Towards Multi-dimensional Explanation Alignment for Medical Classification
Authors: Lijie Hu, Songning Lai, Wenshuo Chen, Hongru Xiao, Hongbin Lin, Lu Yu, Jingfeng ZHANG, Di Wang
NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | To demonstrate the effectiveness and interpretability of Med-MICN, we apply it to four benchmark datasets and compare it with baselines. The results clearly demonstrate the superior performance and interpretability of our Med-MICN. |
| Researcher Affiliation | Collaboration | Lijie Hu ,1,2, Songning Lai ,1,2,3, Wenshuo Chen ,1,2, Hongru Xiao4 Hongbin Lin3, Lu Yu6, Jingfeng Zhang5,7, and Di Wang1,2 1Provable Responsible AI and Data Analytics (PRADA) Lab 2King Abdullah University of Science and Technology 3HKUST(GZ) 4Tongji University 5The University of Auckland 6Ant Group, 7RIKEN Center for Advanced Intelligence Project (AIP) |
| Pseudocode | No | The paper does not contain any explicitly labeled 'Pseudocode' or 'Algorithm' blocks. |
| Open Source Code | No | Code [Yes] DATA [No] Justification: We will soon open-source the code, and the datasets used are publicly available for download by anyone. |
| Open Datasets | Yes | Datasets. We consider four benchmark medical datasets: COVID-CT [29] for CT images, DDI [10] for dermatology images, Chest X-Ray [14], and Fitzpatrick17k [15] for a dermatological dataset with skin colors. |
| Dataset Splits | No | For each dataset, the paper specifies training and test splits (e.g., 'We divided this dataset into a training set and a test set with an 8:2 ratio'), but does not explicitly define a separate validation dataset split. |
| Hardware Specification | Yes | We utilized only a single Ge Force RTX 4090 GPU, and the training duration did not exceed half an hour. |
| Software Dependencies | No | The paper mentions models like ResNet-50 and optimizers like Adam, but it does not specify software dependencies such as programming languages, libraries, or frameworks with their version numbers (e.g., Python, PyTorch, TensorFlow versions). |
| Experiment Setup | Yes | During training, we adopted the Adam optimizer with a learning rate of 5e-5 throughout the training stages. For hyperparameter selection, we set λ1 = λ2 = 0.1. Additionally, all images were resized to a uniform dimension of (256, 256). |