beSpacific2d ago
Via #LLRX - #AI in #Finance & #Banking, 5/15/26 - highlights: AI Managed #Household Portfolios: Report; #ClimateRisk & AI: Banking’s Next Regulatory Frontier'; Microstructure of AI #Diffusion: Evidence from Firms, Business Functions, Worker Tasks; #Financial Stability #Risks Mount as AI Fuels #Cyberattacks; #Anthropic deepens finance push w 10 new #AI agents 4 warns over private credit industry fuelling AI boom; #OpenAI launched new set of #personalfinance tools https://www.llrx.com/2026/05/ai-in-finance-and-banking-may-15-2026/
André Ourednik3d ago

Ma conférence "#Robopïèses, #robotectures: les #IA de la #nature et de l'espace construit", donnée à l'ENSA Strasbourg est désormais en ligne dans son intégralité.

Un chaleureux merci @dorianne dupré-tabti pour l'invitation et l'organisation de mon intervention, ainsi qu'à l'ENSA pour l'agréable accueil!

#intelligence #smartcity #circulation #urbanisme #géographie #fractales #lsystems #architecture #modèlesbasésagents #GAN #diffusion #slop #cloud #impensé

https://www.youtube.com/watch?v=igDN-60yHi4

André OUREDNIK. Robopoïèses, robotectures : les IA de la nature et de l’espace construit

YouTube
sayzard4d ago

DramaBox: An Open-Weight TTS That Reads Stage Directions

DramaBox는 텍스트가 아닌 연극 대본 형식의 무대 지시문을 입력받아 감정과 연기 톤을 반영한 음성을 생성하는 오픈 웨이트 TTS 모델입니다. Lightricks의 LTX-2.3 비디오용 확산 트랜스포머를 IC-LoRA로 파인튜닝하고, Gemma 3 12B 텍스트 임베딩으로 자연어 지시문을 해석해 대사 전달 방식을 조절합니다. 10초 음성 샘플을 주면 해당 목소리와 연기 스타일을 모방할 수 있으며, 게임, 오디오 드라마, 더빙 등 표현력이 중요한 오디오 네이티브 경험에 적합합니다. 다만 24GB VRAM 요구와 상업적 이용 제한이 있어 주의가 필요합니다.

https://firethering.com/dramabox-open-weights-tts-voice-cloning/

#tts #diffusion #llm #voicecloning #audiogeneration

DramaBox: An Open-Weight TTS Model Built Around Stage Directions - Firethering

Dramabox just landed on Hugging Face and the demo space is live. Resemble AI built it on top of Lightricks' LTX-2.3, and the thing that makes it different from every other TTS model is simpler than you'd expect, you don't give it text to read. You write it a scene.

Firethering
sayzardMay 11

fly51fly (@fly51fly)

ByteDance Seed가 'Continuous Latent Diffusion Language Model' 논문을 소개했다. 연속 잠재공간에서 확산(diffusion) 방식을 활용하는 언어모델로, 기존 autoregressive LLM과 다른 생성 패러다임을 제시하는 연구다. 새로운 텍스트 생성 접근법으로 주목할 만하다.

https://x.com/fly51fly/status/2053589032333152693

#llm #diffusion #languagemodel #bytedance #research

fly51fly (@fly51fly) on X

[CL] Continuous Latent Diffusion Language Model H Guo, Q Zhao, Y Zhao, S Nie… [Bytedance Seed] (2026) https://t.co/WliGZc9lMy

X (formerly Twitter)
Courgette EditionsMay 8
#diffusion #distribution #coulisse #edition
https://courgette-editions.com/une-tournee-locale-pour-courgette-editions/
Une tournée locale pour Courgette Éditions !

Ces derniers jours, Courgette Éditions est parti à la rencontre des commerçants de Herstal et de Liège pour annoncer la sortie de Présomption de culpabilité et sa soirée d’inauguration, ce samedi 9…

Courgette Editions
N-gated Hacker NewsMay 6
🤯 Oh wow, 84 minutes to learn that "integral of a #diffusion model" is just fancy math-talk for "fancy noise-guessing game." DeepMind's secret sauce: taking baby steps in random noise-land until 🎯—voilà!—magic samples appear! 🍻 Cheers to #Sander for making #calculus look like it needs a therapist.
https://sander.ai/2026/05/06/flow-maps.html #DeepMind #Integral #Math #Noise #HackerNews #ngated
Learning the integral of a diffusion model

A deep dive on flow maps.

Sander Dieleman
Hacker NewsMay 6

Learning the Integral of a Diffusion Model

https://sander.ai/2026/05/06/flow-maps.html

#HackerNews #Learning #Integral #Diffusion #Model #Flow #Maps #DataScience

Learning the integral of a diffusion model

A deep dive on flow maps.

Sander Dieleman
L'agenda du LibreMay 3

Émission | Radio FM-Plus | Entrée Libre | Diffusion

6 mai 2026, 12:00:00 UTC+02:00 - GMT+02:00 - Radio FM-Plus, 4 rue Saint Barthelemy, Montpellier, Occitanie, France

https://keskonfai.fr/events/865203f8-df99-4002-b02a-f6c16063217d

Agenda du LibreMay 2
Montpellier: Émission | Radio FM-Plus | Entrée Libre | Diffusion, Le mercredi 6 mai 2026 de 12h00 à 13h00. https://www.agendadulibre.org/events/35096 #montpelLibre #radio #fmPlus #diffusion #libre
Émission | Radio FM-Plus | Entrée Libre | Diffusion

Montpel'libre réalise une série d’émissions régulières à la Radio FM-Plus intitulées « Entrée Libre ». Ces émissions sont la présentation hebdomadaire des activités de Montpel’libre. Après le jingle où l’on présente brièvement Montpel'libre, nous donnerons un coup de projecteur sur les activités qui

sayzardMay 1

Hugging Models (@HuggingModels)

Qwen3 기반에 새로운 diffusion-based speculative decoding을 결합한 z-lab/Qwen3.6-27B-DFlash가 소개됐다. flash decoding을 통해 텍스트 생성 속도와 효율성을 높인 모델로, AI 커뮤니티에서 주목받고 있다.

https://x.com/HuggingModels/status/2049772758771646701

#qwen3 #speculativedecoding #textgeneration #diffusion #flashdecoding

Hugging Models (@HuggingModels) on X

Imagine a model that combines the power of Qwen3 with a new diffusion-based speculative decoding. That's z-lab/Qwen3.6-27B-DFlash. It's a text-generation transformer that uses flash decoding for speed and efficiency. The AI community is buzzing about this one.

X (formerly Twitter)