Ce chercheur génial @gabrielpeyre porte une vision brillante sur les liens entre IA et mathématiques @arthurmensch a été son élève avant la création de @MistralAI Abonnez-vous pour suivre la révolution de l’IA qui est d’abord une révolution mathématique ! Chapeau Bas !
RESEARCH
-
AI Resilience: Beyond Safety Through Strength Building
By
–
AI Safety isn't enough – we need AI Resilience. Focusing only on restrictions makes us vulnerable. The key to long-term survival isn't control – it's building strength. AI Resilience!
-
Kimi k1.5: Multimodal Reasoning Model Analysis
By
–
Very good mind path about the kimi k1.5 project, from its contributor, though in Chinese and requiring signing in… 如何评价 Kimi 发布的多模态推理模型 k1.5? – Flood Sung的回答 – 知乎 https://
zhihu.com/question/10114
790245/answer/84028353434
… -
AI Models Learning from Experience: Emerging Breakthrough
By
–
AI models that learn from experience? Could be an unlock in the coming year. https://t.co/p53Mkizrxd
— Paul Roetzer (@paulroetzer) 23 janvier 2025AI models that learn from experience? Could be an unlock in the coming year.
-

Future of Programming with AI Podcast Coming Soon
By
–
new pod with @rauchg on the future of programming with AI dropping soon! pic.twitter.com/oGv0fVAyBB
— Dan Shipper 📧 (@danshipper) 23 janvier 2025new pod with @rauchg on the future of programming with AI dropping soon!
-

Reasoning Models Show Increased Robustness Against Adversarial Attacks
By
–
Interesting & encouraging result: reasoning models such as o1 become more robust to adversarial attacks as they think for longer.
-
One-Prompt Consistent Text-to-Image Generation Method
By
–
3/3 Paper "One-Prompt-One-Story: Free-Lunch Consistent Text-to-Image Generation Using a Single Prompt"
-
InterLCM: Latent Consistency Models for Blind Face Restoration
By
–
2/3 Paper "InterLCM: Low-Quality Images as Intermediate States of Latent Consistency Models for Effective Blind Face Restoration"
-

Diffusion Classifier Scores for Personalized Text-to-Image Models
By
–
1/3 For paper “Mining Your Own Secrets: Diffusion Classifier Scores for Continual Personalization of Text-to-Image Diffusion Models” https://
arxiv.org/abs/2410.00700 -

Sony and collaborators celebrate three ICLR 2025 paper acceptances
By
–
3 ICLR 2025 accepted! Big congrats to the first authors (
@saurav_j_ who is one of Sony's as always talented interns, Tao an amazing master student and Senmao yet another talented PhD student.) along with coauthors. More details will be out soon! #ICLR2025
