sayzard3d ago

fly51fly (@fly51fly)

확산 모델(diffusion models)의 일반화 성질을 다시 생각해야 한다는 연구 논문이 공개되었습니다. 생성형 AI의 핵심 기술인 diffusion model에 대한 이론적 이해를 재정비하는 내용으로, 모델 해석과 연구 방향에 중요한 시사점을 줄 수 있습니다.

https://x.com/fly51fly/status/2052865384320237938

#diffusionmodels #generativeai #research #machinelearning #ai

fly51fly (@fly51fly) on X

[LG] Understanding diffusion models requires rethinking (again) generalization P Marion, Y Wu [PSL Research University & Sorbonne University] (2026) https://t.co/xCh86Voo3d

X (formerly Twitter)
sayzardMay 5

fly51fly (@fly51fly)

의사결정 트리와 확산 모델을 하나의 틀로 통합하는 새로운 연구를 소개합니다. 트리를 플로우로 바꾸고 다시 되돌리는 관점으로 두 모델 계열의 연결을 제시합니다.

https://x.com/fly51fly/status/2051418728764723228

#decisiontrees #diffusionmodels #machinelearning #research #ai

fly51fly (@fly51fly) on X

[LG] Trees to Flows and Back: Unifying Decision Trees and Diffusion Models S N Ramachandran, S Sra [ Technical University of Munich] (2026) https://t.co/kTzywnz25A

X (formerly Twitter)
PsychoticSheepApr 19
A stylized cat in manga / ink-style with black-and-white fur, sharp brush lines, and striking green eyes. The background adds subtle green and blue accents, giving it a modern creative vibe.

Z-Image is really cool for creations like this.

#CatArt #MangaStyle #DigitalArt #AIArt #GreenEyes #BlackAndWhite #CreativeTools #ZImage #GenerativeAI #CreativeAI #DiffusionModels #Localference #ComputerVision #InkStyle
PsychoticSheepApr 16
Once again experimented with my local Z-Image setup, this time exploring different scenes in a monochrome manga / ink-style aesthetic.

#InkStyle #MangaAesthetic #AIArt #GenerativeAI #CreativeAI #Pixelfed #MangaStyle #Monochrome #DigitalArt #LocalInference #ZImage #DiffusionModels #MonochromeRendering #ComputerVision
N-gated Hacker NewsApr 14
🔍 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
I-DLM: Introspective Diffusion Language Models

First DLM to match AR quality with 2.9-4.1x throughput. Code, models, and interactive analysis.

Pure AcetoneApr 5

Вот 23 хэштега под твой пост (с уклоном в ИИ, VFX, dev и немного хайпа):

#AI #ArtificialIntelligence #GenerativeAI #VideoAI #MachineLearning #DeepLearning #ComputerVision #VFX #CGI #VideoEditing #ContentCreation #NeuralNetworks #DiffusionModels #CogVideoX #SAM2 #GeminiAI #OpenSource #GitHub #TechNews #FutureTech #AItools #AIVideo #DigitalReality

https://t.me/+7-bs1d6x6hRlNDli

Expanse

https://matrix.to/#/#Expanse:matrix.org

Telegram
Hacker NewsMar 30

Hamilton-Jacobi-Bellman Equation: Reinforcement Learning and Diffusion Models

https://dani2442.github.io/posts/continuous-rl/

#HackerNews #HamiltonJacobiBellman #ReinforcementLearning #DiffusionModels #AIResearch #MachineLearning

Hamilton-Jacobi-Bellman Equation: Reinforcement Learning and Diffusion Models

Why the HJB is Bellman's equation in continuous time, why continuous time matters, and how to solve the resulting control problem with neural policy iteration.

dani2442's Blog
Max BonhommeMar 28

Nouveau papier sur HAL ! https://hal.science/hal-05564515v1

"Réel ou symbolique. Le bruit des images techniques" dans Georges Didi-Huberman et Peter Geimer (dir.), Message ou bruit ? De la lisibilité des images, Paris, Éditions de la MSH / Centre allemand d’histoire de l’art (coll. Passages), à paraître 2026/2027.

Où il est question du bruit de Perlin, du bruit en photographie et de l'IA.

C'est le pre-print (sans les images pour l'instant).

#mediastudies #noise #cgihistory #diffusionmodels

sayzardMar 27

Saar Huberman (@HubermanSaar)

‘Image Generation from Contextually Contradictory Prompts’ 연구가 CVPR 2026에 채택됐다. 확산 모델이 겉보기엔 단순한 프롬프트에서도 지시를 무시하는 이유를 분석하고, 이를 훈련 없이 해결하는 방법을 제시한다.

https://x.com/HubermanSaar/status/2037176218886308075

#cvpr2026 #diffusionmodels #imagegeneration #research

Saar Huberman (@HubermanSaar) on X

Excited to share that our work, “Image Generation from Contextually Contradictory Prompts,” was accepted to CVPR 2026 🎉 Diffusion models fail on some seemingly simple prompts. Why do they ignore what you asked for? We show why, and how to fix it with a simple, training-free

X (formerly Twitter)
sayzardFeb 28

TechRadar (@techradar)

Nano Banana 2의 이미지 프롬프트 예시들이 공개되어 해당 버전의 성능 향상과 개선된 이미지 생성 능력을 보여줍니다. (이미지 프롬프트를 통해 새 모델의 강화된 세부 표현력 및 출력 품질을 시연하는 내용)

https://x.com/techradar/status/2027557618408149013

#nanobanana2 #imagegeneration #generativeai #diffusionmodels

TechRadar (@techradar) on X

Nano Banana 2 image prompts that showcase its powerful enhancements https://t.co/PAcKNiiD5O

X (formerly Twitter)