Publications

BackTranslation2.0: A Linguistically Motivated Metric to Assess Sign Language Production

Published in ECCV 2026, 2026

A linguistically grounded, agentic evaluation metric for text-to-sign translation that scores grammatical correctness, phonological accuracy, motion fluency, and generation fidelity, validated on a BSL dataset annotated by native Deaf raters.

Recommended citation: Cory, O., Ivashechkin, M., Ranum, O., Low, J., Fish, E., Pelykh, A., Sahin, K., Mercanoglu Sincan, O., Bowden, R. (2026). "BackTranslation2.0: A Linguistically Motivated Metric to Assess Sign Language Production." ECCV 2026. https://arxiv.org/pdf/2606.28673

The NGT200 Dataset: Geometric Multi-View Isolated Sign Recognition

Published in GRaM Workshop @ ICML 2024, 2024

A novel spatio-temporal multi-view benchmark for isolated sign recognition of NGT. SE(2)-equivariant Temporal-PONITA improves multi-view recognition by 8–22% over the SL-GCN baseline.

Recommended citation: Ranum, O., Wessels, D., Otterspeer, G., Bekkers, E., Roelofsen, F., Andersen, J. (2024). "The NGT200 Dataset." GRaM Workshop @ ICML 2024. https://openreview.net/forum?id=idkNzTC67X

3D-LEX v1.0 — 3D Lexicons for American Sign Language and Sign Language of the Netherlands

Published in LREC-COLING 2024 · Sign Language Workshop, 2024

High-resolution motion capture dataset of 2,000 signs across ASL and NGT, with semi-automatic phonetic annotation that matches or exceeds expert accuracy.

Recommended citation: Ranum, O., Otterspeer, G., Andersen, J., Belleman, R., Roelofsen, F. (2024). "3D-LEX v1.0." LREC-COLING 2024 Sign Language Workshop. pp. 290–301. https://aclanthology.org/2024.signlang-1.33/

Reproducibility Study: Joint Multisided Exposure Fairness for Recommendation

Published in ML Reproducibility Challenge 2022 · ReScience, 2023

Reproducibility study of joint multisided exposure fairness for recommendation systems, as part of the ML Reproducibility Challenge 2022.

Recommended citation: Hu, A., Ranum, O., Pozrikidou, C., Zhou, M. (2023). "Reproducibility Study of Joint Multisided Exposure Fairness for Recommendation." ML Reproducibility Challenge 2022. https://openreview.net/forum?id=A0Sjs3IJWb-

A fiber detector to monitor ortho-Ps formation and decay at the AEgIS experiment, CERN

Published in Nuclear Instruments and Methods in Physics Research Section A, 2022

Development of a scintillating fiber detector system for monitoring positronium formation and decay at the AEgIS antimatter experiment at CERN.

Recommended citation: Rienäcker, B., Brusa, R.S., Caravita, R., Mariazzi, S., Penasa, L., Pino, F., Ranum, O.A., Nebbia, G. (2022). "A fiber detector to monitor ortho-Ps formation and decay." NIM-A. 1027, 166275. https://doi.org/10.1016/j.nima.2021.166275