Pascal Mettes — accepted AI-conference papers

2026

The Human Brain as a Dynamic Mixture of Expert Models in Video Understanding

Christina Sartzetaki, Anne W. Zonneveld, Pablo Oyarzo, Alessandro Thomas Gifford, Radoslaw Martin Cichy, Pascal Mettes +1

ICLR 2026poster

The human brain is the most efficient and versatile system for processing dynamic visual input. By comparing representations from deep video models to brain activity, we can gain insights into mechanistic solutions for effective video processing, important to better understand the brain and to build…

Cited by 0 Source ↗code ↗Scholar ↗

2025

BrainACTIV: Identifying visuo-semantic properties driving cortical selectivity using diffusion-based image manipulation

Diego Garcia Cerdas, Christina Sartzetaki, Magnus Petersen, Gemma Roig, Pascal Mettes, Iris Groen

ICLR 2025poster

The human brain efficiently represents visual inputs through specialized neural populations that selectively respond to specific categories. Advancements in generative modeling have enabled data-driven discovery of neural selectivity using brain-optimized image synthesis. However, current methods in…

Cited by 0 Source ↗PDF ↗Scholar ↗

2025

Compositional Entailment Learning for Hyperbolic Vision-Language Models

Avik Pal, Max van Spengler, Guido Maria D'Amely di Melendugno, Alessandro Flaborea, Fabio Galasso, Pascal Mettes

ICLR 2025oral

Image-text representation learning forms a cornerstone in vision-language models, where pairs of images and textual descriptions are contrastively aligned in a shared embedding space. Since visual and textual concepts are naturally hierarchical, recent work has shown that hyperbolic space can serve…

Cited by 0 Source ↗PDF ↗code ↗Scholar ↗

2025

Hyperbolic Safety-Aware Vision-Language Models

Tobia Poppi, Tejaswi Kasarla, Pascal Mettes, Lorenzo Baraldi, Rita Cucchiara

CVPR 2025highlight

Addressing the retrieval of unsafe content from vision-language models such as CLIP is an important step towards real-world integration. Current efforts have relied on unlearning techniques that try to erase the model's knowledge of unsafe concepts. While effective in reducing unwanted outputs, unle…

Cited by 0 Source ↗PDF ↗code ↗Scholar ↗

2025

Large Language Models Are Natural Video Popularity Predictors

Pratik Kayal, Pascal Mettes, Nima Dehmamy, Minsu Park

ACL 2025finding

Predicting video popularity is often framed as a supervised learning task, relying heavily on meta-information and aggregated engagement data. However, video popularity is shaped by complex cultural and social factors that such approaches often overlook. We argue that Large Language Models (LLMs), w…

Cited by 0 Source ↗PDF ↗code ↗Scholar ↗

2025

Low-distortion and GPU-compatible Tree Embeddings in Hyperbolic Space

Max van Spengler, Pascal Mettes

ICML 2025poster

Embedding tree-like data, from hierarchies to ontologies and taxonomies, forms a well-studied problem for representing knowledge across many domains. Hyperbolic geometry provides a natural solution for embedding trees, with vastly superior performance over Euclidean embeddings. Recent literature has…

Cited by 0 Source ↗PDF ↗Scholar ↗

2025

Union-over-Intersections: Object Detection beyond Winner-Takes-All

Aritra Bhowmik, Pascal Mettes, Martin R. Oswald, Cees G. M. Snoek

ICLR 2025spotlight

This paper revisits the problem of predicting box locations in object detection architectures. Typically, each box proposal or box query aims to directly maximize the intersection-over-union score with the ground truth, followed by a winner-takes-all non-maximum suppression where only the highest sc…

Cited by 0 Source ↗PDF ↗code ↗Scholar ↗

2024

Hyp2Nav: Hyperbolic Planning and Curiosity for Crowd Navigation

Guido Maria D'Amely di Melendugno, Alessandro Flaborea, Pascal Mettes, Fabio Galasso

IROS 2024

Autonomous robots are increasingly becoming a strong fixture in social environments. Effective crowd navigation requires not only safe yet fast planning, but should also enable interpretability and computational efficiency for working in real-time on embedded devices. In this work, we advocate for h

Cited by 2 Source ↗code ↗Scholar ↗

2024

Latent Space Editing in Transformer-Based Flow Matching

Vincent Tao Hu, Wei Zhang, Meng Tang, Pascal Mettes, Deli Zhao, Cees Snoek

AAAI 2024technical

This paper strives for image editing via generative models. Flow Matching is an emerging generative modeling technique that offers the advantage of simple and efficient training. Simultaneously, a new transformer-based U-ViT has recently been proposed to replace the commonly used UNet for better sca…

Cited by 31 Source ↗PDF ↗Scholar ↗

2023

Poincare ResNet

Max van Spengler, Erwin Berkhout, Pascal Mettes

ICCV 2023poster

This paper introduces an end-to-end residual network that operates entirely on the Poincare ball model of hyperbolic space. Hyperbolic learning has recently shown great potential for visual understanding, but is currently only performed in the penultimate layer(s) of deep networks. All visual repres…

Cited by 25 PDF ↗code ↗Scholar ↗

2022

Hyperbolic Image Segmentation

Mina Ghadimi Atigh, Julian Schoep, Erman Acar, Nanne van Noord, Pascal Mettes

CVPR 2022poster

For image segmentation, the current standard is to perform pixel-level optimization and inference in Euclidean output embedding spaces through linear hyperplanes. In this work, we show that hyperbolic manifolds provide a valuable alternative for image segmentation and propose a tractable formulation…

Cited by 114 PDF ↗code ↗Scholar ↗

2022

Less than Few: Self-Shot Video Instance Segmentation

Pengwan Yang, Yuki M. Asano, Pascal Mettes, Cees G. M. Snoek

ECCV 2022poster

"The goal of this paper is to bypass the need for labelled examples in few-shot video understanding at run time. While proven effective, in many practical video settings even labelling a few examples appears unrealistic. This is especially true as the level of details in spatio-temporal video unders…

Cited by 3 Source ↗PDF ↗code ↗Scholar ↗

2022

Maximum Class Separation as Inductive Bias in One Matrix

Tejaswi Kasarla, Gertjan J. Burghouts, Max van Spengler, Elise van der Pol, Rita Cucchiara, Pascal Mettes

NeurIPS 2022accept

Maximizing the separation between classes constitutes a well-known inductive bias in machine learning and a pillar of many traditional algorithms. By default, deep networks are not equipped with this inductive bias and therefore many alternative solutions have been proposed through differential opti…

Cited by 31 Source ↗PDF ↗code ↗Scholar ↗

2021

Few-Shot Transformation of Common Actions Into Time and Space

Pengwan Yang, Pascal Mettes, Cees G. M. Snoek

CVPR 2021poster

This paper introduces the task of few-shot common action localization in time and space. Given a few trimmed support videos containing the same but unknown action, we strive for spatio-temporal localization of that action in a long untrimmed query video. We do not require any class labels, interval…

Cited by 13 PDF ↗Scholar ↗

2021

Hyperbolic Busemann Learning with Ideal Prototypes

Mina Ghadimi Atigh, Martin Keller-Ressel, Pascal Mettes

NeurIPS 2021poster

Hyperbolic space has become a popular choice of manifold for representation learning of various datatypes from tree-like structures and text to graphs. Building on the success of deep learning with prototypes in Euclidean and hyperspherical spaces, a few recent works have proposed hyperbolic prototy…

Cited by 44 Source ↗PDF ↗code ↗Scholar ↗

2021

Social Fabric: Tubelet Compositions for Video Relation Detection

Shuo Chen, Zenglin Shi, Pascal Mettes, Cees G. M. Snoek

ICCV 2021poster

This paper strives to classify and detect the relationship between object tubelets appearing within a video as a <subject-predicate-object> triplet. Where existing works treat object proposals or tubelets as single entities and model their relations a posteriori, we propose to classify and detect pr…

Cited by 30 PDF ↗code ↗Scholar ↗

2020

Localizing the Common Action Among a Few Videos

Pengwan Yang, Vincent Tao Hu, Pascal Mettes, Cees G. M. Snoek

ECCV 2020poster

This paper strives to localize the temporal extent of an action in a long untrimmed video. Where existing work leverages many examples with their start, their ending, and/or the class of the action during training time, we propose few-shot common action localization. The start and end of an action i…

Cited by 29 Source ↗PDF ↗code ↗Scholar ↗

2020

PointMixup: Augmentation for Point Clouds

Yunlu Chen, Vincent Tao Hu, Efstratios Gavves, Thomas Mensink, Pascal Mettes, Pengwan Yang +1

ECCV 2020poster

This paper introduces data augmentation for point clouds by interpolation between examples. Data augmentation by interpolation has shown to be a simple and effective approach in the image domain. Such a mixup is however not directly transferable to point clouds, as we do not have a one-to-one corres…

Cited by 216 Source ↗PDF ↗code ↗Scholar ↗

2020

Searching for Actions on the Hyperbole

Teng Long, Pascal Mettes, Heng Tao Shen, Cees G. M. Snoek

CVPR 2020poster

In this paper, we introduce hierarchical action search. Starting from the observation that hierarchies are mostly ignored in the action literature, we retrieve not only individual actions but also relevant and related actions, given an action name or video example as input. We propose a hyperbolic a…

Cited by 61 PDF ↗code ↗Scholar ↗

2019

Counting With Focus for Free

Zenglin Shi, Pascal Mettes, Cees G. M. Snoek

ICCV 2019poster

This paper aims to count arbitrary objects in images. The leading counting approaches start from point annotations per object from which they construct density maps. Then, their training objective transforms input images to density maps through deep convolutional networks. We posit that the point an…

Cited by 137 PDF ↗code ↗Scholar ↗

2019

Hyperspherical Prototype Networks

Pascal Mettes, Elise van der Pol, Cees Snoek

NeurIPS 2019poster

This paper introduces hyperspherical prototype networks, which unify classification and regression with prototypes on hyperspherical output spaces. For classification, a common approach is to define prototypes as the mean output vector over training examples per class. Here, we propose to use hypers…

Cited by 180 Source ↗PDF ↗code ↗Scholar ↗

2019

SILCO: Show a Few Images, Localize the Common Object

Tao Hu, Pascal Mettes, Jia-Hong Huang, Cees G. M. Snoek

ICCV 2019poster

Few-shot learning is a nascent research topic, motivated by the fact that traditional deep learning requires tremendous amounts of data. In this work, we propose a new task along this research direction, we call few-shot common-localization. Given a few weakly-supervised support images, we aim to lo…

Cited by 56 PDF ↗Scholar ↗

2017

Spatial-Aware Object Embeddings for Zero-Shot Localization and Classification of Actions

Pascal Mettes, Cees G. M. Snoek

ICCV 2017oral

We aim for zero-shot localization and classification of human actions in video. Where traditional approaches rely on global attribute or object classification scores for their zero-shot knowledge transfer, our main contribution is a spatial-aware object embedding. To arrive at spatial awareness, we…

Cited by 109 PDF ↗Scholar ↗