Towards Trustworthy Explanation: On Causal Rationalization

Authors: Wenbo Zhang, Tong Wu, Yunlong Wang, Yong Cai, Hengrui Cai

ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The superior performance of the proposed causal rationalization is demonstrated on real-world review and medical datasets with extensive experiments compared to state-of-the-art methods.
Researcher Affiliation Collaboration 1Department of Statistics, University of California Irvine, California, USA 2Advanced Analytics, IQVIA, Pennsylvania, USA.
Pseudocode Yes Algorithm 1 Causal Rationalization
Open Source Code Yes Our code is publicly available online.1
Open Datasets Yes Beer Review Data. We use the publicly available version of the Beer review dataset also adopted by Bao et al. (2018) and Chen et al. (2022).
Dataset Splits Yes We follow the same train/validation/test split as Chen et al. (2022) and it is summarized in Table 5.
Hardware Specification Yes All of our experiments are conducted with PyTorch on 4 V100 GPU.
Software Dependencies No The paper mentions "PyTorch" and "BERT-base-uncased" but does not specify their version numbers.
Experiment Setup Yes For all experiments, we utilize a batch size of 256 and choose the learning rate α {1e-5, 5e-4, 1e-4}. We train for 10 epochs all the datasets. For training the causal component, we tune the values of the Lagrangian multiplier µ {0.01, 0.1, 1} and set k = 5. We set the temperature of Gumbel-softmax to be 0.5.