Do you want to make your programs easier to parallelise? Yes, you do. Do you enjoy debugging concurrency? I bet you don't. Good news: you can now turn your favourite sequential structures into high-throughput concurrent ones without writing a single line of concurrent code ⇒

El Tecnológico de Monterrey y la University of Twente (inactive) firmaron un acuerdo para continuar colaborando en tres grandes ejes: educación, investigación y transferencia de conocimiento. conecta.tec.mx/es/noticias/na…

🔴 Coming soon 23th Mexican International Conference on Artificial Intelligence Keynote Speaker: Efrén Mezura-Montes / Comunicación UV 🗓️October 21th - 25th, 2024 📍CCC_INAOE Tonantzintla, Puebla, México Registration ccc.inaoep.mx/~tndi/registro… More Info micai.org/2024/ #RT

New paper📢 LLM folks have been supervised finetuning their models with data from large and expensive models (e.g., Gemini Pro). However, we achieve better perf. by finetuning on the samples from the smaller and weaker LLMs (e.g., Flash)! w/Mehran Kazemi Arian Hosseini Rishabh Agarwal Vinh Q. Tran

awesome-point-clouds-registration A list of papers about point clouds registration github.com/wsunid/awesome…

Structured Prompting is a key requirement for building real-world LLM applications or agents, but does it harm the performance and ability to reason? 🤔 ”Let Me Speak Freely” studies the impact of structured formats (JSON, XML, YAML) versus generating free-form responses across

#AcademicTwitter Yep! #phdlife #phdvoice

Free-DyGS: Camera-Pose-Free Scene Reconstruction based on Gaussian Splatting for Dynamic Surgical Videos Paper: arxiv.org/abs/2409.01003

🔴 Coming soon 23th Mexican International Conference on Artificial Intelligence Keynote Speaker: Dr. Saiph Savage / Northeastern U. 🗓️October 21th - 25th, 2024 📍CCC_INAOE UDLAP Tonantzintla, Puebla, México Registration ccc.inaoep.mx/~tndi/registro… More Info micai.org/2024/ #RT

arxiv.org/abs/2409.02426 This is our latest work that tries to reveal what diffusion/denoising model is really doing in terms of learning a data distribution: Subspace Clustering and Subspace Denoising. What surprises me the most is the number of samples required to learn the

Welcome to CV-INSIDE Lab lab Eyobed Birhanu Paulos! Eyobed from Ethiopia 🇪🇹 will be working the kidney sones autmated identification project using images from endosopes, our work with CRAN 🇫🇷 Eyobed will continue stuff on eXplaible AI and report generation!

🚨 Medical AI Research alert! 🚨 Can smaller LLMs advance medical decisions in resource-constrained hospitals? Zhejiang University presents: Med-MoE: Mixture of Domain-Specific Experts for Lightweight Medical Vision-Language Models Authors (Songtao Jiang.., Yeying Jin) Here's

Reading for this trip to France 🤩

Yo había leído algo de Banana Yoshimoto desde la secundaria, cuando leí novela Kitchen y me vio los sesos 😱 Este lo compré en la FIL 2019, pre pandemia y había quedado escondido en un rincón, y me está encantando 🤩

Book buying spree in Nancy and this only the second day 😱 Two of my fav French writers… I might try to buy new stuff as well 🤩

Tutorial on Diffusion Models for Imaging and Vision 2nd edition is up on arXiv arxiv.org/abs/2403.18103 (51 pages --> 89 pages) - Expanded VAE - More detailed DDPM - New section: physics of diffusion - and more Feedback is welcome!

This week I am visiting the CRAN lab in Nancy (🇫🇷), where my PhD students Francisco Lopez-Tiro and Ricardo Espinosa are doing double diplomas with my colleague Prof Christian Daul ☺️We are having several meeting with our urologist collaborators in the CHRU Brabois and finishing some articles in

For as along as there've been sensors, there's been noise to deal with - hence the importance of denoisers Recently we're using denoisers to solve far bigger problems -in imaging, inverse problems, & machine learning. So what are denoisers and what properties do they have? 1/n

Súper buenas perspectivas de colabs con el CHRU Nancy, en particular con el depto de urología, en nuevos proyectos!

Large language models (#LLMs) can effectively summarize medical evidence, but proprietary LLMs often lack transparency and create vendor dependency. This study showed that fine-tuning open-source LLMs significantly improves their performance, bringing some models, like LongT5,