in their original design. We propose modifications to its objective to make it capable of generating creative art by maximizing deviation from established styles and minimizing deviation from art distribution. We conducted experiments to compare the response of human subjects to the generated art with their response to art created by artists. The results show that human subjects could not distinguish art generated by the proposed system from art generated by contemporary artists and shown in top art fairs. Human subjects even ...
We propose a deep learning algorithm that can detect content and discover co-occurring patterns o... more We propose a deep learning algorithm that can detect content and discover co-occurring patterns of the content in fine art paintings. The following intellectual merits are the motivations of our project. First, the content detection provides a baseline of Computational Iconography (CI), which is to understand what objects/subjects can be seen in fine art paintings. Second, we argue that the found co-occurring patterns chart meaningful connectivity across content in art. Third, we imbed our system in Computational Creativity (CC) in a broad sense. By the nature of our system of machine learning, it creates informative connections between different modalities (images/words), which are not initially constructed or intentionally specified. Our system is automatically trained to discover the connective patterns reflecting artists’ creativity, which are latent in the large dataset of paintings. To build a content detector, we adopted an InceptionV3 (ImageNet) and fine-tuned it over 40,000...
Technoetic Arts: a Journal of Speculative Research, 2009
... piece of 1965 called Letters to the Population, containing a series of perverse instructions ... more ... piece of 1965 called Letters to the Population, containing a series of perverse instructions mailed to 1,000 recipients, selected randomly from the phonebook (Crane and Stofflet 1984: 69 ... in 1967 included a work entitled Telephone for Steve Abrams, that Friedman described thus ...
Artnodes, 2020
Artificial intelligence researchers and artists have trained machines and generative processes to... more Artificial intelligence researchers and artists have trained machines and generative processes to produce visually interesting and novel works, thereby devising machinic means of creativity. At Artrendex, Playform was developed as an easy-to-use programme specifically to be used by a broad range of artists, from beginners to those with advanced technical skills. This essay focuses on the motivations behind the development of Playform and the early reception and use of it by some artists. Our aim is to better understand both human and machine-based creativity at their intersection in an art-generating system such as Playform.
Leonardo, 2020
This article positions Andy Warhol as a model for computational thinking and art-making, linking ... more This article positions Andy Warhol as a model for computational thinking and art-making, linking him to concepts in new media art. Warhol's work is analyzed for its variability in form generation and output, both in painting and on the early Amiga computer. His work becomes a simulation of the abstraction of process and methods of production familiar to us in electronic computational art of today. Rather than seen as banal mass production on the modern assembly line, Warhol's work can be seen as inspiration for new media arts practitioners.
Arts, 2019
Our essay discusses an AI process developed for making art (AICAN), and the issues AI creativity ... more Our essay discusses an AI process developed for making art (AICAN), and the issues AI creativity raises for understanding art and artists in the 21st century. Backed by our training in computer science (Elgammal) and art history (Mazzone), we argue for the consideration of AICAN’s works as art, relate AICAN works to the contemporary art context, and urge a reconsideration of how we might define human and machine creativity. Our work in developing AI processes for art making, style analysis, and detecting large-scale style patterns in art history has led us to carefully consider the history and dynamics of human art-making and to examine how those patterns can be modeled and taught to the machine. We advocate for a connection between machine creativity and art broadly defined as parallel to but not in conflict with human artists and their emotional and social intentions of art making. Rather, we urge a partnership between human and machine creativity when called for, seeing in this collaboration a means to maximize both partners’ creative strengths.
How does the machine classify styles in art? And how does it relate to art historians's methods f... more How does the machine classify styles in art? And how does it relate to art historians's methods for analyzing style? Several studies have shown the ability of the machine to learn and predict style categories, such as Renaissance, Baroque, Impressionism, etc., from images of paintings. This implies that the machine can learn an internal representation encoding discriminative features through its visual analysis. However, such a representation is not necessarily interpretable. We conducted a comprehensive study of several of the state-of-the-art convolutional neural networks applied to the task of style classification on 77K images of paintings, and analyzed the learned representation through correlation analysis with concepts derived from art history. Surprisingly, the networks could place the works of art in a smooth temporal arrangement mainly based on learning style labels, without any a priori knowledge of time of creation, the historical time and context of styles, or relations between styles. The learned representations showed that there are few underlying factors that explain the visual variations of style in art. Some of these factors were found to correlate with style patterns suggested by Heinrich Wölfflin (1846-1945). The learned representations also consistently highlighted certain artists as the extreme distinctive representative of their styles, which quantitatively confirms art historian observations.
In July of 1888 Vincent van Gogh produced a series of drawings of the plain the Crau. Two drawing... more In July of 1888 Vincent van Gogh produced a series of drawings of the plain the Crau. Two drawings from this series are particularly shaped by the circumstances surrounding van Gogh at that time, and what he wanted to communicate about the French countryside. Wanting to produce drawings that would sell, van Gogh turned to methods of composition and style based on Dutch seventeenth-century panoramic landscapes, which were themselves shaped by the practices of map making. Van Gogh produced representations of the French countryside that reveal his nostalgic attitude and the biases of his class. What van Gogh saw in France was the old Holland of the seventeenth-century landscape artists, not France of the late nineteenth century. The drawings re-connect the artist to his Dutch visual heritage. They also reveal van Gogh's nostalgic view of the rural landscape, and his particularly Dutch attitude toward changes in this landscape caused by nineteenth-century modernization. (Abstract shortened with permission of author.
UMI Dissertation Services eBooks, 1997
Proceedings of the AAAI Conference on Artificial Intelligence
We present a machine learning system that can quantify fine art paintings with a set of visual el... more We present a machine learning system that can quantify fine art paintings with a set of visual elements and principles of art. The formal analysis is fundamental for understanding art, but developing such a system is challenging. Paintings have high visual complexities, but it is also difficult to collect enough training data with direct labels. To resolve these practical limitations, we introduce a novel mechanism, called proxy learning, which learns visual concepts in paintings through their general relation to styles. This framework does not require any visual annotation, but only uses style labels and a general relationship between visual concepts and style. In this paper, we propose a novel proxy model and reformulate four pre-existing methods in the context of proxy learning. Through quantitative and qualitative comparison, we evaluate these methods and compare their effectiveness in quantifying the artistic visual concepts, where the general relationship is estimated by langu...
arXiv: Artificial Intelligence, Jun 21, 2017
We propose a new system for generating art. The system generates art by looking at art and learni... more We propose a new system for generating art. The system generates art by looking at art and learning about style; and becomes creative by increasing the arousal potential of the generated art by deviating from the learned styles. We build over Generative Adversarial Networks (GAN), which have shown the ability to learn to generate novel images simulating a given distribution. We argue that such networks are limited in their ability to generate creative products in their original design. We propose modifications to its objective to make it capable of generating creative art by maximizing deviation from established styles and minimizing deviation from art distribution. We conducted experiments to compare the response of human subjects to the generated art with their response to art created by artists. The results show that human subjects could not distinguish art generated by the proposed system from art generated by contemporary artists and shown in top art fairs.
ArXiv, 2022
We present a machine learning system that can quantify fine art paintings with a set of visual el... more We present a machine learning system that can quantify fine art paintings with a set of visual elements and principles of art. This formal analysis is fundamental for understanding art, but developing such a system is challenging. Paintings have high visual complexities, but it is also difficult to collect enough training data with direct labels. To resolve these practical limitations, we introduce a novel mechanism, called proxy learning, which learns visual concepts in paintings though their general relation to styles. This framework does not require any visual annotation, but only uses style labels and a general relationship between visual concepts and style. In this paper, we propose a novel proxy model and reformulate four pre-existing methods in the context of proxy learning. Through quantitative and qualitative comparison, we evaluate these methods and compare their effectiveness in quantifying the artistic visual concepts, where the general relationship is estimated by langu...
Artnodes, 2020
Artificial intelligence researchers and artists have trained machines and generative processes to... more Artificial intelligence researchers and artists have trained machines and generative processes to produce visually interesting and novel works, thereby devising machinic means of creativity. At Artrendex, Playform was developed as an easy-to-use program specifically to be used by a broad range of artists, from beginners to those with advanced technical skills. This essay focuses on the motivations behind the development of Playform and the early reception and use of it by some artists. Our aim is to better understand both human and machine-based creativity at their intersection in an art generating system such as Playform.
Leonardo, 2020
This article positions Andy Warhol as a model for computational thinking and art-making, linking ... more This article positions Andy Warhol as a model for computational thinking and art-making, linking him to concepts in new media art. Warhol's work is analyzed for its variability in form generation and output, both in painting and on the early Amiga computer. His work becomes a simulation of the abstraction of process and methods of production familiar to us in electronic computational art of today. Rather than seen as banal mass production on the modern assembly line, Warhol's work can be seen as inspiration for new media arts practitioners.
Arts, 2019
Our essay discusses an AI process developed for making art (AICAN), and the issues AI creativity ... more Our essay discusses an AI process developed for making art (AICAN), and the issues AI creativity raises for understanding art and artists in the 21st century. Backed by our training in computer science (Elgammal) and art history (Mazzone), we argue for the consideration of AICAN’s works as art, relate AICAN works to the contemporary art context, and urge a reconsideration of how we might define human and machine creativity. Our work in developing AI processes for art making, style analysis, and detecting large-scale style patterns in art history has led us to carefully consider the history and dynamics of human art-making and to examine how those patterns can be modeled and taught to the machine. We advocate for a connection between machine creativity and art broadly defined as parallel to but not in conflict with human artists and their emotional and social intentions of art making. Rather, we urge a partnership between human and machine creativity when called for, seeing in this c...
2018 IEEE 12th International Conference on Semantic Computing (ICSC), 2018
We investigate a novel mechanism to interpret each coordinate of Deep Neural Network (DNN) activa... more We investigate a novel mechanism to interpret each coordinate of Deep Neural Network (DNN) activations with concepts aligned with art principles such as light, shape, pattern, line, and textures. After training state-of-the-art DNN architectures as paintings' style classifiers, we collect activations from the second to last layer of DNN and conduct a data analysis. Based on the interpretation results, we can demonstrate which art principles are essential when machines understand styles and also can decipher each coordinates value in the context of relatedness to specific art principles.
The short-lived Prague Spring of 1968 in Czechoslovakia was a demonstration of the impossibility ... more The short-lived Prague Spring of 1968 in Czechoslovakia was a demonstration of the impossibility of reform from within the Socialist system, and a lesson to the Czechoslovak population on the lack of constructive possibilities within that order. In the work of the Czech Milan Kní ák and the Slovak Július Koller, I explore how the artists internalised that lesson of failed revolution. Both artists launched new phases in their work in 1968–1969, prompting my inquiry into how and why they changed practices and imagery in this critical period. I approach their work as an example of how to communicate within a political and social culture undergoing ‘normalisation’. Within this context, I argue, the artists began employing techniques that the contemporary artworld would come to know as Conceptualist. Throughout the 1960s, culminating in the Prague Spring, many had believed that the Socialist system could be reformed from within; that it had the flexibility to be shaped to suit the needs ...
We propose a new system for generating art. The system generates art by looking at art and learni... more We propose a new system for generating art. The system generates art by looking at art and learning about style; and becomes creative by increasing the arousal potential of the generated art by deviating from the learned styles. We build over Generative Adversarial Networks (GAN), which have shown the ability to learn to generate novel images simulating a given distribution. We argue that such networks are limited in their ability to generate creative products in their original design. We propose modifications to its objective to make it capable of generating creative art by maximizing deviation from established styles and minimizing deviation from art distribution. We conducted experiments to compare the response of human subjects to the generated art with their response to art created by artists. The results show that human subjects could not distinguish art generated by the proposed system from art generated by contemporary artists and shown in top art fairs. Human subjects even ...
We propose a deep learning algorithm that can detect content and discover co-occurring patterns o... more We propose a deep learning algorithm that can detect content and discover co-occurring patterns of the content in fine art paintings. The following intellectual merits are the motivations of our project. First, the content detection provides a baseline of Computational Iconography (CI), which is to understand what objects/subjects can be seen in fine art paintings. Second, we argue that the found co-occurring patterns chart meaningful connectivity across content in art. Third, we imbed our system in Computational Creativity (CC) in a broad sense. By the nature of our system of machine learning, it creates informative connections between different modalities (images/words), which are not initially constructed or intentionally specified. Our system is automatically trained to discover the connective patterns reflecting artists’ creativity, which are latent in the large dataset of paintings. To build a content detector, we adopted an InceptionV3 (ImageNet) and fine-tuned it over 40,000...
Technoetic Arts: a Journal of Speculative Research, 2009
... piece of 1965 called Letters to the Population, containing a series of perverse instructions ... more ... piece of 1965 called Letters to the Population, containing a series of perverse instructions mailed to 1,000 recipients, selected randomly from the phonebook (Crane and Stofflet 1984: 69 ... in 1967 included a work entitled Telephone for Steve Abrams, that Friedman described thus ...
Artnodes, 2020
Artificial intelligence researchers and artists have trained machines and generative processes to... more Artificial intelligence researchers and artists have trained machines and generative processes to produce visually interesting and novel works, thereby devising machinic means of creativity. At Artrendex, Playform was developed as an easy-to-use programme specifically to be used by a broad range of artists, from beginners to those with advanced technical skills. This essay focuses on the motivations behind the development of Playform and the early reception and use of it by some artists. Our aim is to better understand both human and machine-based creativity at their intersection in an art-generating system such as Playform.
Leonardo, 2020
This article positions Andy Warhol as a model for computational thinking and art-making, linking ... more This article positions Andy Warhol as a model for computational thinking and art-making, linking him to concepts in new media art. Warhol's work is analyzed for its variability in form generation and output, both in painting and on the early Amiga computer. His work becomes a simulation of the abstraction of process and methods of production familiar to us in electronic computational art of today. Rather than seen as banal mass production on the modern assembly line, Warhol's work can be seen as inspiration for new media arts practitioners.
Arts, 2019
Our essay discusses an AI process developed for making art (AICAN), and the issues AI creativity ... more Our essay discusses an AI process developed for making art (AICAN), and the issues AI creativity raises for understanding art and artists in the 21st century. Backed by our training in computer science (Elgammal) and art history (Mazzone), we argue for the consideration of AICAN’s works as art, relate AICAN works to the contemporary art context, and urge a reconsideration of how we might define human and machine creativity. Our work in developing AI processes for art making, style analysis, and detecting large-scale style patterns in art history has led us to carefully consider the history and dynamics of human art-making and to examine how those patterns can be modeled and taught to the machine. We advocate for a connection between machine creativity and art broadly defined as parallel to but not in conflict with human artists and their emotional and social intentions of art making. Rather, we urge a partnership between human and machine creativity when called for, seeing in this collaboration a means to maximize both partners’ creative strengths.
How does the machine classify styles in art? And how does it relate to art historians's methods f... more How does the machine classify styles in art? And how does it relate to art historians's methods for analyzing style? Several studies have shown the ability of the machine to learn and predict style categories, such as Renaissance, Baroque, Impressionism, etc., from images of paintings. This implies that the machine can learn an internal representation encoding discriminative features through its visual analysis. However, such a representation is not necessarily interpretable. We conducted a comprehensive study of several of the state-of-the-art convolutional neural networks applied to the task of style classification on 77K images of paintings, and analyzed the learned representation through correlation analysis with concepts derived from art history. Surprisingly, the networks could place the works of art in a smooth temporal arrangement mainly based on learning style labels, without any a priori knowledge of time of creation, the historical time and context of styles, or relations between styles. The learned representations showed that there are few underlying factors that explain the visual variations of style in art. Some of these factors were found to correlate with style patterns suggested by Heinrich Wölfflin (1846-1945). The learned representations also consistently highlighted certain artists as the extreme distinctive representative of their styles, which quantitatively confirms art historian observations.
A cura di aliCe Barale A fine 2018 un evento inaspettato si verifica presso la prestigiosa casa d... more A cura di aliCe Barale A fine 2018 un evento inaspettato si verifica presso la prestigiosa casa d'aste Christie's: per la prima volta viene battuta, con grande successo, un'opera d'arte assistita dall'intelligenza artificiale, è Edmond de Belamy del collettivo francese Obvious. La sorpresa e lo sconcerto della stampa rivelano quanto possa essere problematica l'idea stessa di intelligenza artificiale per il mondo dell'arte. Pochi mesi dopo, anche Sotheby's mette in asta un'installazione di Mario Klingemann, pioniere nell'utilizzo delle gAN (Generative Adversarial Networks), le reti generative avversarie inventate nel 2014 dal giovane studioso Ian Goodfellow. Le opere che usano l'intelligenza artificiale suscitano un dibattito sfaccettato e di grande interesse. Questo volume raccoglie le voci di filosofi, artisti della scena internazionale, informatici, studiosi d'arte e di musica che si confrontano, ognuno in modi e prospettive differenti, su domande di fondo: che cos'è la creatività? Le macchine possono essere creative o la creatività è solo una caratteristica umana? Chi è l'artefice e chi lo spettatore? Il processo generativo di un sistema di intelligenza artificiale può essere qualificato come originale? Possiamo chiamare "estetici" gli algoritmi delle gAN che discriminano fra milioni di "opere" generate da un computer?... Dalle pagine del libro, capitolo dopo capitolo, prende forma un panorama ricchissimo, anche dal punto di vista estetico, e di enorme fascino per chi ami percorrere le strade dell'arte, comunque imprevedibili.