[{"id":"c45d07d3-e458-491d-9007-35441fe749be","thread_id":"6007e725-7f58-4c7c-a889-24daf8ece3df","created_at":"2025-07-27T17:58:11Z","state":"done","prompt":"First, do research on diffusion models and give me a quick summary. Then, extract the text in the image and put it into a code block.","references_html":"<ol data-ref-list>\n \n <li>\n <a href=\"https://lilianweng.github.io/posts/2021-07-11-diffusion-models/\" >What are Diffusion Models? | Lil'Log</a>\n <span class=\"__domain-name\">lilianweng.github.io</span>\n <div class=\"trigger\" tabindex=\"0\">\n <span class=\"contribution\" style=\"background: linear-gradient(to right, var(--primary-70) , transparent )\"></span>\n <div class=\"tooltip top end tooltip_base tw-auto\">\n This percentage indicates how much a source has contributed to the answer.\n </div>\n </div>\n </li>\n \n <li>\n <a href=\"https://jiwasawa.github.io/blog/posts/difusion-basics/index.html\" >Introduction to Diffusion Models: From Core Concepts to Cutting-Edge ...</a>\n <span class=\"__domain-name\">jiwasawa.github.io</span>\n <div class=\"trigger\" tabindex=\"0\">\n <span class=\"contribution\" style=\"background: linear-gradient(to right, var(--primary-70) , transparent )\"></span>\n <div class=\"tooltip top end tooltip_base tw-auto\">\n This percentage indicates how much a source has contributed to the answer.\n </div>\n </div>\n </li>\n \n <li>\n <a href=\"https://assemblyai.com/blog/diffusion-models-for-machine-learning-introduction\" >Introduction to Diffusion Models for Machine Learning - AssemblyAI</a>\n <span class=\"__domain-name\">assemblyai.com</span>\n <div class=\"trigger\" tabindex=\"0\">\n <span class=\"contribution\" style=\"background: linear-gradient(to right, var(--primary-70) , transparent )\"></span>\n <div class=\"tooltip top end tooltip_base tw-auto\">\n This percentage indicates how much a source has contributed to the answer.\n </div>\n </div>\n </li>\n \n <li>\n <a href=\"https://arxiv.org/pdf/2209.00796v9\" >Diffusion Models: A Comprehensive Survey of Methods and Applications</a>\n <span class=\"__domain-name\">arxiv.org</span>\n <div class=\"trigger\" tabindex=\"0\">\n <span class=\"contribution\" style=\"background: linear-gradient(to right, var(--primary-70) , transparent )\"></span>\n <div class=\"tooltip top end tooltip_base tw-auto\">\n This percentage indicates how much a source has contributed to the answer.\n </div>\n </div>\n </li>\n \n <li>\n <a href=\"https://en.wikipedia.org/wiki/Diffusion_model\" >Diffusion model - Wikipedia</a>\n <span class=\"__domain-name\">en.wikipedia.org</span>\n <div class=\"trigger\" tabindex=\"0\">\n <span class=\"contribution\" style=\"background: linear-gradient(to right, var(--primary-70) , transparent )\"></span>\n <div class=\"tooltip top end tooltip_base tw-auto\">\n This percentage indicates how much a source has contributed to the answer.\n </div>\n </div>\n </li>\n \n <li>\n <a href=\"https://www.superannotate.com/blog/diffusion-models\" >Introduction to Diffusion Models for Machine Learning | SuperAnnotate</a>\n <span class=\"__domain-name\">www.superannotate.com</span>\n <div class=\"trigger\" tabindex=\"0\">\n <span class=\"contribution\" style=\"background: linear-gradient(to right, var(--primary-70) , transparent )\"></span>\n <div class=\"tooltip top end tooltip_base tw-auto\">\n This percentage indicates how much a source has contributed to the answer.\n </div>\n </div>\n </li>\n \n <li>\n <a href=\"https://arxiv.org/abs/2209.00796\" >Diffusion Models: A Comprehensive Survey of Methods and ... - arXiv</a>\n <span class=\"__domain-name\">arxiv.org</span>\n <div class=\"trigger\" tabindex=\"0\">\n <span class=\"contribution\" style=\"background: linear-gradient(to right, var(--primary-70) , transparent )\"></span>\n <div class=\"tooltip top end tooltip_base tw-auto\">\n This percentage indicates how much a source has contributed to the answer.\n </div>\n </div>\n </li>\n \n</ol>\n","references_md":"[^1]: [What are Diffusion Models? | Lil'Log](https://lilianweng.github.io/posts/2021-07-11-diffusion-models/)\n[^2]: [Introduction to Diffusion Models: From Core Concepts to Cutting-Edge ...](https://jiwasawa.github.io/blog/posts/difusion-basics/index.html)\n[^3]: [Introduction to Diffusion Models for Machine Learning - AssemblyAI](https://assemblyai.com/blog/diffusion-models-for-machine-learning-introduction)\n[^4]: [Diffusion Models: A Comprehensive Survey of Methods and Applications](https://arxiv.org/pdf/2209.00796v9)\n[^5]: [Diffusion model - Wikipedia](https://en.wikipedia.org/wiki/Diffusion_model)\n[^6]: [Introduction to Diffusion Models for Machine Learning | SuperAnnotate](https://www.superannotate.com/blog/diffusion-models)\n[^7]: [Diffusion Models: A Comprehensive Survey of Methods and ... - arXiv](https://arxiv.org/abs/2209.00796)\n","reply":"<details><summary>Searching with Kagi:<p></p>\n<ul>\n<li>diffusion models summary</li></ul></summary> Sources (10):<br>\n<strong>What are Diffusion Models? | Lil'Log</strong><br>\n<a href=\"https://lilianweng.github.io/posts/2021-07-11-diffusion-models/\" target=\"_blank\">https://lilianweng.github.io/posts/2021-07-11-diffusion-models/</a><br>\nJul 11, 2021 Diffusion models are inspired by non-equilibrium thermodynamics. They define a Markov chain of diffusion steps to slowly add random noise to data.\n\n<p><strong><a href=\"https://arxiv.org/abs/2404.07771\" target=\"_blank\">https://arxiv.org/abs/2404.07771</a>..</strong><br>\n<a href=\"https://arxiv.org/abs/2404.07771\" target=\"_blank\">https://arxiv.org/abs/2404.07771</a><br>\n In this paper, we review emerging applications of diffusion models, understanding their sample generation under various controls. Next, we overview the existing theories of diffusion models, covering their statistical properties and sampling capabilities.</p>\n<p><strong>Diffusion Models: A Comprehensive Survey of Methods and ... - arXiv</strong><br>\n<a href=\"https://arxiv.org/abs/2209.00796\" target=\"_blank\">https://arxiv.org/abs/2209.00796</a><br>\n Diffusion models have emerged as a powerful new family of deep generative models with record-breaking performance in many applications.</p>\n<p><strong>Diffusion Models: A Comprehensive Survey of Methods and Applications</strong><br>\n<a href=\"https://arxiv.org/pdf/2209.00796v9\" target=\"_blank\">https://arxiv.org/pdf/2209.00796v9</a><br>\n Difusion models have emerged as a powerful new family of deep generative models with record-breaking performance in many applications, including image synthesis, video generation, and molecule design.</p>\n<p><strong>Interpretable Diffusion via Information</strong><br>\n<a href=\"https://arxiv.org/html/2310.07972v3\" target=\"_blank\">https://arxiv.org/html/2310.07972v3</a><br>\nApr 13, 2025 We find that on the ARO benchmark (Yuksekgonul et al., 2022 ) , diffusion models are significantly underestimated due to sub-optimal alignment ...</p>\n<p><strong>An Overview of Diffusion Models: Applications, Guided Generation ...</strong><br>\n<a href=\"https://arxiv.org/html/2404.07771v1\" target=\"_blank\">https://arxiv.org/html/2404.07771v1</a><br>\n In this paper, we review emerging applications of diffusion models, understanding their sample generation under various controls. Next, we overview the existing theories of diffusion models, covering their statistical properties and sampling capabilities.</p>\n<p><strong>Diffusion model - Wikipedia</strong><br>\n<a href=\"https://en.wikipedia.org/wiki/Diffusion_model\" target=\"_blank\">https://en.wikipedia.org/wiki/Diffusion_model</a><br>\n Diffusion models are mainly used for computer vision tasks, including image denoising, inpainting, super-resolution, image generation, and video generation.</p>\n<p><strong>Introduction to Diffusion Models for Machine Learning - AssemblyAI</strong><br>\n<a href=\"https://assemblyai.com/blog/diffusion-models-for-machine-learning-introduction\" target=\"_blank\">https://assemblyai.com/blog/diffusion-models-for-machine-learning-introduction</a><br>\nMay 12, 2022 Diffusion Models are generative models, meaning that they are used to generate data similar to the data on which they are trained.</p>\n<p><strong>Introduction to Diffusion Models: From Core Concepts to Cutting-Edge ...</strong><br>\n<a href=\"https://jiwasawa.github.io/blog/posts/difusion-basics/index.html\" target=\"_blank\">https://jiwasawa.github.io/blog/posts/difusion-basics/index.html</a><br>\n Diffusion models represent a powerful and flexible class of generative models based on systematically destroying data structure with noise and then learning to reverse the process.</p>\n<p><strong>Introduction to Diffusion Models for Machine Learning | SuperAnnotate</strong><br>\n<a href=\"https://www.superannotate.com/blog/diffusion-models\" target=\"_blank\">https://www.superannotate.com/blog/diffusion-models</a><br>\nFeb 28, 2025 Diffusion models are advanced machine learning algorithms that uniquely generate high-quality data by progressively adding noise to a dataset ...</p></details>\n<p><em>Gathering key details from</em> <em>image.png</em></p><p>I'll provide a summary of diffusion models based on the research, and then I'll attempt to extract text from the image.</p>\n<p>Summary of Diffusion Models:</p>\n<p>Diffusion models are an innovative class of deep generative machine learning models with several key characteristics:</p>\n<ul>\n<li>Inspiration: Derived from non-equilibrium thermodynamics <sup class=\"trigger\"><a href=\"https://lilianweng.github.io/posts/2021-07-11-diffusion-models/\">1</a><span class=\"tooltip top center citation\"><a href=\"https://lilianweng.github.io/posts/2021-07-11-diffusion-models/#:~:text=Diffusion%20models,to%20data.\"><span class=\"source\"><img alt=\"\" src=\"/proxy/favicons?c=lQ30-cxV_E2Vdyqb6hrbuYgId7qu7hecjO7o1T7aqQDFc7XpcLttL1Ryala-GNHJiR_POeWoSqVnjce3BUW24vSiEtoNmVVLXlAe8h9cbequsidzlnKIAoMsyBUH2wxm\"> lilianweng.github.io</span><b>1. What are Diffusion Models? | Lil'Log</b></a><span class=\"passage\">Diffusion models are inspired by non-equilibrium thermodynamics. They define a Markov chain of diffusion steps to slowly add random noise to data.</span></span></sup></li>\n<li>Core Mechanism: They systematically add random noise to data and then learn to reverse this process <sup class=\"trigger\"><a href=\"https://jiwasawa.github.io/blog/posts/difusion-basics/index.html\">2</a><span class=\"tooltip top center citation\"><a href=\"https://jiwasawa.github.io/blog/posts/difusion-basics/index.html#:~:text=Diffusion%20models,the%20process.\"><span class=\"source\"><img alt=\"\" src=\"/proxy/favicons?c=lQ30-cxV_E2Vdyqb6hrbuYgId7qu7hecjO7o1T7aqQDFc7XpcLttL1Ryala-GNHJNobgAGiC49r-TzW77K38R16DvlRemCzLhT1Iqu1kn7I%3D\"> jiwasawa.github.io</span><b>2. Introduction to Diffusion Models: From Core Concepts to Cutting-Edge ...</b></a><span class=\"passage\">Diffusion models represent a powerful and flexible class of generative models based on systematically destroying data structure with noise and then learning to reverse the process.</span></span></sup></li>\n<li>Purpose: Generate high-quality data similar to their training dataset <sup class=\"trigger\"><a href=\"https://assemblyai.com/blog/diffusion-models-for-machine-learning-introduction\">3</a><span class=\"tooltip top center citation\"><a href=\"https://assemblyai.com/blog/diffusion-models-for-machine-learning-introduction#:~:text=Diffusion%20Models,are%20trained.\"><span class=\"source\"><img alt=\"\" src=\"/proxy/favicons?c=lQ30-cxV_E2Vdyqb6hrbuYgId7qu7hecjO7o1T7aqQDFc7XpcLttL1Ryala-GNHJyjhCHSYeOhiuUs3ubRFMRXF8naKYpkp0an99YSI3uyQ%3D\"> assemblyai.com</span><b>3. Introduction to Diffusion Models for Machine Learning - AssemblyAI</b></a><span class=\"passage\">Diffusion Models are generative models, meaning that they are used to generate data similar to the data on which they are trained.</span></span></sup></li>\n<li>Applications:\n<ul>\n<li>Image synthesis</li>\n<li>Video generation</li>\n<li>Molecule design</li>\n<li>Image denoising</li>\n<li>Inpainting</li>\n<li>Super-resolution <sup class=\"trigger\"><a href=\"https://arxiv.org/pdf/2209.00796v9\">4</a><span class=\"tooltip top center citation\"><a href=\"https://arxiv.org/pdf/2209.00796v9#:~:text=Difusion%20models,molecule%20design.\"><span class=\"source\"><img alt=\"\" src=\"/proxy/favicons?c=lQ30-cxV_E2Vdyqb6hrbuYgId7qu7hecjO7o1T7aqQDFc7XpcLttL1Ryala-GNHJpx390kEVKKSm35Jlb8zJxiE_gOHMoXvEd2qVZP40_zA%3D\"> arxiv.org</span><b>4. Diffusion Models: A Comprehensive Survey of Methods and Applications</b></a><span class=\"passage\">Difusion models have emerged as a powerful new family of deep generative models with record-breaking performance in many applications, including image synthesis, video generation, and molecule design.</span></span></sup><sup class=\"trigger\"><a href=\"https://en.wikipedia.org/wiki/Diffusion_model\">5</a><span class=\"tooltip top center citation\"><a href=\"https://en.wikipedia.org/wiki/Diffusion_model#:~:text=Diffusion%20models,video%20generation.\"><span class=\"source\"><img alt=\"\" src=\"/proxy/favicons?c=lQ30-cxV_E2Vdyqb6hrbuYgId7qu7hecjO7o1T7aqQDFc7XpcLttL1Ryala-GNHJuWyqBKj8MT4ZBZYnSH4uzuNLNQdbiyz2-gUDYI6hGs0%3D\"> en.wikipedia.org</span><b>5. Diffusion model - Wikipedia</b></a><span class=\"passage\">Diffusion models are mainly used for computer vision tasks, including image denoising, inpainting, super-resolution, image generation, and video generation.</span></span></sup></li>\n</ul>\n</li>\n</ul>\n<p>Key Distinguishing Feature: Diffusion models uniquely generate data by progressively adding noise to a dataset and then learning to reconstruct the original data <sup class=\"trigger\"><a href=\"https://www.superannotate.com/blog/diffusion-models\">6</a><span class=\"tooltip top center citation\"><a href=\"https://www.superannotate.com/blog/diffusion-models#:~:text=Diffusion%20models,a%20dataset\"><span class=\"source\"><img alt=\"\" src=\"/proxy/favicons?c=lQ30-cxV_E2Vdyqb6hrbuYgId7qu7hecjO7o1T7aqQDFc7XpcLttL1Ryala-GNHJ3YOWpDV7M75W5AknSyvdIi4xPWIcptZ_fE-82yS-5t1u0P8AYlKUnZDouTvTddBS\"> www.superannotate.com</span><b>6. Introduction to Diffusion Models for Machine Learning | SuperAnnotate</b></a><span class=\"passage\">Diffusion models are advanced machine learning algorithms that uniquely generate high-quality data by progressively adding noise to a dataset ...</span></span></sup> <sup class=\"trigger\"><a href=\"https://arxiv.org/abs/2209.00796\">7</a><span class=\"tooltip top center citation\"><a href=\"https://arxiv.org/abs/2209.00796#:~:text=Diffusion%20models,many%20applications.\"><span class=\"source\"><img alt=\"\" src=\"/proxy/favicons?c=lQ30-cxV_E2Vdyqb6hrbuYgId7qu7hecjO7o1T7aqQDFc7XpcLttL1Ryala-GNHJpx390kEVKKSm35Jlb8zJxiE_gOHMoXvEd2qVZP40_zA%3D\"> arxiv.org</span><b>7. Diffusion Models: A Comprehensive Survey of Methods and ... - arXiv</b></a><span class=\"passage\">Diffusion models have emerged as a powerful new family of deep generative models with record-breaking performance in many applications.</span></span></sup></p>\n<p>Regarding the image text extraction, I apologize, but I cannot see any image in the current context. Could you please upload the image again? Without a visible image, I cannot extract its text and place it in a code block.</p>","md":"I'll provide a summary of diffusion models based on the research, and then I'll attempt to extract text from the image.\n\nSummary of Diffusion Models:\n\nDiffusion models are an innovative class of deep generative machine learning models with several key characteristics:\n\n- Inspiration: Derived from non-equilibrium thermodynamics [^1]\n- Core Mechanism: They systematically add random noise to data and then learn to reverse this process [^2]\n- Purpose: Generate high-quality data similar to their training dataset [^3]\n- Applications: \n - Image synthesis\n - Video generation\n - Molecule design\n - Image denoising\n - Inpainting\n - Super-resolution [^4][^5]\n\nKey Distinguishing Feature: Diffusion models uniquely generate data by progressively adding noise to a dataset and then learning to reconstruct the original data [^6] [^7]\n\nRegarding the image text extraction, I apologize, but I cannot see any image in the current context. Could you please upload the image again? Without a visible image, I cannot extract its text and place it in a code block.","profile":{"id":null,"name":null,"model":"claude-3-haiku","model_name":"Claude 3.5 Haiku","model_provider":"anthropic","model_input_limit":1000000,"recommended":false,"model_info":"\n\n<div class=\"heading\">\n <div class=\"left\">\n \n \n \n <svg class=\"model-icon icon-xs\" data-model=\"claude-3-haiku\">\n <use href=\"#icon-anthropic\"/>\n </svg>\n <h3>\n \n <span class=\"model-provider\">Anthropic</span>\n \n <span class=\"model-name\">none</span>\n </h3>\n </div>\n\n \n <a href=\"/settings?p=custom_assistant&id=none\">\n <svg class=\"icon-sm edit-icon\" aria-hidden=\"true\">\n <use href=\"#icon-pencil\" />\n </svg>\n </a>\n \n</div>\n\n\n<hr />\n<ul>\n \n <li>\n <h4>\n Cost\n </h4>\n <span>\n \n \n \n\n \n \n \n\n <div class=\"score\" aria-label=\"Moderate\">\n \n \n \n \n <div aria-hidden=\"true\" class=\"filled-cost --medium\"></div>\n \n \n \n \n \n <div aria-hidden=\"true\" class=\"filled-cost --medium\"></div>\n \n \n \n \n \n <div aria-hidden=\"true\" class=\"filled-cost --medium\"></div>\n \n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-cost --medium\"></div>\n \n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-cost --medium\"></div>\n \n \n </div>\n</span>\n </li>\n <li>\n <h4>\n Quality\n </h4>\n <span>\n \n \n \n\n \n \n \n\n <div class=\"score\" aria-label=\"Rated 1 out of 5\">\n \n \n \n <div aria-hidden=\"true\" class=\"filled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-bubble\"></div>\n \n \n </div>\n</span>\n </li>\n <li>\n <h4>\n Speed\n </h4>\n <span>\n \n \n \n\n \n \n \n\n <div class=\"score\" aria-label=\"Rated 4 out of 5\">\n \n \n \n <div aria-hidden=\"true\" class=\"filled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"filled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"filled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"filled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-bubble\"></div>\n \n \n </div>\n</span>\n </li>\n <li>\n <h4>\n Privacy\n </h4>\n <span>\n \n \n \n\n \n \n \n\n <div class=\"score\" aria-label=\"Rated 2 out of 5\">\n \n \n \n <div aria-hidden=\"true\" class=\"filled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"filled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-bubble\"></div>\n \n \n \n \n <div aria-hidden=\"true\" class=\"unfilled-bubble\"></div>\n \n \n </div>\n</span>\n </li>\n <li class=\"divider\" aria-hidden=\"true\"></li>\n <li>\n <h4>Last updated</h4>\n <span class=\"date\">\n \n \n 2024-10-22\n \n \n </span>\n </li>\n \n</ul>\n","model_provider_name":"Anthropic","internet_access":true,"personalizations":true,"shortcut":null,"is_default_profile":false},"metadata":"<li>\n <span class=\"attribute\">Model</span>\n <span class=\"value\"><i class=\"icon-xs mr-2\"><svg><use href=\"#icon-anthropic\"></use></svg></i>Claude 3.5 Haiku</span>\n</li>\n<li>\n <span class=\"attribute\">Version</span>\n <span class=\"value\">claude-3-5-haiku-20241022</span>\n</li>\n<li>\n <span class=\"attribute\">Speed (tok/s)</span>\n <span class=\"value\">50</span>\n</li>\n<li>\n <span class=\"attribute\">Tokens</span>\n <span class=\"value\">2,209</span>\n</li>\n<li>\n <span class=\"attribute\">Cost / Total ($)</span>\n <span class=\"value\">0.007 / 0.007</span>\n</li>\n<li>\n <span class=\"attribute\">End to end time (s)</span>\n <span class=\"value\">9.18</span>\n</li>\n\n","documents":[{"id":"ec00a6ca-0216-4d3a-865e-3b7124a0e069","name":"image.png","type":"file","mime":"image/png","data":"data:image/png;base64,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"}]}]