The Other Side of Compression: Measuring Bias in Pruned Transformers
Social media platforms have become popular worldwide. Online discussion forums attract users because of their easy access, speech freedom, and ease of communication. Yet there are also possible negative aspects of such communication, including hostile and hate language. While fast and effective solutions for detecting inappropriate language online are constantly being developed, there is little research focusing on the bias of compressed language models that are commonly used nowadays. In this work, we evaluate bias in compressed models trained on Gab and Twitter hatespeech data and estimate to which extent these pruned models capture the relevant context when classifying the input text as hateful, offensive or neutral. Results of our experiments show that transformer-based encoders with 70% or fewer preserved weights are prone to gender, racial, and religious identity-based bias, even if the performance loss is insignificant. We suggest a supervised attention mechanism to counter bias amplification using ground truth per-token hate speech annotation. The proposed method allows pruning BERT, RoBERTa and their distilled versions up to 50% while preserving 90% of their initial performance according to bias and plausibility scores.
Code (1)
Tasks
Hate Speech DetectionNetwork PruningMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Bias in Pruned Vision Models: In-Depth Analysis and Countermeasures
Pruning - that is, setting a significant subset of the parameters of a neural network to zero - is one of the most popular methods of model compression. Yet, several recent works have raised the issue that pruning may in…
Model CompressionNetwork PruningLinearity-based neural network compression
In neural network compression, most current methods reduce unnecessary parameters by measuring importance and redundancy. To augment already highly optimized existing solutions, we propose linearity-based compression as …
Efficient Neural NetworkNeural Network CompressionData-Free Backbone Fine-Tuning for Pruned Neural Networks
Model compression techniques reduce the computational load and memory consumption of deep neural networks. After the compression operation, e.g. parameter pruning, the model is normally fine-tuned on the original trainin…
2D Human Pose Estimationimage-classificationImage ClassificationModel Compression+3Win the Lottery Ticket via Fourier Analysis: Frequencies Guided Network Pruning
With the remarkable success of deep learning recently, efficient network compression algorithms are urgently demanded for releasing the potential computational power of edge devices, such as smartphones or tablets. Howev…
Knowledge DistillationNetwork PruningPruning Convolutional Neural Networks for Image Instance Retrieval
In this work, we focus on the problem of image instance retrieval with deep descriptors extracted from pruned Convolutional Neural Networks (CNN). The objective is to heavily prune convolutional edges while maintaining r…
Image Instance RetrievalRetrievalTriplet