Papers with anisotropy
Anisotropy Is Inherent to Self-Attention in Transformers (2024.eacl-long)
Copied to clipboard
| Challenge: | despite their success, Transformers models suffer from a representation degeneration problem . anisotropy is a property of hidden representations that makes them unexpectedly close to each other . |
| Approach: | They investigate the representation degeneration problem in a self-supervised learning model based on Transformers . anisotropy is a property of hidden representations which makes them unexpectedly close to each other . |
| Outcome: | The representation degeneration problem is a phenomenon widely observed among self-supervised learning methods based on Transformers. |
Is Anisotropy Truly Harmful? A Case Study on Text Clustering (2023.acl-short)
Copied to clipboard
| Challenge: | Contextualized pre-trained representations are widely used as input to various tasks such as information retrieval, anomaly detection and document clustering. |
| Approach: | They propose to examine the impact of different transformations on isotropy and performance to assess the true impact of anisotropi. |
| Outcome: | The proposed model is based on a clustering task and shows that it has limited impact on expressiveness and closeness. |
Anisotropy is Not Inherent to Transformers (2024.naacl-long)
Copied to clipboard
| Challenge: | Existing work has shown that Transformer embeddings are anisotropic, which is called the representation degradation problem. |
| Approach: | They identify a set of Transformer models with isotropic embedding spaces, the large Pythia models. |
| Outcome: | The proposed model sets show that isotropic models do not develop as previously theorized. |
Is anisotropy really the cause of BERT embeddings not being semantic? (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing approaches to train contextual language models for NLP use a lightweight approach called bi-encoder, which takes two sentences as input, but does not perform well with vanilla pre-trained Transformers. |
| Approach: | They conduct a set of experiments to improve our understanding of the lack of semantic isometry in contextualized word representations in BERT. |
| Outcome: | The proposed approach does not perform well with vanilla pre-trained Transformers. |
Mitigating Frequency Bias and Anisotropy in Language Model Pre-Training with Syntactic Smoothing (2024.emnlp-main)
Copied to clipboard
| Challenge: | Language models rely on frequency information because they maximize the likelihood of tokens during training. |
| Approach: | They propose a method for quantifying the frequency bias of a language model by assessing sentence-level perplexity with respect to token-level frequency. |
| Outcome: | The proposed method reduces the frequency bias of a language model by inducing a syntactic prior over token representations during pre-training. |
HIL: Hybrid Isotropy Learning for Zero-shot Performance in Dense retrieval (2024.naacl-long)
Copied to clipboard
| Challenge: | Recent advances in dense retrieval models have brought ColBERT to prominence in information retrieval, but it is underperforming in zero-shot tasks. |
| Approach: | They propose a Hybrid Isotropy Learning architecture that integrates isotropic and anisotropic representations to improve zero-shot retrieval performance. |
| Outcome: | The proposed model outperforms the baseline ColBERT model in BEIR benchmarks. |
Better Embeddings with Coupled Adam (2025.acl-long)
Copied to clipboard
| Challenge: | Anisotropic Embeddings Large Language Models exhibit undesirable yet poorly understood feature of anisotropy. |
| Approach: | They propose an algorithm that uses the second moment in Adam to mitigate anisotropic embeddings . they propose an embeddable matrix and unembedding matrix to map the input and output tokens based on weight tying . |
| Outcome: | The proposed model improves quality and performance on large datasets. |