#diffusionmodels — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #diffusionmodels, aggregated by home.social.
-
🔍 Look, another #AI #research paper where the title is longer than the list of authors! 🤯 Enjoy this thrilling saga where 15 scientists and their army of acronyms perform a meticulous autopsy on diffusion models, only to conclude that numbers are indeed bigger than other numbers. 📈 Spoiler: the authors' names are as long as the codebase isn't. 🙄
https://introspective-diffusion.github.io/ #DiffusionModels #ScientificPaper #HackerNews #TechHumor #Academia #HackerNews #ngated -
Hamilton-Jacobi-Bellman Equation: Reinforcement Learning and Diffusion Models
https://dani2442.github.io/posts/continuous-rl/
#HackerNews #HamiltonJacobiBellman #ReinforcementLearning #DiffusionModels #AIResearch #MachineLearning
-
Steerling-8B, the first interpretable model that can trace any token it generates to its input context, concepts a human can understand, and its training data.
https://www.guidelabs.ai/post/steerling-8b-base-model-release/
-
Steerling-8B, the first interpretable model that can trace any token it generates to its input context, concepts a human can understand, and its training data.
https://www.guidelabs.ai/post/steerling-8b-base-model-release/
-
Steerling-8B, the first interpretable model that can trace any token it generates to its input context, concepts a human can understand, and its training data.
https://www.guidelabs.ai/post/steerling-8b-base-model-release/
-
Steerling-8B, the first interpretable model that can trace any token it generates to its input context, concepts a human can understand, and its training data.
https://www.guidelabs.ai/post/steerling-8b-base-model-release/
-
Steerling-8B, the first interpretable model that can trace any token it generates to its input context, concepts a human can understand, and its training data.
https://www.guidelabs.ai/post/steerling-8b-base-model-release/
-
STARFlow-V normalizing flows bring faster, stable long clips across text, image, and video, using sliding windows and 640x480/16fps limits.
-
Apple STARFlow-V video model uses normalizing flows for stable, fast 640x480/16fps clips and 30s demos via sliding window, with three modes.
-
🚀✨ Behold, the groundbreaking Qwen-Image-Layered: Now you can finally edit images like an onion, one tear-inducing layer at a time! 😂 Who knew #transparency in diffusion models would be such a *layered* joke? 🤡
https://huggingface.co/papers/2512.15603 #QwenImageLayered #ImageEditing #DiffusionModels #TechHumor #HackerNews #ngated -
Meta V-JEPA model learns physical intuition from everyday videos, signaling new progress for self-supervised video AI and safer robotics.
https://www.aistory.news/generative-ai/meta-v-jepa-model-shows-intuitive-physics-from-video/
-
NeurIPS 2025 revealed a turning point: the top-four papers argue that bigger isn't always better. From revamped attention mechanisms to leaner diffusion models and smarter RL benchmarks, researchers are redefining performance metrics. Curious how the AI community is shifting focus? Dive into the highlights and see what's next for open-source innovation. #NeurIPS2025 #LLM #DiffusionModels #BenchmarkRevisions
🔗 https://aidailypost.com/news/neurips-2025-top-4-papers-highlight-shift-from-bigger-models-limits
-
NeurIPS 2025 revealed a turning point: the top-four papers argue that bigger isn't always better. From revamped attention mechanisms to leaner diffusion models and smarter RL benchmarks, researchers are redefining performance metrics. Curious how the AI community is shifting focus? Dive into the highlights and see what's next for open-source innovation. #NeurIPS2025 #LLM #DiffusionModels #BenchmarkRevisions
🔗 https://aidailypost.com/news/neurips-2025-top-4-papers-highlight-shift-from-bigger-models-limits
-
NeurIPS 2025 revealed a turning point: the top-four papers argue that bigger isn't always better. From revamped attention mechanisms to leaner diffusion models and smarter RL benchmarks, researchers are redefining performance metrics. Curious how the AI community is shifting focus? Dive into the highlights and see what's next for open-source innovation. #NeurIPS2025 #LLM #DiffusionModels #BenchmarkRevisions
🔗 https://aidailypost.com/news/neurips-2025-top-4-papers-highlight-shift-from-bigger-models-limits
-
NeurIPS 2025 revealed a turning point: the top-four papers argue that bigger isn't always better. From revamped attention mechanisms to leaner diffusion models and smarter RL benchmarks, researchers are redefining performance metrics. Curious how the AI community is shifting focus? Dive into the highlights and see what's next for open-source innovation. #NeurIPS2025 #LLM #DiffusionModels #BenchmarkRevisions
🔗 https://aidailypost.com/news/neurips-2025-top-4-papers-highlight-shift-from-bigger-models-limits
-
DiverGen uses accurate SAM-based annotation methods, generative models, and a variety of prompts to improve AI segmentation. https://hackernoon.com/divergen-proves-ai-models-learn-better-with-variety #diffusionmodels
-
DiverGen's ablation research show how a variety of prompts, models, and categories improve AI accuracy above and above what is currently possible. https://hackernoon.com/the-impact-of-generative-data-diversity-on-ai-accuracy #diffusionmodels
-
DiverGen demonstrates that superior instance segmentation performance is driven by data diversity rather than quantity. https://hackernoon.com/data-diversity-matters-more-than-data-quantity-in-ai #diffusionmodels
-
DiverGen reduces distribution bias in instance segmentation by diversifying generative data among models, prompts, and categories. https://hackernoon.com/how-generative-data-expands-ais-understanding-of-the-real-world #diffusionmodels
-
Latest Deep Learning Breakthroughs Shaping Our World
https://jivoice.com/deep-learning-breakthroughs/
#artificialintelligenceadvancements #transformerarchitecture #AlphaFold #diffusionmodels #computervision #neuralnetworktechnology #AIinhealthcare #naturallanguageprocessing #generativeAI #deeplearningbreakthroughs
-
Latest Deep Learning Breakthroughs Shaping Our World
https://jivoice.com/deep-learning-breakthroughs/
#artificialintelligenceadvancements #transformerarchitecture #AlphaFold #diffusionmodels #computervision #neuralnetworktechnology #AIinhealthcare #naturallanguageprocessing #generativeAI #deeplearningbreakthroughs
-
Latest Deep Learning Breakthroughs Shaping Our World
https://jivoice.com/deep-learning-breakthroughs/
#artificialintelligenceadvancements #transformerarchitecture #AlphaFold #diffusionmodels #computervision #neuralnetworktechnology #AIinhealthcare #naturallanguageprocessing #generativeAI #deeplearningbreakthroughs
-
Latest Deep Learning Breakthroughs Shaping Our World
https://jivoice.com/deep-learning-breakthroughs/
#artificialintelligenceadvancements #transformerarchitecture #AlphaFold #diffusionmodels #computervision #neuralnetworktechnology #AIinhealthcare #naturallanguageprocessing #generativeAI #deeplearningbreakthroughs
-
Latest Deep Learning Breakthroughs Shaping Our World
https://jivoice.com/deep-learning-breakthroughs/
#artificialintelligenceadvancements #transformerarchitecture #AlphaFold #diffusionmodels #computervision #neuralnetworktechnology #AIinhealthcare #naturallanguageprocessing #generativeAI #deeplearningbreakthroughs
-
"The new diffusion-based image generator works by first using a digital encoder (that has been trained on publicly available datasets) to create the static that will ultimately make the picture. This requires a small amount of energy. Then, a liquid crystal screen known as a spatial light modulator (SLM) imprints this pattern onto a laser beam. The beam is then passed through a second decoding SLM, which turns the pattern in the laser into the final image.
Unlike conventional AI, which relies on millions of computer calculations, this process uses light to do all the heavy lifting. Consequently, the system uses almost no power. "Our optical generative models can synthesize countless images with almost no computing power, offering a scalable and energy-efficient alternative to digital AI models," said Shiqi Chen, lead author.
The researchers tested their system on various images used to train AI models, including those of celebrities and butterflies, as well as full-color pictures in the style of Dutch painter Vincent Van Gogh.
The results were comparable to those of conventional image generators, but were created with much less energy. This breakthrough has the potential to reduce the carbon footprint of AI-generated content significantly."
https://techxplore.com/news/2025-08-ai-breakthrough-power-images.html
-
Reusing Computation in Text-to-Image Diffusion for Efficient Image Generation
https://arxiv.org/abs/2508.21032
#HackerNews #Reusing #Computation #in #Text-to-Image #Diffusion #for #Efficient #Image #Generation #AI #Innovation #ImageGeneration #DiffusionModels #TechResearch #ComputationEfficiency
-
Reusing Computation in Text-to-Image Diffusion for Efficient Image Generation
https://arxiv.org/abs/2508.21032
#HackerNews #Reusing #Computation #in #Text-to-Image #Diffusion #for #Efficient #Image #Generation #AI #Innovation #ImageGeneration #DiffusionModels #TechResearch #ComputationEfficiency
-
Reusing Computation in Text-to-Image Diffusion for Efficient Image Generation
https://arxiv.org/abs/2508.21032
#HackerNews #Reusing #Computation #in #Text-to-Image #Diffusion #for #Efficient #Image #Generation #AI #Innovation #ImageGeneration #DiffusionModels #TechResearch #ComputationEfficiency
-
Reusing Computation in Text-to-Image Diffusion for Efficient Image Generation
https://arxiv.org/abs/2508.21032
#HackerNews #Reusing #Computation #in #Text-to-Image #Diffusion #for #Efficient #Image #Generation #AI #Innovation #ImageGeneration #DiffusionModels #TechResearch #ComputationEfficiency
-
Reusing Computation in Text-to-Image Diffusion for Efficient Image Generation
https://arxiv.org/abs/2508.21032
#HackerNews #Reusing #Computation #in #Text-to-Image #Diffusion #for #Efficient #Image #Generation #AI #Innovation #ImageGeneration #DiffusionModels #TechResearch #ComputationEfficiency
-
"To generate images, diffusion models use a process known as denoising. They convert an image into digital noise (an incoherent collection of pixels), then reassemble it. It’s like repeatedly putting a painting through a shredder until all you have left is a pile of fine dust, then patching the pieces back together. For years, researchers have wondered: If the models are just reassembling, then how does novelty come into the picture? It’s like reassembling your shredded painting into a completely new work of art.
Now two physicists have made a startling claim: It’s the technical imperfections in the denoising process itself that leads to the creativity of diffusion models. In a paper(opens a new tab) that will be presented at the International Conference on Machine Learning 2025, the duo developed a mathematical model of trained diffusion models to show that their so-called creativity is in fact a deterministic process — a direct, inevitable consequence of their architecture.
By illuminating the black box of diffusion models, the new research could have big implications for future AI research — and perhaps even for our understanding of human creativity. “The real strength of the paper is that it makes very accurate predictions of something very nontrivial,” said Luca Ambrogioni(opens a new tab), a computer scientist at Radboud University in the Netherlands."
https://www.quantamagazine.org/researchers-uncover-hidden-ingredients-behind-ai-creativity-20250630/
-
🔥 AI just leveled up! Apple’s **DiffuCoder-7B** is here, rocking the world of **Diffusion Large Language Models (DLLMs)**. Faster, smarter, and rewriting the rules of gen AI! 🚀
🌐 Learn why diffusion is the future we deserve: https://smsk.dev/2025/07/06/apples-here-with-diffusion-dllms/
-
"There are a number of diffusion models out there, but I have tended to use Midjourney, which has been around longer than many other AI tools. Using Midjourney allows us to see how diffusion models have developed over time, as you can see with the simple prompt “otter on a plane using wifi” (for every image and video in this post, I pick the best out of the first four images generated). We go from melted fur at the start of 2022 to a visible otter (with too many fingers and a weird keyboard) at the end of that year. In 2023, we get a photorealistic otter, but still a weird keyboard and plane windows. In 2024, the lighting and positioning become better, and by 2025 we have excellent photorealism.
But what makes diffusion models interesting is not their increasing ability to make photorealistic images, but rather the fact that they can create images in various styles. This cuts to the heart of why AI image generation is so controversial, as many AI models are trained on images from throughout the web, including copyrighted work, and can thus replicate images in the style of living artists without their permission or compensation."
#AU #GenerativeAI #GeneratedImages #DiffusionModels #OpenWeight
-
🎨🤯 Ah yes, diffusion models are "simple," if you have a PhD in rocket science and a decoder ring. Forget hobbling through incomprehensible papers—just squint hard enough, and these pixel-pushing wizards will reveal themselves to you, like a magic eye poster from 1993. 😂📚
https://www.seangoedecke.com/diffusion-models-explained/ #diffusionmodels #rocketScience #pixelArt #magicEye #humor #HackerNews #ngated -
🎨🤯 Ah yes, diffusion models are "simple," if you have a PhD in rocket science and a decoder ring. Forget hobbling through incomprehensible papers—just squint hard enough, and these pixel-pushing wizards will reveal themselves to you, like a magic eye poster from 1993. 😂📚
https://www.seangoedecke.com/diffusion-models-explained/ #diffusionmodels #rocketScience #pixelArt #magicEye #humor #HackerNews #ngated -
🎨🤯 Ah yes, diffusion models are "simple," if you have a PhD in rocket science and a decoder ring. Forget hobbling through incomprehensible papers—just squint hard enough, and these pixel-pushing wizards will reveal themselves to you, like a magic eye poster from 1993. 😂📚
https://www.seangoedecke.com/diffusion-models-explained/ #diffusionmodels #rocketScience #pixelArt #magicEye #humor #HackerNews #ngated -
🎨🤯 Ah yes, diffusion models are "simple," if you have a PhD in rocket science and a decoder ring. Forget hobbling through incomprehensible papers—just squint hard enough, and these pixel-pushing wizards will reveal themselves to you, like a magic eye poster from 1993. 😂📚
https://www.seangoedecke.com/diffusion-models-explained/ #diffusionmodels #rocketScience #pixelArt #magicEye #humor #HackerNews #ngated -
Block Diffusion: Interpolating Autoregressive and Diffusion Language Models
#HackerNews #BlockDiffusion #Autoregressive #LanguageModels #DiffusionModels #MachineLearning #AIResearch
-
In today's episode of "How to Make #Gibberish Sound Smart," we dive into an article that bravely attempts to bridge the vast chasm between #autoregressive models and diffusion models with absolutely zero context 🧠🔍. Meanwhile, Hugging Face is here to remind us that no amount of tech jargon can hide the fact they're just as desperate for your money as your local neighborhood pyramid scheme 💸🎩.
https://arxiv.org/abs/2503.09573 #techjargon #diffusionmodels #HuggingFace #HackerNews #ngated -
Block Diffusion: Interpolating Between Autoregressive and Diffusion Models
https://arxiv.org/abs/2503.09573
#HackerNews #Block #Diffusion #Autoregressive #DiffusionModels #MachineLearning #AIResearch #Interpolation
-
MIT 6.S184: Introduction to Flow Matching and Diffusion Models — https://diffusion.csail.mit.edu
#HackerNews #MIT #6S184 #Introduction #FlowMatching #DiffusionModels -
Microsoft’s New BioEmu-1 AI Model Can Predict How Proteins Move and Change
#AI #BioEmu1 #Microsoft #MicrosoftResearch #DiffusionModels #BioScience #Science #AIResearch #Alphafold #BigTech
-
Revolutionizing Video Generation: Meet Diffusion as Shader (DaS)
In a groundbreaking development, researchers unveil Diffusion as Shader (DaS), a game-changing approach that empowers users to control video generation like never before. By harnessing 3D tracking, Da...
https://news.lavx.hu/article/revolutionizing-video-generation-meet-diffusion-as-shader-das
#news #tech #DiffusionModels #3DVideoControl #AIContentCreation
-
AI Lecture on 11 March 2024, 18:15—19:45 with Prof. Dr. Jong Chul Ye from KAIST, exploring advancements in solving inverse problems with diffusion models, including 3D extensions and guidance by text prompts. Location: Theresienstraße 39, Room B 006. Open to the public. #AI #DiffusionModels #InverseProblems
https://www.ai-news.lmu.de/guestlecture/ -
OpenAI’s new generative tool Sora could revolutionize marketing and content creation
#AI #ArtificialIntelligence #OpenAI #Sora #TextToVideo #Innovation #ContentCreation #Marketing #Education #DigitalChallenges #Marketing #GenerativeAI #DiffusionModels #ArtificialIntelligenceEthics
https://the-14.com/openais-new-generative-tool-sora-could-revolutionize-marketing-and-content-creation/ -
Latest #AI #Freakout Dept: So yeah. #OpenAI #Sora . Should this be surprising ? From what #DiffusionModels are and the fact that video is simply a sequence of frames ? Try utterly inevitable. All the bits to do the heavy lifting are already several years old. So why the meltdowns ? Part of it is cognitive boundaries in how we think about images becoming a thing that walks, talks and dances around the room...like real memories instead of utter fictions.Context is everything
-
Alibaba Animate Anyone ersetzt TikTok jetzt mit KI
#AnimateAnyone #CharacterAnimation #DiffusionModels #Alibaba #ImageToVideoSynthesis #AIAnimation #DigitalArtTechnology #MachineLearningArt #AIinEntertainment #TechnologyInnovation
-
I'm not at @websummit this year as I'm busy on various writing projects, but pleased to see AI in fashion mentioned at the opening ceremony last night.
It's excellent timing - here's my new in-depth piece for the @eveningstandard today on the viability of AI models & virtual try-on tech 🛍️: https://www.standard.co.uk/news/tech/ai-fashion-models-virtual-tryon-diffusion-models-b1120382.html
#AI #aiimages #diffusionmodels #virtualtryon #retail #AImodels
-
Teaching a Robot to Hallucinate - Training robots to execute tasks in the real world requires data — the more, the b... - https://hackaday.com/2023/02/26/teaching-a-robot-to-hallucinate/ #artificialintelligence #machinehallucination #diffusionmodels #imagegenerators #trainingdata #robotshacks #robotics #imagen #ai