Üretken Yapay Zekâ İle Yapılan Sanatsal Stil Taklidini Engelleme
Özet
Metinden görüntü üreten yapay zekâ modelleri, kullanıcıdan doğal dil istemi biçiminde girdi alır ve bu istemle eşleşen bir görüntü üretir. Bunun için modelin; webden alınan çok sayıda resim, medya ve metin açıklamalarından oluşan bir veri kümesi üzerinden eğitilmesi gerekir. İnternetteki görseller ve metin açıklamaları, veri madenciliği veya veri kazıma adı verilen bir uygulamayla toplanır ve alınır. Bu verilerin bir kısmı sanatçıların telif hakkıyla korunan eserleri ve kamuya ait özel verilerdir. Modellere girilen istemler çoğunlukla belirli sanatçıların tarzında çalışmalar oluşturmak üzerinedir. Bu durum, yapay zekâ uygulamalarını kullanan kişiler için önem teşkil etmese de kendine özgü stili taklit edilen sanatçı ve tasarımcılar için olumsuz bir tablo ortaya çıkarmaktadır. Stil taklidini engellemek isteyen sanatçı/tasarımcılar ve bilim insanları bu doğrultuda çeşitli önlemler geliştirmektedir. Çalışma, stil taklidini engellemek üzere geliştirilen önlemleri derlemek amacıyla hazırlanmıştır. Bu bağlamda stil taklidinin sanatçılar ve tasarımcılar üzerindeki olumsuz etkileri, stil taklidinin hukuki boyutu gibi konulara değinilmiştir.
Referanslar
Arte es Ética. (2023). Un manifiesto para comprender y regular la I.A. generativa. (21 Haziran 2024 tarihinde https://arteesetica.org adresinden ulaşılmıştır).
Back, S. S. & Hun, K. Y. (2018). Nonfacial Portrait. (16 Haziran 2024 tarihinde https://ssbkyh.com/works/nonfacial_portrait/ adresinden ulaşılmıştır).
Belsky, S. & Rao, D. (2024). Updating Adobe’s Terms of Use. (11 Haziran 2024 tarihinde https://blog.adobe.com/en/publish/2024/06/10/updating-adobes-terms-of-use adresinden ulaşılmıştır).
Bridy, A. (2016). “The Evolution of Authorship: Work Made by Code”, Columbia Journal of Law & ARTS, No 39, p.395-401.
CAA. (t.y.). Advocacy. (12 Haziran 2024 tarihinde https://www.conceptartassociation.com/advocacy adresinden ulaşılmıştır).
Cara. (2023). (10 Haziran 2024 tarihinde https://cara.app/faq. Adresinden ulaşılmıştır).
Cherepanova, V., Goldblum, M., Foley, H., et al. (2021). LowKey: Leveraging Adversarial Attacks to Protect Social Media Users from Facial Recognition. International Conference on Learning Representations (ICLR 2021), May 04 2021, Vienna, Austria.
Coyle, M. (2024). Getty Images vs. Stability AI: The High Court Showdown over AI and Copyright Infringement. (12 Haziran 2024 tarihinde https://lawdit.co.uk/readingroom/getty-images-stability-ai-high-court-case adresinden ulaşılmıştır).
David, E. (2023). Getty lawsuit against Stability AI to go to trial in the UK. (17 Haziran 2024 tarihinde https://www.theverge.com/2023/12/4/23988403/getty-lawsuit-stability-ai-copyright-infringement adresinden ulaşılmıştır).
Deltorn, J. M. & Macrez, F. (2018). Authorship in the Age of Machine Learning and Artificial Intelligence, Center for International Intellectual Property Studies Research Paper, No 10, p.1- 24.
EGAIR. (2023). A proposal to regulate AI in EU: Our Manifesto for AI companies regulation in Europe. (13 Haziran 2024 tarihinde https://www.egair.eu/ adresinden ulaşılmıştır).
European Parliament News. (2024). Artificial Intelligence Act: MEPs adopt landmark law. (10 Haziran 2024 tarihinde https://www.europarl.europa.eu/news/en/press-room/20240308IPR19015/artificial-intelligence-act-meps-adopt-landmark-law adresinden ulaşılmıştır).
EU DSM Digital Single Market. (18 Haziran 2024 tarihinde https://eur-lex.europa.eu/eli/dir/2019/790/oj adresinden ulaşılmıştır).
Feng, W., He, J., Zhang, J., et al. (2023). Catch You Everything Everywhere: Guarding Textual Inversion via Concept Watermarking. https://doi.org/10.48550/arXiv.2309.05940
Fikir ve Sanat Eserleri Kanunu. (2014). (11 Haziran 2024 tarihinde https://www.mevzuat.gov.tr/mevzuatmetin/1.3.5846-20141212.pdf adresinden ulaşılmıştır).
Freedman, R. (2023). Microsoft seeks to reassure Copilot users with genAI copyright legal defense. (12 Haziran 2024 tarihinde https://www.legaldive.com/news/microsoft-copilot-user-copyright-legal-indemnification-ip-law/693163/ adresinden ulaşılmıştır).
Gandikota, R., Materzynska, J., Kaufman, J. F., et al. (2023) Erasing Concepts from Diffusion Models. Proceedings of the 2023 IEEE International Conference on Computer Vision. ICCV 2023.
Gervais, D. J. (2019). The Machine as Author. Iowa Law Review, 105, 2053-2106.
Goodfellow, I. J., Shlens & J., Szegedy, C. (2014). Explaining and harnessing adversarial examples. https://doi.org/10.48550/arXiv.1412.6572
Guadamuz, A. (2024). A scanner darkly: copyright liability and exceptions in artificial intelligence inputs and outputs. GRUR International. https://doi.org/10.1093/grurint/ikad140
Harvey, A. (2015). CV Dazzle. (14 Haziran 2024 tarihinde https://adam.harvey.studio/cvdazzle adresinden ulaşılmıştır).
Heikkilä, M. (2022). Artists can now opt out of the next version of Stable Diffusion. (17 Haziran 2024 tarihinde https://www.technologyreview.com/2022/12/16/1065247/artists-can-now-opt-out-of-the-next-version-of-stable-diffusion/ adresinden ulaşılmıştır).
Heikkila, M. (2023). We need to bring consent to AI. MIT Technology Review. (12 Haziran 2024 tarihinde https://www.technologyreview.com/2023/05/02/1072556/we-need-to-bring-consent-to-ai/ adresinden ulaşılmıştır).
HSK. İnsan Hakları Evrensel Beyannamesi. (13 Nisan 2024 tarihinde https://www.hsk.gov.tr/Eklentiler/Dosyalar/9a3bfe74-cdc4-4ae4-b876-8cb1d7eeae05.pdf adresinden ulaşılmıştır).
Jiang, H., Brown, L., Cheng, J., et al. (2023). AI Art and its Impact on Artists. AIES ’23, August 08–10, 2023, Montréal, QC, Canada. https://doi.org/10.1145/3600211.3604681
Laird, J. (2023). New anti-AI tool 'poisons' generative models to protect artwork from unauthorized robo-Rembrandts. (5 Haziran 2024 tarihinde https://www.pcgamer.com/new-anti-ai-tool-poisons-generative-models-to-protect-artwork-from-unauthorized-robo-rembrandts/ adresinden ulaşılmıştır).
Le, T. V., Phung, H., Nguyen, T. H., et al. (2023). Anti-DreamBooth: Protecting users from personalized text-to-image synthesis. https://doi.org/10.48550/arXiv.2303.15433
Leffer, L. (2024). Artists Are Slipping Anti-AI ‘Poison’ into Their Art. Here’s How It Works. (5 Haziran 2024 tarihinde https://www.scientificamerican.com/article/art-anti-ai-poison-heres-how-it-works/ adresinden ulaşılmıştır).
Luo, G., Huang, J., Zhang, M., et al. (2023). Steal My Artworks for Fine-tuning? A Watermarking Framework for Detecting Art Theft Mimicry in Text-to-Image Models. https://doi.org/10.48550/arXiv.2311.13619
McCormack, J., Gambardella, C. C., Rajcic, N., et al. (2023). Is Writing Prompts Really Making Art? https://doi.org/10.48550/arXiv.2301.13049.
Mearian, L. (2023). ‘Data poisoning’ anti-AI theft tools emerge - but are they ethical? (7 Haziran 2024 tarihinde https://www.computerworld.com/article/1638694/data-poisoning-anti-ai-theft-tools-emerge-but-are-they-ethical.html adresinden ulaşılmıştır).
Niemann, C. (2023). (5 Mayıs 2024 tarihinde https://x.com/abstractsunday/status/1798361022962823446 adresinden ulaşılmıştır).
Park, Y. S. (2023) Creative and Critical Entanglements With AI in Art Education. Studies in Art Education, 64:4, 406-425, DOI: 10.1080/00393541.2023.2255084
Philippot D. (2023). AI called to replace journalists at the German group Springer. Le Figaro. (2 Haziran 2024 tarihinde https://www.lefigaro.fr/ medias/l-ia-appelee-a-remplacer-des-journalistes-chez-le-groupe-allemand-springer-20230302 adresinden ulaşılmıştır).
Popli, N. (2022). He made a children’s book using AI. Artists are not happy. Time. (17 Nisan 2024 tarihinde https://time.com/6240569/ai- childrens-book-alice-and-sparkle-artists-unhappy/ adresinden ulaşılmıştır).
Reese, B. (2020). Yapay Zekâ Çağı Dördüncü Çağ: Akıllı Robotlar, Bilinçli Bilgisayarlar ve İnsanlığın Geleceği. (Çev. Mihriban Doğan). İstanbul: Say Yayınları.
Rose, J. (2022). Inside Midjourney, The Generative Art AI That Rivals DALL-E. (12 Mayıs 2024 tarihinde https://www.vice.com/en/article/wxn5wn/inside-midjourney-the-generative-art-ai-that-rivals-dall-e adresinden ulaşılmıştır).
Roose, K. (2022). An AI-Generated Picture Won an Art Prize. Artists Aren’t Happy. (5 Haziran 2024 tarihinde https://www.nytimes.com/2022/09/02/technology/ai-artificial-intelligence-artists.html adresinden ulaşılmıştır).
Salman, H., Khaddaj, A., Leclerc, G., et al. (2023). Raising the Cost of Malicious AI-Powered Image Editing. https://doi.org/10.48550/arXiv.2302.06588
Samuelson, P. (2013). The Quest for a Sound Conception of Copyright's Derivative Work Right. Georgetown Law Journal 101 (6). 1505-1564.
Sanative. (2023). Building tools to help people thrive in a world blended with AI. (9 Haziran 2024 tarihinde https://sanative.ai/ adresinden ulaşılmıştır).
Saveri, J. & Butterrick, M. (2024). Image Generator Litigation. (20 Haziran 2024 tarihinde https://imagegeneratorlitigation.com/# adresinden ulaşılmıştır).
Shan, S., Wenger, E., Zhang, J., et al. (2020). Fawkes: Protecting Privacy against Unauthorized Deep Learning Models. Proceedings of USENIX Security Symposium, August 12-14, 2020.
Shan, S., Cryan, J., Wenger, E., et al. (2023). Glaze: Protecting Artists from Style Mimicry by Text-to-Image Models. 32nd USENIX Security Symposium, 9-11 August, 2023, Anaheim, CA, USA
Shan, S., Ding, W., Passananti, J., et al. (2024). Nightshade: Prompt-Specific Poisoning Attacks on Text-to-Image Generative Models. 2024 IEEE Symposium on Security and Privacy.
Shan, S. & Zhao, B. (2024).What Is Nightshade? Why Does It Work, and Limitations. https://nightshade.cs.uchicago.edu/whatis.html adresinden ulaşılmıştır).
Sonnet, S. (2023). Nightshade: A defensive tool for artists against AI Art Generators. (15 Haziran 2024 tarihinde https://amt-lab.org/reviews/2023/11/nightshade-a-defensive-tool-for-artists-against-ai-art-generators adresinden ulaşılmıştır).
Vincent, J. (2023). Getty Images is suing the creators of AI art tool Stable Diffusion for scraping its content. (23 Mayıs 2024 tarihinde https://www.theverge.com/2023/1/17/23558516/ai-art-copyright-stable-diffusion-getty-images-lawsuit adresinden ulaşılmıştır).
Vinchon, F., Lubart, T., Bartolotta, S., et al. (2023). Artificial Intelligence & Creativity: A Manifesto for Collaboration. The Journal of Creative Behavior, 57 (4), 472–484. DOI: 10.1002/jocb.597
Vyas, N., Kakade, S. & Barak, B. (2023). On Provable Copyright Protection for Generative Models https://doi.org/10.48550/arXiv.2302.10870
Yerushalmy, J. (2023). German publisher Axel Springer says journalists could be replaced by AI. (14 Haziran 2024 tarihinde https://www.theguardian.com/technology/2023/mar/01/german-publisher-axel-springer-says-journalists-could-be-replaced-by-ai adresinden ulaşılmıştır).
Zhai, S., Wang, W., Li, J., et al. (2024). Discovering Universal Semantic Triggers for Text-to-Image Synthesis. https://doi.org/10.48550/arXiv.2402.07562
Zhang, E., Wang, K., Xu, X., et al. (2023). Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models. https://doi.org/10.48550/arXiv.2303.17591
Referanslar
Arte es Ética. (2023). Un manifiesto para comprender y regular la I.A. generativa. (21 Haziran 2024 tarihinde https://arteesetica.org adresinden ulaşılmıştır).
Back, S. S. & Hun, K. Y. (2018). Nonfacial Portrait. (16 Haziran 2024 tarihinde https://ssbkyh.com/works/nonfacial_portrait/ adresinden ulaşılmıştır).
Belsky, S. & Rao, D. (2024). Updating Adobe’s Terms of Use. (11 Haziran 2024 tarihinde https://blog.adobe.com/en/publish/2024/06/10/updating-adobes-terms-of-use adresinden ulaşılmıştır).
Bridy, A. (2016). “The Evolution of Authorship: Work Made by Code”, Columbia Journal of Law & ARTS, No 39, p.395-401.
CAA. (t.y.). Advocacy. (12 Haziran 2024 tarihinde https://www.conceptartassociation.com/advocacy adresinden ulaşılmıştır).
Cara. (2023). (10 Haziran 2024 tarihinde https://cara.app/faq. Adresinden ulaşılmıştır).
Cherepanova, V., Goldblum, M., Foley, H., et al. (2021). LowKey: Leveraging Adversarial Attacks to Protect Social Media Users from Facial Recognition. International Conference on Learning Representations (ICLR 2021), May 04 2021, Vienna, Austria.
Coyle, M. (2024). Getty Images vs. Stability AI: The High Court Showdown over AI and Copyright Infringement. (12 Haziran 2024 tarihinde https://lawdit.co.uk/readingroom/getty-images-stability-ai-high-court-case adresinden ulaşılmıştır).
David, E. (2023). Getty lawsuit against Stability AI to go to trial in the UK. (17 Haziran 2024 tarihinde https://www.theverge.com/2023/12/4/23988403/getty-lawsuit-stability-ai-copyright-infringement adresinden ulaşılmıştır).
Deltorn, J. M. & Macrez, F. (2018). Authorship in the Age of Machine Learning and Artificial Intelligence, Center for International Intellectual Property Studies Research Paper, No 10, p.1- 24.
EGAIR. (2023). A proposal to regulate AI in EU: Our Manifesto for AI companies regulation in Europe. (13 Haziran 2024 tarihinde https://www.egair.eu/ adresinden ulaşılmıştır).
European Parliament News. (2024). Artificial Intelligence Act: MEPs adopt landmark law. (10 Haziran 2024 tarihinde https://www.europarl.europa.eu/news/en/press-room/20240308IPR19015/artificial-intelligence-act-meps-adopt-landmark-law adresinden ulaşılmıştır).
EU DSM Digital Single Market. (18 Haziran 2024 tarihinde https://eur-lex.europa.eu/eli/dir/2019/790/oj adresinden ulaşılmıştır).
Feng, W., He, J., Zhang, J., et al. (2023). Catch You Everything Everywhere: Guarding Textual Inversion via Concept Watermarking. https://doi.org/10.48550/arXiv.2309.05940
Fikir ve Sanat Eserleri Kanunu. (2014). (11 Haziran 2024 tarihinde https://www.mevzuat.gov.tr/mevzuatmetin/1.3.5846-20141212.pdf adresinden ulaşılmıştır).
Freedman, R. (2023). Microsoft seeks to reassure Copilot users with genAI copyright legal defense. (12 Haziran 2024 tarihinde https://www.legaldive.com/news/microsoft-copilot-user-copyright-legal-indemnification-ip-law/693163/ adresinden ulaşılmıştır).
Gandikota, R., Materzynska, J., Kaufman, J. F., et al. (2023) Erasing Concepts from Diffusion Models. Proceedings of the 2023 IEEE International Conference on Computer Vision. ICCV 2023.
Gervais, D. J. (2019). The Machine as Author. Iowa Law Review, 105, 2053-2106.
Goodfellow, I. J., Shlens & J., Szegedy, C. (2014). Explaining and harnessing adversarial examples. https://doi.org/10.48550/arXiv.1412.6572
Guadamuz, A. (2024). A scanner darkly: copyright liability and exceptions in artificial intelligence inputs and outputs. GRUR International. https://doi.org/10.1093/grurint/ikad140
Harvey, A. (2015). CV Dazzle. (14 Haziran 2024 tarihinde https://adam.harvey.studio/cvdazzle adresinden ulaşılmıştır).
Heikkilä, M. (2022). Artists can now opt out of the next version of Stable Diffusion. (17 Haziran 2024 tarihinde https://www.technologyreview.com/2022/12/16/1065247/artists-can-now-opt-out-of-the-next-version-of-stable-diffusion/ adresinden ulaşılmıştır).
Heikkila, M. (2023). We need to bring consent to AI. MIT Technology Review. (12 Haziran 2024 tarihinde https://www.technologyreview.com/2023/05/02/1072556/we-need-to-bring-consent-to-ai/ adresinden ulaşılmıştır).
HSK. İnsan Hakları Evrensel Beyannamesi. (13 Nisan 2024 tarihinde https://www.hsk.gov.tr/Eklentiler/Dosyalar/9a3bfe74-cdc4-4ae4-b876-8cb1d7eeae05.pdf adresinden ulaşılmıştır).
Jiang, H., Brown, L., Cheng, J., et al. (2023). AI Art and its Impact on Artists. AIES ’23, August 08–10, 2023, Montréal, QC, Canada. https://doi.org/10.1145/3600211.3604681
Laird, J. (2023). New anti-AI tool 'poisons' generative models to protect artwork from unauthorized robo-Rembrandts. (5 Haziran 2024 tarihinde https://www.pcgamer.com/new-anti-ai-tool-poisons-generative-models-to-protect-artwork-from-unauthorized-robo-rembrandts/ adresinden ulaşılmıştır).
Le, T. V., Phung, H., Nguyen, T. H., et al. (2023). Anti-DreamBooth: Protecting users from personalized text-to-image synthesis. https://doi.org/10.48550/arXiv.2303.15433
Leffer, L. (2024). Artists Are Slipping Anti-AI ‘Poison’ into Their Art. Here’s How It Works. (5 Haziran 2024 tarihinde https://www.scientificamerican.com/article/art-anti-ai-poison-heres-how-it-works/ adresinden ulaşılmıştır).
Luo, G., Huang, J., Zhang, M., et al. (2023). Steal My Artworks for Fine-tuning? A Watermarking Framework for Detecting Art Theft Mimicry in Text-to-Image Models. https://doi.org/10.48550/arXiv.2311.13619
McCormack, J., Gambardella, C. C., Rajcic, N., et al. (2023). Is Writing Prompts Really Making Art? https://doi.org/10.48550/arXiv.2301.13049.
Mearian, L. (2023). ‘Data poisoning’ anti-AI theft tools emerge - but are they ethical? (7 Haziran 2024 tarihinde https://www.computerworld.com/article/1638694/data-poisoning-anti-ai-theft-tools-emerge-but-are-they-ethical.html adresinden ulaşılmıştır).
Niemann, C. (2023). (5 Mayıs 2024 tarihinde https://x.com/abstractsunday/status/1798361022962823446 adresinden ulaşılmıştır).
Park, Y. S. (2023) Creative and Critical Entanglements With AI in Art Education. Studies in Art Education, 64:4, 406-425, DOI: 10.1080/00393541.2023.2255084
Philippot D. (2023). AI called to replace journalists at the German group Springer. Le Figaro. (2 Haziran 2024 tarihinde https://www.lefigaro.fr/ medias/l-ia-appelee-a-remplacer-des-journalistes-chez-le-groupe-allemand-springer-20230302 adresinden ulaşılmıştır).
Popli, N. (2022). He made a children’s book using AI. Artists are not happy. Time. (17 Nisan 2024 tarihinde https://time.com/6240569/ai- childrens-book-alice-and-sparkle-artists-unhappy/ adresinden ulaşılmıştır).
Reese, B. (2020). Yapay Zekâ Çağı Dördüncü Çağ: Akıllı Robotlar, Bilinçli Bilgisayarlar ve İnsanlığın Geleceği. (Çev. Mihriban Doğan). İstanbul: Say Yayınları.
Rose, J. (2022). Inside Midjourney, The Generative Art AI That Rivals DALL-E. (12 Mayıs 2024 tarihinde https://www.vice.com/en/article/wxn5wn/inside-midjourney-the-generative-art-ai-that-rivals-dall-e adresinden ulaşılmıştır).
Roose, K. (2022). An AI-Generated Picture Won an Art Prize. Artists Aren’t Happy. (5 Haziran 2024 tarihinde https://www.nytimes.com/2022/09/02/technology/ai-artificial-intelligence-artists.html adresinden ulaşılmıştır).
Salman, H., Khaddaj, A., Leclerc, G., et al. (2023). Raising the Cost of Malicious AI-Powered Image Editing. https://doi.org/10.48550/arXiv.2302.06588
Samuelson, P. (2013). The Quest for a Sound Conception of Copyright's Derivative Work Right. Georgetown Law Journal 101 (6). 1505-1564.
Sanative. (2023). Building tools to help people thrive in a world blended with AI. (9 Haziran 2024 tarihinde https://sanative.ai/ adresinden ulaşılmıştır).
Saveri, J. & Butterrick, M. (2024). Image Generator Litigation. (20 Haziran 2024 tarihinde https://imagegeneratorlitigation.com/# adresinden ulaşılmıştır).
Shan, S., Wenger, E., Zhang, J., et al. (2020). Fawkes: Protecting Privacy against Unauthorized Deep Learning Models. Proceedings of USENIX Security Symposium, August 12-14, 2020.
Shan, S., Cryan, J., Wenger, E., et al. (2023). Glaze: Protecting Artists from Style Mimicry by Text-to-Image Models. 32nd USENIX Security Symposium, 9-11 August, 2023, Anaheim, CA, USA
Shan, S., Ding, W., Passananti, J., et al. (2024). Nightshade: Prompt-Specific Poisoning Attacks on Text-to-Image Generative Models. 2024 IEEE Symposium on Security and Privacy.
Shan, S. & Zhao, B. (2024).What Is Nightshade? Why Does It Work, and Limitations. https://nightshade.cs.uchicago.edu/whatis.html adresinden ulaşılmıştır).
Sonnet, S. (2023). Nightshade: A defensive tool for artists against AI Art Generators. (15 Haziran 2024 tarihinde https://amt-lab.org/reviews/2023/11/nightshade-a-defensive-tool-for-artists-against-ai-art-generators adresinden ulaşılmıştır).
Vincent, J. (2023). Getty Images is suing the creators of AI art tool Stable Diffusion for scraping its content. (23 Mayıs 2024 tarihinde https://www.theverge.com/2023/1/17/23558516/ai-art-copyright-stable-diffusion-getty-images-lawsuit adresinden ulaşılmıştır).
Vinchon, F., Lubart, T., Bartolotta, S., et al. (2023). Artificial Intelligence & Creativity: A Manifesto for Collaboration. The Journal of Creative Behavior, 57 (4), 472–484. DOI: 10.1002/jocb.597
Vyas, N., Kakade, S. & Barak, B. (2023). On Provable Copyright Protection for Generative Models https://doi.org/10.48550/arXiv.2302.10870
Yerushalmy, J. (2023). German publisher Axel Springer says journalists could be replaced by AI. (14 Haziran 2024 tarihinde https://www.theguardian.com/technology/2023/mar/01/german-publisher-axel-springer-says-journalists-could-be-replaced-by-ai adresinden ulaşılmıştır).
Zhai, S., Wang, W., Li, J., et al. (2024). Discovering Universal Semantic Triggers for Text-to-Image Synthesis. https://doi.org/10.48550/arXiv.2402.07562
Zhang, E., Wang, K., Xu, X., et al. (2023). Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models. https://doi.org/10.48550/arXiv.2303.17591
