Invariance Principle Meets Information Bottleneck for Out-of-Distribution Generalization

Authors: Kartik Ahuja, Ethan Caballero, Dinghuai Zhang, Jean-Christophe Gagnon-Audet, Yoshua Bengio, Ioannis Mitliagkas, Irina Rish

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We propose an approach that combines both these principles and demonstrate its effectiveness in several experiments.
Researcher Affiliation Academia Mila Quebec AI Institute, Université de Montréal. Correspondence to: kartik.ahuja@mila.quebec.
Pseudocode No The paper does not contain a clearly labeled pseudocode or algorithm block.
Open Source Code Yes The code for experiments can be found at https://github.com/ahujak/IB-IRM.
Open Datasets Yes We use all the datasets in Table 2, Terra Incognita dataset (Beery et al., 2018), and COCO (Ahmed et al., 2021).
Dataset Splits Yes We follow the same protocol for tuning hyperparameters from Aubin et al. (2021); Arjovsky et al. (2019) for their respective datasets (see the Appendix for more details).
Hardware Specification Yes All the experiments were run on a server with an Intel(R) Xeon(R) Gold 6130 CPU @ 2.10GHz processor and NVIDIA Tesla V100 GPU.
Software Dependencies No The paper mentions using code from external GitHub repositories, but it does not specify software dependencies like programming language versions or library versions (e.g., Python 3.x, PyTorch 1.x) that were used for their experiments.
Experiment Setup Yes We follow the same protocol for tuning hyperparameters from Aubin et al. (2021); Arjovsky et al. (2019) for their respective datasets (see the Appendix for more details). For reproducibility, we use a fixed random seed of 0 across all experiments.