SEBRA : Debiasing through Self-Guided Bias Ranking

Authors: Adarsh Kappiyath, Abhra Chaudhuri, AJAY JAISWAL, Ziquan Liu, Yunpeng Li, Xiatian Zhu, Lu Yin

ICLR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments show that Sebra consistently outperforms previous state-of-the-art unsupervised debiasing techniques across multiple standard benchmarks, including Urban Cars, BAR, Celeb A, Multi NLI, and Image Net-1K. Code, pre-trained models, and training logs are available at https://kadarsh22.github.io/sebra_iclr25/. [...] 4 EXPERIMENTAL SETUP
Researcher Affiliation Collaboration 1University of Surrey, 2Fujitsu Research of Europe, 3University of Texas at Austin 4Queen Mary University of London
Pseudocode Yes We provide the pseudocode for Sebra in Algorithm 1.
Open Source Code Yes Code, pre-trained models, and training logs are available at https://kadarsh22.github.io/sebra_iclr25/.
Open Datasets Yes We evaluate the proposed ranking and debiasing strategy on one synthetic and four natural datasets with spurious correlations. The synthetic dataset, Urban Cars Li et al. (2023), focuses on car-type classification with spurious correlations involving the background and co-occurring objects. [...] For natural datasets, we use Celeb A Liu et al. (2015), [...] Additionally, we test on two natural vision datasets, BAR (Nam et al., 2020) and Image Net-1K Deng et al. (2009), and the natural language dataset Multi NLI Williams et al. (2018) [...] All the datasets used are publicly available or can be generated with publicly available resources.
Dataset Splits Yes For BAR, no bias annotations are used, even during validation and validation set is obtained by random split of training set in 80:20 ratio.
Hardware Specification Yes All the measurements were done using a single Nvidia RTX 3090 GPU. [...] All experiments were conducted using a single RTX 3090 GPU.
Software Dependencies No The paper mentions using specific ResNet architectures (Res Net-50, Res Net-18) but does not provide specific software dependencies like Python, PyTorch/TensorFlow versions, or CUDA versions.
Experiment Setup Yes Table 7: Optimal hyper-parameters for the BAR, Urban Cars, Celeb A, and Image Net datasets determined through hyper-parameter search. Parameter Urban Cars BAR Celeb A Image Net Learning Rate (LR) 1.0 10 3 1.0 10 4 1.0 10 3 1.0 10 3 batch Size 128 64 64 512 optimiser SGD Adam SGD SGD momentum 0.1 0.8 0.5 weight decay 0.001 0 1.0 10 4 1.0 10 4 pcritical 0.75 0.75 0.7 0.1 β 1.25 1.42 1.25 1.42 γ 0.5 0.5 1 1 τ 0.05 0.15 0.05 0.1