A. Sophia Koepke

Almut Sophia Koepke

I am a junior research group leader at the Technical University of Munich and the University of Tübingen. Furthermore, I am a postdoctoral scholar at UC Berkeley, advised by Alexei (Alyosha) Efros. My research focusses on multi-modal learning problems with sound, vision, and text.

I completed my DPhil in the Visual Geometry Group (VGG) at the University of Oxford, supervised by Andrew Zisserman. I spent a summer at Reichman University working with Yael Moses. Following my DPhil, I was a postdoctoral researcher with Zeynep Akata at the University of Tübingen.

I am an ELLIS member, attached to the ELLIS unit Munich. Furthermore, I am a Zia fellow for 2025-2026.

Email GitHub LinkedIn Twitter Bluesky Google Scholar

Publications

*co-first/-last authors

	Back into Plato's Cave: Examining Cross-modal Representational Convergence at Scale A. Sophia Koepke, Daniil Zverev, Shiry Ginosar, Alexei A. Efros arXiv preprint, 2026 paper / project page / code


	LAION-BVD: A 10-Million-Hour Open Video Dataset for Multimodal Pre-training Andreas Hochlehnert, Marianna Nezhurina, Mehdi Cherti, Andrej Radonjic, Thaddäus Wiedemer, Christoph Schuhmann, Romain Beaumont, Wieland Brendel, Bernhard Schölkopf, A. Sophia Koepke, Jenia Jitsev, Matthias Bethge* Advances in Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track, 2026 paper / project page / dataset download / code / blog post


	It's never too late: Noise optimization for collapse recovery in trained diffusion models Anne Harrington, A. Sophia Koepke*, Shyamgopal Karthik, Trevor Darrell, Alexei A. Efros IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)*, 2026 paper / project page / code


	VGGSounder: Audio-visual evaluations for foundation models Daniil Zverev, Thaddäus Wiedemer, Ameya Prabhu, Matthias Bethge, Wieland Brendel, A. Sophia Koepke IEEE/CVF International Conference on Computer Vision (ICCV), 2025 paper / project page / code VGGSounder v1.0 received an additional round of manual verification (released on June 30, 2026). Per-modality multi-class annotations for the audio-visual VGGSound data.


	On the dangers of bootstrapping generation for continual learning and beyond Daniil Zverev, A. Sophia Koepke, João F. Henriques DAGM German Conference on Pattern Recognition (GCPR), 2025 paper


	Dissecting temporal understanding in text-to-audio retrieval Andreea-Maria Oncescu, João F. Henriques, A. Sophia Koepke ACM Multimedia (ACMMM), 2024 paper / project page / code


	Fantastic gains and where to find them: On the existence and prospect of general knowledge transfer between any pretrained model Karsten Roth, Lukas Thede, A. Sophia Koepke, Oriol Vinyals, Olivier J. Hénaff, Zeynep Akata International Conference on Learning Representations (ICLR), 2024 paper / code / openreview Spotlight.


	Audio-visual generalized zero-shot learning using pre-trained large multi-modal models David Kurzendörfer, Otniel-Bogdan Mercea, A. Sophia Koepke, Zeynep Akata Workshop on Learning with Limited Labelled Data for Image and Video Understanding at the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024 paper / code


	A sound approach: Using large language models to generate audio descriptions for egocentric text-audio retrieval Andreea-Maria Oncescu, João F. Henriques, Andrew Zisserman, Samuel Albanie, A. Sophia Koepke The International Conference on Acoustics, Speech, & Signal Processing (ICASSP), 2024 paper / project page / code


	Addressing caveats of neural persistence with deep graph persistence Leander Girrbach, Anders Christensen, Ole Winther, Zeynep Akata, A. Sophia Koepke Transactions on Machine Learning Research (TMLR), 2023 paper / code / openreview / video A part of this was also presented at the TAG-ML Workshop at ICML 2023.


	Zero-shot audio captioning with audio-language model guidance and audio context keywords Leonard Salewski, Stefan Fauth, A. Sophia Koepke, Zeynep Akata NeurIPS Workshop on Machine Learning for Audio, 2023 paper / code


	Video-adverb retrieval with compositional adverb-action embeddings Thomas Hummel, Otniel-Bogdan Mercea, A. Sophia Koepke, Zeynep Akata British Machine Vision Conference (BMVC), 2023 paper / project page / code Oral presentation.


	Waffling around for performance: Visual classification with random words and broad concepts Karsten Roth, Jae Myung Kim, A. Sophia Koepke, Oriol Vinyals, Cordelia Schmid, Zeynep Akata IEEE/CVF International Conference on Computer Vision (ICCV), 2023 paper / code


	Image-free classifier injection for zero-shot classification Anders Christensen, Massimiliano Mancini, A. Sophia Koepke, Ole Winther, Zeynep Akata IEEE/CVF International Conference on Computer Vision (ICCV), 2023 paper / code


	Text-to-feature diffusion for audio-visual few-shot learning Otniel-Bogdan Mercea, Thomas Hummel, A. Sophia Koepke, Zeynep Akata DAGM German Conference on Pattern Recognition (GCPR), 2023 paper / code


	Zero-shot translation of attention patterns in VQA models to natural language Leonard Salewski, A. Sophia Koepke, Hendrik Lensch, Zeynep Akata DAGM German Conference on Pattern Recognition (GCPR), 2023 paper / code


	Exposing and mitigating spurious correlations for cross-modal retrieval Jae Myung Kim, A. Sophia Koepke, Cordelia Schmid, Zeynep Akata Multimodal Learning and Applications Workshop at the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPRW), 2023 paper / code


	PlanT: Explainable planning transformers via object-level representations Katrin Renz, Kashyap Chitta, Otniel-Bogdan Mercea, A. Sophia Koepke, Zeynep Akata, Andreas Geiger Conference on Robot Learning (CoRL), 2022 paper / project page / code


	Temporal and cross-modal attention for audio-visual zero-shot learning Otniel-Bogdan Mercea, Thomas Hummel, A. Sophia Koepke, Zeynep Akata European Conference on Computer Vision (ECCV), 2022 paper / code


	Audio-visual generalised zero-shot learning with cross-modal attention and language Otniel-Bogdan Mercea, Lukas Riesch, A. Sophia Koepke, Zeynep Akata IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022 paper / code


	CLEVR-X: A visual reasoning dataset for natural language explanations Leonard Salewski, A. Sophia Koepke, Hendrik Lensch, Zeynep Akata Springer Lecture Notes on Artificial Intelligence, 2022 paper / project page / code This was also presented at the CVPR 2022 Workshop on Explainable AI for Computer Vision (XAI4CV).


	Audio retrieval with natural language queries: A benchmark study A. Sophia Koepke, Andreea-Maria Oncescu, João F. Henriques, Zeynep Akata, Samuel Albanie Transactions on Multimedia, 2022 paper / project page / code Extension of the INTERSPEECH paper with a new dataset and new results.


	Audio retrieval with natural language queries Andreea-Maria Oncescu, A. Sophia Koepke*, João F. Henriques, Zeynep Akata, Samuel Albanie INTERSPEECH*, 2021 paper / project page / code Shortlisted for best student paper award.


	Distilling audio-visual knowledge by compositional contrastive learning Yanbei Chen, Yongqin Xian, A. Sophia Koepke, Ying Shan, Zeynep Akata IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021 paper / code


	Sight to sound: An end-to-end approach for visual piano transcription A. Sophia Koepke, Olivia Wiles, Yael Moses, Andrew Zisserman The International Conference on Acoustics, Speech, & Signal Processing (ICASSP), 2020 paper / project page Oral presentation.


	Self-supervised learning of class embeddings from video Olivia Wiles, A. Sophia Koepke, Andrew Zisserman Compact and Efficient Feature Representation and Learning in Computer Vision Workshop at the IEEE/CVF International Conference on Computer Vision (ICCV Workshop), 2019 paper


	Visual pitch estimation A. Sophia Koepke, Olivia Wiles, Andrew Zisserman Sound and Music Computation Conference (SMC), 2019 paper / project page


	Self-supervised learning of a facial attribute embedding from video Olivia Wiles, A. Sophia Koepke*, Andrew Zisserman British Machine Vision Conference (BMVC)*, 2018 paper / supplementary material / project page / code Oral presentation.


	X2Face: A network for controlling face generation by using images, audio, and pose codes Olivia Wiles, A. Sophia Koepke*, Andrew Zisserman European Conference on Computer Vision (ECCV)*, 2018 paper / project page / code

Teaching

Introduction to Machine Learning (University of Tübingen, Summer semester 2023)

Community service

Area chair:

Outstanding reviewer:

Reviewer:

ICCV (2023, 2025)
NeurIPS (2023)
ECCV (2022, 2026)
CVPR (2021, 2022, 2023, 2026)
ACCV (2020)
WACV (2026)
CVPR workshop: Learning with Limited Labelled Data for Image and Video Understanding (2022)
ICCV workshop: Closing the loop between Vision and Language (2021)
NeurIPS workshop: The preregistration experiment: an alternative publication model for machine learning research (2020)
CVPR workshop: Women in Computer Vision (2019, 2020)
ICCV workshop: Neural Architects (2019)
IJCV
TPAMI
JMLR
IEEE Access
Eurographics

Workshop organisation:

Misc

Tübingen AI Center
Munich Center for Machine Learning (MCML)
ELLIS Society
KI macht Schule (AI education for pupils)