Whitening Sentence Representations for Better Semantics and Faster Retrieval

Abstract

Pre-training models such as BERT have achieved great success in many natural language processing tasks. However, how to obtain better sentence representation through these pre-training models is still worthy to exploit. Previous work has shown that the anisotropy problem is an critical bottleneck for BERT-based sentence representation which hinders the model to fully utilize the underlying semantic features. Therefore, some attempts of boosting the isotropy of sentence distribution, such as flow-based model, have been applied to sentence representations and achieved some improvement. In this paper, we find that the whitening operation in traditional machine learning can similarly enhance the isotropy of sentence representations and achieve competitive results. Furthermore, the whitening technique is also capable of reducing the dimensionality of the sentence representation. Our experimental results show that it can not only achieve promising performance but also significantly reduce the storage cost and accelerate the model retrieval speed.

arXiv commentThe source code of this paper is available at https://github.com/bojone/BERT-whitening

References in corpus (5)

• Language Models are Few-Shot Learners
• NICE: Non-linear Independent Components Estimation
• Universal Sentence Encoder
• Poly-encoders: Transformer Architectures and Pre-training Strategies for Fast and Accurate Multi-sentence Scoring
• Representation Degeneration Problem in Training Natural Language Generation Models

Cited by in corpus (20)

• A Review-aware Graph Contrastive Learning Framework for Recommendation
• TSDAE: Using Transformer-based Sequential Denoising Auto-Encoder for Unsupervised Sentence Embedding Learning
• GraphFormers: GNN-nested Transformers for Representation Learning on Textual Graph
• SimCSE: Simple Contrastive Learning of Sentence Embeddings
• Improving Speech Translation by Cross-Modal Multi-Grained Contrastive Learning
• Improving Contrastive Learning of Sentence Embeddings with Case-Augmented Positives and Retrieved Negatives
• SimCSE++: Improving Contrastive Learning for Sentence Embeddings from Two Perspectives
• DebCSE: Rethinking Unsupervised Contrastive Sentence Embedding Learning in the Debiasing Perspective
• DistillCSE: Distilled Contrastive Learning for Sentence Embeddings
• Towards human-compatible autonomous car: A study of non-verbal Turing test in automated driving with affective transition modelling
• Leveraging Advantages of Interactive and Non-Interactive Models for Vector-Based Cross-Lingual Information Retrieval
• Learnable Adaptive Cosine Estimator (LACE) for Image Classification
• Positional Artefacts Propagate Through Masked Language Model Embeddings
• DialogueCSE: Dialogue-based Contrastive Learning of Sentence Embeddings
• Emotions are Subtle: Learning Sentiment Based Text Representations Using Contrastive Learning
• Identical and Fraternal Twins: Fine-Grained Semantic Contrastive Learning of Sentence Representations
• Text analysis and deep learning: A network approach
• Aligning Cross-lingual Sentence Representations with Dual Momentum Contrast
• Towards Zero and Few-shot Knowledge-seeking Turn Detection in Task-orientated Dialogue Systems
• WhiteningBERT: An Easy Unsupervised Sentence Embedding Approach