Contextual Lensing of Universal Sentence Representations
arXiv:2002.08866
Abstract
What makes a universal sentence encoder universal? The notion of a generic encoder of text appears to be at odds with the inherent contextualization and non-permanence of language use in a dynamic world. However, mapping sentences into generic fixed-length vectors for downstream similarity and retrieval tasks has been fruitful, particularly for multilingual applications. How do we manage this dilemma? In this work we propose Contextual Lensing, a methodology for inducing context-oriented universal sentence vectors. We break the construction of universal sentence vectors into a core, variable length, sentence matrix representation equipped with an adaptable `lens' from which fixed-length vectors can be induced as a function of the lens context. We show that it is possible to focus notions of language similarity into a small number of lens parameters given a core universal matrix representation. For example, we demonstrate the ability to encode translation similarity of sentences across several languages into a single weight matrix, even when the core encoder has not seen parallel data.
10 pages
References in corpus (10)
- Skip-Thought Vectors
- Word Translation Without Parallel Data
- Insertion Transformer: Flexible Sequence Generation via Insertion Operations
- Parameter-Efficient Transfer Learning for NLP
- How multilingual is Multilingual BERT?
- BERT and PALs: Projected Attention Layers for Efficient Adaptation in Multi-Task Learning
- No Training Required: Exploring Random Encoders for Sentence Classification
- Zero-shot Text Classification With Generative Language Models
- KERMIT: Generative Insertion-Based Modeling for Sequences
- Learning Compressed Sentence Representations for On-Device Text Processing