پیوندهای مفید
اخبار و اعلانات
آمار
| تعداد نشریات | 25 |
| تعداد شماره‌ها | 939 |
| تعداد مقالات | 7,734 |
| تعداد مشاهده مقاله | 12,721,686 |
| تعداد دریافت فایل اصل مقاله | 9,063,892 |
هنر و هوش مصنوعی از منظر رفتارگرایی و آزمون تورینگ | ||
| جلوه هنر | ||
| مقاله 6، دوره 16، شماره 2 - شماره پیاپی 43، شهریور 1403، صفحه 77-88 اصل مقاله (730.97 K) | ||
| نوع مقاله: مقاله پژوهشی | ||
| شناسه دیجیتال (DOI): 10.22051/jjh.2023.44059.1997 | ||
| نویسندگان | ||
| لیلا رضوانی1؛ بهنام کامرانی* 2؛ محمد رضا شریف زاده3 | ||
| 1دانشجوی دکتری پژوهش هنر، دانشکده هنر و معماری، واحد تهران مرکزی، دانشگاه آزاد اسلامی، تهران، ایران. | ||
| 2، استادیار گروه نقاشی، دانشکده هنرهای تجسمی، دانشگاه هنر ایران، تهران، ایران، نویسنده مسئول. | ||
| 3استاد گروه پژوهش هنر، دانشکده هنر و معماری، واحد تهران مرکزی، دانشگاه آزاد اسلامی، تهران، ایران. | ||
| چکیده | ||
| در سالهای اخیر هوش مصنوعی با ارائه امکانات، ابزارها و تکنیک‌های نوآورانه، تحولات چشمگیری را در حوزه هنر پدید آورده است. از اینرو، مانند هر فناوری دگرگون‌کننده دیگری، ادغام هوشمصنوعی در دنیای هنر و نقش غیرقابلانکار آن در خلق آثار هنری، سؤالات و ملاحظات متعددی را مطرح می‌کند که ممکن است تاکنون با دقت لازم مورد بررسی و ارزیابی قرار نگرفته باشند. از جمله مهمترین این پرسشها به عملکرد هوشمصنوعی در خلق اثر هنری اصیل باز میگردد. بهطور مشخص، این پرسش که آیا هوشمصنوعی میتواند آثاری همتراز آثار هنری انسان بهوجود آورد؟ هوشمصنوعی با رفتارگرایی، به عنوان یکی از مهمترین نظریههای فلسفه و روانشناسی، پیوند نزدیکی دارد. تاکید این نظریه بر عملکرد یا «رفتار قابل مشاهده» به عنوان کانون اصلی درک و تحلیل میباشد. از اینرو عملکرد هوشمصنوعی بر اساس این رویکرد در سطح رفتار مورد ارزیابی قرار خواهد گرفت. به این منظور در ابتدا بهطور مختصر به رویکرد رفتارگرایی و ارتباط آن با هوشمصنوعی پرداخته خواهد شد. سپس آزمون تورینگ به عنوان یک آزمون محوری در زمینه هوش مصنوعی شرح داده خواهد شد. این آزمون معیاری رفتارگرایانه در جهت سنجش هوش و خلاقیت ماشین به دست میدهد. در بخش بعدی با نظر به معیار رفتارگرایانه آزمون تورینگ و ملاحظه عناصر اصلی آثار هنری اصیل، دو دسته از آثار هنری مورد سنجش قرار خواهند گرفت. این دو دسته شامل آثار تولید شده توسط هوشمصنوعی و آثار خلق شده توسط انسان میباشد. این بررسی زمینه مناسبی را برای ارزیابی عملکرد ماشین در مقایسه با انسان به دست خواهد داد. بنابر این روش پژوهش حاضر به این صورت است که با استفاده از منابع مختلف، دادههایی در ارتباط با آثار هنری حاصل از هوشمصنوعی جمعآوری و به شیوهای کیفی توصیف میگردند. سپس این داده ها با عناصر موجود در آثار هنری انسان مقایسه میشوند و بر اساس معیار اصلی آزمون تورینگ مورد تحلیل قرار میگیرند. در نهایت این نتیجه حاصل میگردد که عملکرد هوشمصنوعی دستکم بر اساس معیارهای رفتارگرایانه یعنی از لحاظ رفتار مشاهدهپذیر، میتواند منجر به پدید آمدن آثاری همتراز آثار هنری خلق شده توسط انسان شود. | ||
| کلیدواژهها | ||
| کلیدواژهها: آزمون تورینگ؛ اثر هنری؛ نقاشی؛ هوش ماشینی؛ رفتارگرایی | ||
| موضوعات | ||
| هنرهای تجسمی و کاربردی | ||
| عنوان مقاله [English] | ||
| Art and Artificial Intelligence from the Perspective of Behaviorism and the Turing Test | ||
| نویسندگان [English] | ||
| Leila Rezvani1؛ Behnam Kamrani2؛ Mohammadreza Sharifzadeh3 | ||
| 1PhD Student of Art Research. Faculty of Arts and Architecture, Central Tehran Branch, Islamic Azad University, Tehran, Iran. | ||
| 2Assistant Professor, Department of Painting, Faculty of Visual Arts, Iran University of Art, Tehran, Iran, Corresponding Author. | ||
| 3Professor, Department of Art Research, Faculty of Arts and Architecture, Central Tehran Branch, Islamic Azad University, Tehran, Iran. | ||
| چکیده [English] | ||
| In recent years, artificial intelligence has created significant developments in various artistic fields, including architecture, cinema, music and other artistic fields, by providing innovative facilities, tools and techniques. Especially in the field of visual arts, it has caused the creation of significant works of art, whose nature requires a more detailed evaluation and understanding. Therefore, like any other transformative technology, the integration of artificial intelligence in the world of art and its undeniable role in the creation of works of art raises many questions and considerations that may not have been carefully examined until now. Among the most important of these questions comes back to the performance of artificial intelligence in creating an original work of art. Specifically, the question of whether artificial intelligence can create works equal to human works of art? Since AI is algorithm-based and often trained on existing art, can its output be considered a creative and authentic work of art? Therefore, the main goal of this article is to investigate the performance of artificial intelligence in the production of original works of art. Artificial intelligence is closely related to behaviorism, as one of the most important theories of philosophy and psychology. The emphasis of this theory is on performance or "observable behavior" as the main focus of understanding and analysis. Therefore, in this research, the performance of artificial intelligence is evaluated based on this approach at the level of behavior. For this purpose, the approach of behaviorism and its connection with artificial intelligence will be briefly discussed at first. Then the Turing test is described as a central test in the field of artificial intelligence. This test provides a behavioral criterion for measuring the intelligence and creativity of the machine. In the next section, two categories of works of art will be evaluated according to the behavioral criteria of the Turing test and considering the main elements of original works of art. These two categories include works produced by artificial intelligence and works created by humans. In this way, by using different sources, data related to art works from artificial intelligence are collected and described in a qualitative way. Then these data are compared with the elements found in human works of art and are analyzed based on the main criteria of the Turing test. This review provides a good basis for evaluating the machine's performance compared to humans. As mentioned, in the present research, in order to investigate the performance of artificial intelligence in creating an original work of art, firstly, the approach of behaviorism and its relationship with artificial intelligence is explained. Behaviorism is a school of thought in philosophy and psychology that claims that mental states such as thoughts, feelings, and beliefs can only be understood in terms of their relationship with "observable behavior". According to behaviorists, behavior can be explained based on causal relationships between environmental stimuli and behavioral responses. The focus of this research is on philosophical behaviorism. Philosophical behaviorism, also known as logical behaviorism, is a theory about the nature of the mind. Behaviorists in the field of philosophy also believe that mental states can only be explained in terms of behavior. In other words, mental states are only descriptions of certain types of behavior. This approach originated from the works of philosophers such as Gilbert Ryle and Ludwig Wittgenstein. From the 1950s until now, the connection between behaviorism and artificial intelligence was formed. Although these two may seem different from each other, they both pay special attention to "observable behavior". The initial research of artificial intelligence was influenced by the principles of behaviorism. For example, the Turing test, which is one of the most popular artificial intelligence tests, is basically a behavioral evaluation. Just as behaviorists consider the only valid data for studying the mind to be "observable behaviors," the Turing test also shows that a machine can be considered "intelligent" based on its "observable behavior" regardless of its internal processes or understanding. Therefore, this test provides a suitable criterion for measuring intelligence based on performance and "observable behavior". The Turing test is conducted through a question and answer process. A human evaluator or judge asks an unseen audience, either a human or a computer, through a text and with a computer, and based on the answers, he must determine whether he is talking to a human or a machine. If the judge cannot distinguish the machine from the human in the answers he receives, it is said that the machine has passed the test successfully. The results of this test depend on how much the machine's answers to the questions are indistinguishable from the answers given by a human or how similar they are to human answers. In this section, the performance stages of the Turing test are formulated as follows: Input behavior: the evaluator provides data to the machine and asks for an output from it. Output behavior: Then the machine responds to the input. This answer is "observable behavior" in this context. At the end of this section, in order to take advantage of the main criteria of the Turing test in order to evaluate the performance of artificial intelligence in the creation of artwork, a new formulation of this test in the field of art is presented, which is as follows: Input behavior: The programmer provides data related to the creation of the artwork to the machine and asks it for an artwork. Output behavior: Then the device produces the artwork in response to the input data. The output of the machine is "observable behavior" in the context of art. In the next section, first, the main elements of the works of art created by humans are listed as "observable behaviors". These elements provide criteria for measuring the artistic performance of artificial intelligence and include: form, content, style, context, intention, impact, innovation and novelty, complexity and depth, adaptation and evolution, individuality and uniqueness. These elements can interact in complex ways to create a rich and multifaceted work of art that engages the viewer's senses, emotions, and intellect. Therefore, they can be evaluated as "observable behaviors" in the works produced by artificial intelligence and compared with human art. In the following, the main components of a number of works produced by artificial intelligence are compared with the components in similar works of art created by humans. In this section, according to the main criteria in the theory of behaviorism and the Turing test, which only focus on "observable behaviors" rather than mental processes, it will be evaluated that the elements of the works of artificial intelligence as "observable behaviors" to what extent and in what ways are similar to the elements in human works of art. In addition, to what extent artificial intelligence has been successful in creating a human-like behavior in order to create a work of art. By comparing the work of Tom White and Wassily Kandinsky's machine, it can be seen that both of them turned to abstractionism in the creation of works of art, and with this method, in their works, they revealed new concepts in relation to visible reality. By taking a creative approach, Gene Kogan and Jackson Pollack have abandoned the usual traditions to create original works of art; They have found art in many things that are not in the art scene or do not even call themselves art. Obvious’s machine and Caspar Netscher consider realism in their works. Light is considered one of the basic elements of their images, as its intensity and angle of radiation emphasize the inner power of the portrait in their paintings. paintings by Robbie Barratt’s machine and Claude Monet are both in impressionism style. The elements in their works have vague and often abstract lines. The colors have been combined with each other in such a way that they have created a soft surface and cover, and the way of using natural light in their paintings has made them attractive. To pass the Turing test, the only thing that was considered was how close the answers of the machine are to the answers of a human. In this research, just like the hypothesis that was raised in the Turing test, by evaluating the similarities between the works of artificial intelligence and humans, it is shown that machines have the same "observable behaviors" as human artists in creating works of art. to be Therefore, they pass the Turing test successfully. Therefore, they can be considered intelligent. In the present research, the result is that artificial intelligence can express itself effectively and with a distinctive style in terms of "observable behavior" in the creation of artwork and create original and creative works, just like human works. | ||
| کلیدواژهها [English] | ||
| Keywords: Turing test, Artwork, Painting, Machine intelligence, Behaviorism | ||
| مراجع | ||
|
منابع
نوزاد، هما. (1400). رهیافت تحلیلیِ هنر دیجیتالی‌تعاملی با تاکید بر ادراک بدنمند مرلوپونتی (مطالعه موردی: نمایشگاه هنرمندان و روبات‌های پاریس). جلوه هنر. (1) 13. 90-80. doi: 10.22051/jjh.2021.32242.1535
References
Arnheim, R., (1954). Art and visual perception: A psychology of the creative eye. Univ of California Press.
Boden, M., (2008), Computers and creativity: models and applications. The Routledge Companion to Creativity, 179-188.
Browne, K., (2022). Who (or what) is an AI Artist? Leonardo, 55(2), 130-134.
Descartes, R., (1996). Discourse on Method and Meditations on First Philosophy. New Haven & London: Yale University Press. ISBN 978-0300067729.
Grba, D., (2022). The Cabinet of Wolfgang von Kempelen: AI Art and Creative Agency, See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/362668667.
Harnad, S., (2008). The annotation game: On turing (1950) on computing, machinery, and intelligence, (published version bowdlerized).
Harnad, S., (1991). Other bodies, other minds: A machine incarnation of an old philosophical problem. Minds and Machines, 1, 43-54.
James’s, W., (1863). Principles of psychology.
Morris, E. K., (1992). The aim, progress, and evolution of behavior analysis. The Behavior Analyst, 15, 3-29.
Natale, S., Henrickson, L., (2022). Perception of Creativity in Machines, New Media and Society, https://mc.manuscriptcentral.com/nms.
Nozad, H., (2021). The Analysis of Interactive Digital Art, Based on Merlea-Ponty Body Perception, (Case Study: Artists and Robots Exhibition of Paris), Glory of Art (Jelve-y Honar), 13(5), (Text in Persian).
Pepperell, R., (2017). Putting the Art in Artificial: Aesthetic Responses to Computer-Generated Art. Psychology of Aesthetics Creativity and the Arts, 12(2).
Pierce, W. D., Cheney, C. D., (2017). Behavior analysis and learning: A biobehavioral approach. Routledge.
Russell, S., Norvig, P., (2003)(1995). Artificial Intelligence: A Modern Approach, (2nd ed), Prentice Hall. ISBN 978-0137903955.
Ryle, G., (2009). The concept of mind. Routledge.
Saygin, A. P., Cicekli, I., Akman, V., (2003). Turing Test: 50 Years Later, Minds and Machines, Reprinted in Moor.
Shieber, S. M., (Ed.). (2004). The Turing test: verbal behavior as the hallmark of intelligence. Mit Press.
Srinivasan, R., Denton, E., Famularo, J., Rostamzadeh, N., Diaz, F., Coleman, B., (2022). Artsheets for Art Datasets, 35th Conference on Neural Information Processing Systems, Datasets and Benchmarks.
Turing, A., (1950). Computing Machinery and Intelligence, Mind, 59(236), 433-460. doi:10.1093/mind/LIX.236.433
White, T., (2018). Perception engines. Blog post at https://medium. com/artists-and-machineintelligence/perception-engines-8a46bc598d57.
Wittgenstein, L., (1953). Philosophical Investigations. Blackwell.
URLs
URL5. http://www.designartpractice.com/blog/20. 2015
URL6. https://www.frieze.com/tags/tom-white. 2023
URL10.https://obvious-art.com/la-famille-belamy/. 2023
URL11. https://plus359gallery.com/en/2019/06/12/tom-white/. 2019
URL12. https://www.theverge.com/2018/8/21/17761424/ai-algorithm-art-machine-vision-perception-tom-white-treachery-imagenet. 2018
URL14. https://twitter.com/genekogan?lang=zh-Hant. 2023
| ||
|
آمار تعداد مشاهده مقاله: 380 تعداد دریافت فایل اصل مقاله: 144 |
||