Measuring Robustness in Deep Learning Based Compressive Sensing
Authors: Mohammad Zalbagi Darestani, Akshay S Chaudhari, Reinhard Heckel
ICML 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this work, we measure the robustness of different approaches for image reconstruction including trained and un-trained neural networks as well as traditional sparsity-based methods. We find, contrary to prior works, that both trained and un-trained methods are vulnerable to adversarial perturbations. Moreover, both trained and un-trained methods tuned for a particular dataset suffer very similarly from distribution shifts. Finally, we demonstrate that an image reconstruction method that achieves higher reconstruction quality, also performs better in terms of accurately recovering fine details. Our results indicate that the state-of-the-art deep-learning-based image reconstruction methods provide improved performance than traditional methods without compromising robustness. |
| Researcher Affiliation | Academia | 1Department of Electrical and Computer Engineering, Rice University 2Department of Radiology and Department of Biomedical Data Science, Stanford University 3Department of Electrical and Computer Engineering, Technical University of Munich. Correspondence to: Mohammad Zalbagi Darestani <mz35@rice.edu>, Akshay S. Chaudhari <akshaysc@stanford.edu>, Reinhard Heckel <rh43@rice.edu>. |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | The code to reproduce all results in this paper is available at https://github.com/MLI-lab/Robustness-CS. |
| Open Datasets | Yes | For all experiments, we use the fast MRI dataset (Zbontar et al., 2018), designed for training and evaluating deeplearning-based MRI reconstruction methods. ... Specifically, we test on the Stanford dataset retrieved by collecting all available 18 knee volumes from mridata.org (Epperson et al., 2013). |
| Dataset Splits | No | The paper mentions using a "fast MRI validation set" and "brain validation set", but it does not specify the exact split percentages or sample counts for training, validation, and test sets. It refers to data being partitioned into training and test sets but does not quantify this. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types with speeds, or memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details with version numbers (e.g., Python 3.8, PyTorch 1.9, or specific solver versions). |
| Experiment Setup | No | The paper describes the general methods and approaches (e.g., U-net, Var Net, ℓ1-norm minimization) but does not provide specific experimental setup details such as concrete hyperparameter values (e.g., learning rate, batch size, number of epochs) or detailed training configurations in the main text. It mentions `λ` for ℓ1-minimization but does not provide its tuned value. |