Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
Spot the Difference: Detection of Topological Changes via Geometric Alignment
Authors: Per Steffen Czolbe, Aasa Feragen, Oswin Krause
NeurIPS 2021 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our approach is validated on two tasks and datasets: detection of topological changes in microscopy images of cells, and unsupervised anomaly detection brain imaging. |
| Researcher Affiliation | Academia | Steffen Czolbe Department of Computer Science University of Copenhagen EMAIL Aasa Feragen DTU Compute Technical University of Denmark EMAIL Oswin Krause Department of Computer Science University of Copenhagen EMAIL |
| Pseudocode | No | The paper describes algorithms but does not include any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | The implementation is available at github.com/Steffen Czolbe/Topological Change Detection. |
| Open Datasets | Yes | For the control set, we combine T1 weighted MRI scans of the healthy subjects from the ABIDE I [11]1, ABIDE II [12] and OASIS3 [27] studies. For the tumor set we use MRI scans from the Bra TS2020 brain tumor segmentation challenge [3, 4, 34] |
| Dataset Splits | Yes | the remaining images of the control dataset are split 2381/149/162 for train/validation/test. |
| Hardware Specification | Yes | We train each model on a single Titan RTX GPU |
| Software Dependencies | No | The paper mentions 'ADAM [25]' (optimizer) and 'U-Net [44]' (architecture) but does not provide specific version numbers for software libraries or dependencies (e.g., Python, PyTorch, TensorFlow versions). |
| Experiment Setup | Yes | For training, the optimization algorithm is ADAM [25] with a learning rate of 10 4. Regularization of all models is performed by applying an L2-penalty to the weights with a factor of 0.01 for the cell dataset and 0.0005 for the brains. ... averaged over the mini-batch of 32 image pairs. |