no code implementations • 15 Mar 2024 • Scott Cheng-Hsin Yang, Baxter Eaves, Michael Schmidt, Ken Swanson, Patrick Shafto
Many metrics exist for evaluating the quality of synthetic tabular data; however, we lack an objective, coherent interpretation of the many metrics.
no code implementations • 5 Mar 2024 • Stefan Hackmann, Haniyeh Mahmoudian, Mark Steadman, Michael Schmidt
This approach, inspired by permutation importance for tabular data, masks each word in the system prompt and evaluates its effect on the outputs based on the available text scores aggregated over multiple user inputs.
no code implementations • 29 Aug 2023 • Alexander Lehner, Stefano Gasperini, Alvaro Marcos-Ramiro, Michael Schmidt, Nassir Navab, Benjamin Busam, Federico Tombari
We conduct extensive experiments across a variety of scenarios on data from KITTI, Waymo, and CrashD for 3D object detection, and on data from SemanticKITTI, Waymo, and nuScenes for 3D semantic segmentation.
no code implementations • 17 Jul 2023 • Aral Hekimoglu, Michael Schmidt, Alvaro Marcos-Ramiro
We propose a novel semi-supervised active learning (SSAL) framework for monocular 3D object detection with LiDAR guidance (MonoLiG), which leverages all modalities of collected data during model development.
no code implementations • 17 Jul 2023 • Aral Hekimoglu, Adrian Brucker, Alper Kagan Kayali, Michael Schmidt, Alvaro Marcos-Ramiro
Curating an informative and representative dataset is essential for enhancing the performance of 2D object detectors.
no code implementations • 21 Jun 2023 • Aral Hekimoglu, Philipp Friedrich, Walter Zimmer, Michael Schmidt, Alvaro Marcos-Ramiro, Alois C. Knoll
In single-task vision-based settings, inconsistency-based active learning has proven to be effective in selecting informative samples for annotation.
no code implementations • ICCV 2023 • Stefano Gasperini, Alvaro Marcos-Ramiro, Michael Schmidt, Nassir Navab, Benjamin Busam, Federico Tombari
By doing so, for the first time in panoptic segmentation with unknown objects, our U3HS is trained without unknown categories, reducing assumptions and leaving the settings as unconstrained as in real-life scenarios.
no code implementations • CVPR 2022 • Alexander Lehner, Stefano Gasperini, Alvaro Marcos-Ramiro, Michael Schmidt, Mohammad-Ali Nikouei Mahani, Nassir Navab, Benjamin Busam, Federico Tombari
Despite training only on a standard dataset, such as KITTI, augmenting with our vector fields significantly improves the generalization to differently shaped objects and scenes.
no code implementations • NeurIPS 2019 • Dominik Linzner, Michael Schmidt, Heinz Koeppl
Instead of sampling and scoring all possible structures individually, we assume the generator of the CTBN to be composed as a mixture of generators stemming from different structures.