iiinova — where your digital self keeps evolving

Marcus Klasson @marcusklasson.bsky.social · 24/04/2026

BayesVLM improves calibration in zero-shot classification without sacrificing accuracy. The uncertainties are also useful for data selection in active fine-tuning, which actually was our target in the beginning, i.e. fetch new samples that will reduce the model's uncertainty on current observations.

100

Marcus Klasson @marcusklasson.bsky.social · 24/04/2026

The paper includes quite some math about estimating Hessians from CLIP efficiently which turned out harder as we initially thought due to the InfoNCE loss being cross-modal and contrastive. Anton, Rui, and Martin did a heckuva job to get this done in a rigorous way!

100

Marcus Klasson @marcusklasson.bsky.social · 24/04/2026

BayesVLM requires estimating Hessians over the image and text proj. layers, where access to the pretraining data (or proxy of this) is needed, but only 10 batches is sufficient. Estimating pseudo-data count and prior precision params is also needed, which is similar to what temp. scaling needs.

100

Marcus Klasson @marcusklasson.bsky.social · 24/04/2026

We put the Laplace approximation on top of CLIP to enable estimating uncertainties without inference overhead. AFAIK this is the first method that can produce zero-shot uncertainties in VLMs without architecture changes or retraining from scratch.

100

Marcus Klasson @marcusklasson.bsky.social · 24/04/2026

👋🇧🇷 If you are at #ICLR2026 today, you should talk to @antonbaumann.bsky.social who is presenting our paper about turning pre-trained VLMs into probabilistic models without retraining or fine-tuning. Poster Session 3 ⌚: 10:30am - 1:00pm (local time) 📍: Pavilion 3 P3 - #313 @iclr-conf.bsky.social

143

Reposted by Marcus Klasson

Martin Trapp @trappmartin.eurosky.social · 02/10/2025

Want to work on Trustworthy AI? 🚀 I'm seeking exceptional candidates to apply for the Digital Futures Postdoctoral Fellowship to work with me on Uncertainty Quantification, Bayesian Deep Learning, and Reliability of ML Systems. The position will be co-advised by Hossein Azizpour or Henrik Boström.

trappmartin.github.io

Home

Martin Trapp - Assistant Professor in Machine Learning at KTH Royal Institute of Technology.

1114

Marcus Klasson @marcusklasson.bsky.social · 13/06/2025

DeSplat has the same FPS and training time as vanilla 3DGS with some additional overhead for storing distractor Gaussians. Extend with MLPs or other models can also be done. Altering DeSplat to video remains to be explored, as distractors barely moving across images can be mistaken as static. [7/8]

100

Marcus Klasson @marcusklasson.bsky.social · 13/06/2025

This decomposed splatting (DeSplat) approach explicitly separates distractors from static parts. Earlier methods (e.g. SpotlessSplats, WildGaussians) use loss masking of detected distractors to avoid overfitting, while DeSplat instead jointly reconstructs distractor elements. [6/8]

100

Marcus Klasson @marcusklasson.bsky.social · 13/06/2025

Knowing how 3DGS treats distractors, we initialize a set of Gaussians close to every camera view for reconstructing view-specific distractors. The Gaussians initialized from the point cloud should reconstruct static stuff. These separately rendered images are alpha-blended during training. [5/8]

100

Marcus Klasson @marcusklasson.bsky.social · 13/06/2025

In a viewer, you can see that these spurious artefacts are thin and are located close to the camera view. For the scene-overfitting approach in 3DGS, this makes sense since an object only appearing in one view must be located as close to the camera such that no other camera view can see it. [4/8]

100

Marcus Klasson @marcusklasson.bsky.social · 13/06/2025

This BabyYoda scene from RobustNeRF is similar to a crowdsourced scenario, where a set of static toys appear together with inconsistently-placed toys between the frames. Vanilla 3DGS is quite robust here, but some views end up being rendered with spurious artefacts (right image). [3/8]

100

Marcus Klasson @marcusklasson.bsky.social · 13/06/2025

Our goal is to learn a scene representation from images that include non-static objects we refer to as distractors. An example is crowdsourced images where different people appear at different locations in the scene, which creates multi-view inconsistencies between the frames. [2/8]

100

Marcus Klasson @marcusklasson.bsky.social · 13/06/2025

👋Interested in Gaussian splatting and removing dynamic content from images? Our DeSplat is presented today at #CVPR2025 at Poster Session 1, ExHall D Poster #52. Yihao will be there to present our fully splatting-based method for separating static and dynamic stuff in images. 🧵[1/8]

161

Reposted by Marcus Klasson

Andrei Bursuc @abursuc.bsky.social · 11/06/2025

You woke up early in the morning jet-lagged and having a hard time deciding for a workshop today @cvprconference.bsky.social ? Here's a reliable choice for you: our workshop on 🛟 Uncertainty Quantification for Computer Vision! 🗓️ Day: Wed, Jun 11 📍Room: 102 B #CVPR2025 #UNCV2025

093

Marcus Klasson @marcusklasson.bsky.social · 23/04/2025

KTH is looking for a *Postdoc* to work on visual domain adaptation for mobile robot perception in a joint project with Ericsson in Stockholm. Apply by May 15 if you are interested in working with computer vision applied to real robots! More info: www.kth.se/lediga-jobb/...

kth.se

KTH | Postdoc in robotics with specialization in visual domain adaptation

KTH jobs is where you search for jobs at www.kth.se.

021

Marcus Klasson @marcusklasson.bsky.social · 17/03/2025

Submission deadline is extended to March 20 for submitting your paper to our #CVPR2025 workshop on Uncertainty Quantification for Computer Vision. Looking forward to see your submissions on recognizing failure scenarios and enabling robust vision systems! More info: uncertainty-cv.github.io/2025/

uncertainty-cv.github.io

UNCV Workshop @ CVPR 2025

CVPR 2025 Workshop on Uncertainty Quantification for Computer Vision.

0105

Reposted by Marcus Klasson

Arno Solin @arnosolin.bsky.social · 08/03/2025

There is still time to submit your papers to our #CVPR2025 workshop on Uncertainty Quantification for Computer Vision, which is part of the workshop lineup at the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) in Nashville, Tennessee.

2136

Reposted by Marcus Klasson

Andrei Bursuc @abursuc.bsky.social · 28/02/2025

Our Workshop on Uncertainty Quantification for Computer Vision goes to @cvprconference.bsky.social this year! We have a super line-up of speakers and a call for papers. This is a chance for your paper to shine at #CVPR2025 ⏲️ Submission deadline: 14 March 💻 Page: uncertainty-cv.github.io/2025/

0337

Reposted by Marcus Klasson

Martin Trapp @trappmartin.eurosky.social · 10/12/2024

I will present ✌️ BDU workshop papers @ NeurIPS: one by Rui Li (looking for internships) and one by Anton Baumann. 🔗 to extended versions: 1. 🙋 "How can we make predictions in BDL efficiently?" 👉 arxiv.org/abs/2411.18425 2. 🙋 "How can we do prob. active learning in VLMs" 👉 arxiv.org/abs/2412.06014

arxiv.org

Post-hoc Probabilistic Vision-Language Models

Vision-language models (VLMs), such as CLIP and SigLIP, have found remarkable success in classification, retrieval, and generative tasks. For this, VLMs deterministically map images and text descripti...

1184