Flat Seeking Bayesian Neural Networks
Authors: Van-Anh Nguyen, Tung-Long Vuong, Hoang Phan, Thanh-Toan Do, Dinh Phung, Trung Le
NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We conduct extensive experiments by leveraging our sharpness-aware posterior with the state-of-the-art and well-known BNNs... In this section, we conduct various experiments to demonstrate the effectiveness of the sharpness-aware approach on Bayesian Neural networks... Our experimental results, presented in Tables 1, 2, 3 for CIFAR-100 and CIFAR-10 dataset, and Table 4 for the Image Net dataset, indicate a notable improvement across all experiments. |
| Researcher Affiliation | Collaboration | Van-Anh Nguyen1 Tung-Long Vuong1,2 Hoang Phan2,3 Thanh-Toan Do1 Dinh Phung 1,2 Trung Le 1 1Department of Data Science and AI, Monash University, Australia 2Vin AI, Vietnam 3New York University, United States |
| Pseudocode | No | The paper does not contain any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | The implementation is provided in https://github.com/anh-ntv/flat_bnn.git |
| Open Datasets | Yes | The experiments are conducted on three benchmark datasets: CIFAR-10, CIFAR-100, and Image Net ILSVRC-2012, and report accuracy, negative log-likelihood (NLL), and Expected Calibration Error (ECE) to estimate the calibration capability and uncertainty of our method against baselines. |
| Dataset Splits | No | The paper states 'The details of the dataset and implementation are described in the supplementary material', but does not provide specific train/validation/test dataset splits (e.g., percentages, sample counts, or explicit references to predefined splits) within the main text. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with versions). |
| Experiment Setup | No | The paper states 'The details of the dataset and implementation are described in the supplementary material' and mentions 'hyper-parameter settings' in the ablation studies ('under the same hyper-parameter settings'). However, concrete hyperparameter values or detailed training configurations are not provided within the main text of the paper. |