PhD student in explainable AI for computer vision @visinf @TUDarmstadt. Prev. intern @AWS and @maxplanckpress.
Joined July 2022
- Tweets31
- Following142
- Followers76
- Likes69
Attending @CVPR ? Don't miss any interesting workshop speakers:
robinhesse.github.io/worksho…
#cvpr2026
Some personal @NeurIPSConf workshop keynote highlights 👇
• Yoshua Bengio — Regulatable ML — Sun
• Dawn Song — LOCK-LLM — Sat
• Max Tegmark — MATH-AI — Sat
• Chelsea Finn — Embodied World Models — Sat
robinhesse.github.io/worksho…
double-check for last-minute changes
#neurips25
@NeurIPSConf is two weeks away!
📢 Stop missing great workshop speakers just because the workshop wasn’t on your radar. Browse them all in one place:
robinhesse.github.io/worksho…
(also available for @EurIPSConf )
#NeurIPS #neurips2025 #EurIPS
Got questions about the state of explainable AI in computer vision? Drop them here, and our speakers might discuss them!
🚨 Call for Questions! 🚨
We are inviting the community and the stakeholders to submit questions, which will be discussed with our experts at the workshop! 🎤💡
👉 Submit your questions: forms.gle/8cYb4Ce3dGHibxTG7
Workshop: excv-workshop.github.io/
#ICCV2025 @ICCVConference #eXCV
Got a strong XAI paper rejected from ICCV? Submit it to our ICCV eXCV Workshop today—we welcome high-quality work!
🗓️ Submissions open until June 26 AoE.
📄 Got accepted to ICCV? Congrats! Consider our non-proceedings track.
#ICCV2025 @ICCVConference
Robin Hesse retweeted
Want to learn about how model design choices affect the attribution quality of vision models? Visit our #NeurIPS2024 poster on Friday afternoon (East Exhibition Hall A-C #2910)! @robinhesse_ @schaub_simone @stefanroth
Paper: arxiv.org/abs/2407.11910
Code: github.com/visinf/idsds
Robin Hesse retweeted
📢 Excited to share our new work: AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers
snap-research.github.io/ac3d
We analyze what pre-trained video diffusion transformers understand about 3D and demonstrate dynamic scene generation with 3D control.
Robin Hesse retweeted
Our work on transforming existing neural networks to be inherently interpretable has been accepted at #NeurIPS2024! This is joint work with @shrebox, @MoritzBoehle, and Bernt Schiele at @cvml_mpiinf.
(1/4)
ALT B-cosification Overview
Robin Hesse retweeted
🎭Recent work shows that models’ inductive biases for 'simpler' features may lead to shortcut learning.
What do 'simple' vs 'complex' features look like? What roles do they play in generalization?
Our new paper explores these questions.
arxiv.org/pdf/2407.06076
#Neurips2024
ALT Qualitative Analysis of “Meta-feature” (cluster of features) Complexity. (Left) A 2D UMAP projection displaying the 10,000 extracted features. The features are organized into 150 clusters using K-means clustering applied to the feature dictionary D. 30 clusters were selected for analysis of features at different complexity levels. (Right) For each Meta-feature cluster, we compute the average complexity score. This allows us to classify the features based on their complexity according to the model. Notably, simple features are often akin to color detectors (e.g., grass, sky) and detectors for low-frequency patterns (e.g., bokeh detector) or lines. In contrast, complex features encompass parts or structured objects, as well as features resembling shape.
Robin Hesse retweeted
After mech int work on vision models and on language models, it’s time to see what’s going on inside VLMs!
Not only that the logit lens on image features works, we show that these features can be linearly edited to reduce hallucinations!
@nickhjiang and @anishk__ are great!
Robin Hesse retweeted
This year's #ECCV2024 recap: Presented our work w @shashankska on Uspampling ⬆️, our work on infinitely large convolutions ♾️ , went to the Doctoral Consortium 🎓 and met with new and old friends 🍻 thanks @eccvconf, it was a blast!
If you're interested in how different model design choices affect attribution quality, check out our (@schaub_simone and @stefanroth) paper "Benchmarking the Attribution Quality of Vision Models," which has been accepted at #NeurIPS2024 @NeurIPSConf D&B track!
Robin Hesse retweeted
What do you see in these images?
These are called hybrid images, originally proposed by Aude Oliva et al. They change appearance depending on size or viewing distance, and are just one kind of perceptual illusion that our method, Factorized Diffusion, can make.
Thanks to all the speakers for the exciting talks and all the authors for their interesting posters! It was a pleasure to have such a great program.
Thank you for the excellent workshop and for giving us the opportunity to talk about our work! It's really valuable to discuss HCI papers at CV/ML conferences and vice versa. Cheers to cross-pollination of perspectives 🐝🌼
@eccvconf #ECCV2024 @XAI_Research
Robin Hesse retweeted
Excited to give a talk about interpretability via reverse engineering tmrw at @eccvconf!
Less than three days to go for the eXCV Workshop at #ECCV2024! Join us on Sunday from 14:00-18:00 in Brown 1 to hear about the state of XAI research from an exciting lineup of speakers!
@orussakovsky, @vidal_rene, @sunniesuhyoung, @YGandelsman, @zeynepakata
@eccvconf
(1/4)
ALT Invited Speakers
Robin Hesse retweeted
Less than three days to go for the eXCV Workshop at #ECCV2024! Join us on Sunday from 14:00-18:00 in Brown 1 to hear about the state of XAI research from an exciting lineup of speakers!
@orussakovsky, @vidal_rene, @sunniesuhyoung, @YGandelsman, @zeynepakata
@eccvconf
(1/4)
ALT Invited Speakers
Robin Hesse retweeted
🏔️🇨🇭"Grüezi" from our summer retreat in the Swiss mountains. It’s been a fantastic week filled with research discussions, hikes, and amazing Swiss food.
Robin Hesse retweeted
Notifications of decisions have now been sent to authors of all submissions. Congratulations to all the authors whose papers were accepted, and a big thanks to all the reviewers! See you at the workshop!
#ECCV2024 @eccvconf
Robin Hesse retweeted
Share your latest published work for explainability in computer vision at our workshop at #ECCV2024! Submit to our Non-Proceedings track, deadline is tomorrow (Aug 5, 23:59 CEST)!
Details: excv-workshop.github.io
@eccvconf
