paper-with-me

Papers

The Low-Dimensional Linear Geometry of Contextualized Word Representations

2021-05-15 · CoNLL (EMNLP) 2021 11 · Evan Hernandez, Jacob Andreas

Black-box probing models can reliably extract linguistic features like tense, number, and syntactic role from pretrained word representations. However, the manner in which these features are encoded in representations remains poorly understood. We present a systematic study of the linear geometry of contextualized word representations in ELMO and BERT. We show that a variety of linguistic features (including structured dependency relationships) are encoded in low-dimensional subspaces. We then refine this geometric picture, showing that there are hierarchical relations between the subspaces encoding general linguistic categories and more specific ones, and that low-dimensional feature encodings are distributed rather than aligned to individual neurons. Finally, we demonstrate that these linear subspaces are causally related to model behavior, and can be used to perform fine-grained manipulation of BERT's output distribution.

📄 PDF Abstract BibTeX arXiv:2105.07109

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Sigmoid Activation 설명 없음
Tanh Activation 설명 없음
LSTM An LSTM is a type of recurrent neural network that addresses the vanishing gradient problem in vanilla…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Linear Warmup With Linear Decay Linear Warmup With Linear Decay is a learning rate schedule in which we increase the learning rate linearly for $n$ updates and then linearly decay afterwards.

Similar Papers 제목 키워드 기반

How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings

2019-09-02 · IJCNLP 2019 11 · Kawin Ethayarajh

Replacing static word embeddings with contextualized word representations has yielded significant improvements on many NLP tasks. However, just how contextual are the contextualized representations produced by models suc…

Word Embeddings

CX-ST-RNM at SemEval-2019 Task 3: Fusion of Recurrent Neural Networks Based on Contextualized and Static Word Representations for Contextual Emotion Detection

2019-06-01 · SEMEVAL 2019 6 · Micha{\l} Pere{\l}kiewicz

In this paper, I describe a fusion model combining contextualized and static word representations for approaching the EmoContext task in the SemEval 2019 competition. The model is based on two Recurrent Neural Networks, …

Frequency-based Distortions in Contextualized Word Embeddings

2021-04-17 · Kaitlyn Zhou, Kawin Ethayarajh, Dan Jurafsky

How does word frequency in pre-training data affect the behavior of similarity metrics in contextualized BERT embeddings? Are there systematic ways in which some word relationships are exaggerated or understated? In this…

Semantic SimilaritySemantic Textual SimilarityWord Embeddings

Improved Word Sense Disambiguation Using Pre-Trained Contextualized Word Representations

2019-10-01 · IJCNLP 2019 11 · Christian Hadiwinoto, Hwee Tou Ng, Wee Chung Gan

Contextualized word representations are able to give different representations for the same word in different contexts, and they have been shown to be effective in downstream natural language processing tasks, such as qu…

named-entity-recognitionNamed Entity RecognitionNamed Entity Recognition (NER)Question Answering+3

Contrastive Learning for Low Resource Machine Translation

2021-11-16 · ACL ARR November 2021 11 · Anonymous

Representation learning plays a vital role in natural language processing tasks. More recent works study the geometry of the representation space for each layer of pre-trained language models. They find that the context…

Contrastive LearningData AugmentationLanguage ModelingLanguage Modelling+3