🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

By popular request, I wrote a separate article on effectively managing and running experiments during a PhD in AI. I learned that effectively running experiments can be a crucial aspect of a PhD and I wanted to share some of my learnings 🧵: davidstutz.de/i3A0y

We're excited to release TAPVid-3D: an evaluation benchmark of 4,000+ real world videos and 2.1 million metric 3D point trajectories, for the task of Tracking Any Point in 3D!

The wait is over 📢 MAST3R is out! DUSt3R+ dense local feature maps & metric depth - 1st in #MapFreeReloc leaderboard, can handle 1000s of images 😀 !! Blog: shorturl.at/9JTH2 Code: github.com/naver/mast3r Paper: arxiv.org/abs/2406.09756

🎉 At ECCV'24 Gizem Unlu Gizem Esra Ünlü will present our work "GroundUp" introducing a sketch-based ideation tool for 3D city massing. We step towards empowering architects to easily switch between 2D sketches and 3D models, making iteration and idea sharing smoother! Details ⬇️

Ominimotion-> SpatialTracker -> Shape of Motion! Cool!

Images aren’t arbitrary collections of pixels -they have complicated structure, even small ones. That’s why it’s hard to generate images well. Let me give you an idea: 3×3 gray images represented as points in ℝ⁹ lie approximately on a 2-D manifold: the Klein bottle! 1/3

arXiv -> alphaXiv Students at Stanford have built alphaXiv, an open discussion forum for arXiv papers. alphaXiv You can post questions and comments directly on top of any arXiv paper by changing arXiv to alphaXiv in any URL!

Our new robot is here! UCL East campus

Many people are in the middle of the #CVPR2025 deadline. So I'm sharing my guide to writing a CVPR paper (or any paper). My students have had this for years but I haven't shared it publicly before. I hope you find it useful and write a great paper. #CVPR2025 medium.com/@black_51980/w…

In the meantime, many EPSRC-funded HPCs for AI research end last/this year without proper replacement. Had a hard time getting alternative GPU resources since July, when JADE2 HPC Service announced they would stop the service in Nov. 2024…

A really great paper! By managing a persistent state/memory, they can do static/dynamic reconstruction/pose estimation; Inferring unseen structures from this compact representation; Refining geometry by revisiting the entire sequence:)

A complete Siamese network with an extra embedding defining coordinate system + per-layer memory. A really interesting architecture with impressive results using only 10-frame training budget:) I am curious about their training frame selection strategy

Pow3R: Empowering Unconstrained 3D Reconstruction with Camera and Scene Priors Wonbong Jang / Won, Weinzaepfel Philippe, Vincent Leroy, Lourdes Agapito, Jerome Revaud tl;dr: DUSt3R with any optional input subset of camera intrinsics, pose and depthmaps arxiv.org/abs/2503.17316

Happy to introduce our new CVPR paper—Pow3R: Empowering Unconstrained 3D Reconstruction with Scene and Camera Priors : arxiv.org/pdf/2503.17316 Inspired by the amazing DUSt3R, which reconstructs 3D from just two unposed images, we explored: (1/n)

Introducing DolphinGemma, an LLM fine-tuned on many years of dolphin sound data 🐬 to help advance scientific discovery. We collaborated with Wild Dolphin Project to train a model that learns vocal patterns to predict what sound they might make next. It’s small enough (~400M params)

We’ll be presenting Pow3R on Friday the 13th, from 10:30 to 12:30, during Poster Session 1 (No. 84). Would love to see you there! I’ll also be at #CVPR2025 until Sunday — happy to grab a coffee if you’re around!

NICE:)

Optimization-based ones might have hit their upper limit, but the task isn’t solved. Since DUSt3R, Learning-based ones are starting to catch up and even surpass in cases like sparse views, textureless scenes—yet still lag in areas where traditional methods excel, e.g. large-scale