CUE Vectors: Modular Training of Language Models Conditioned on Diverse Contextual Signals

TL;DR

This paper introduces CUE vectors, a modular framework for training language models conditioned on diverse external contexts, enabling flexible adaptation and incremental training without joint encoder training.

Contribution

The authors propose a novel modular training approach that separates sentence and context encoding, allowing easy adaptation to new metadata types and incremental learning.

Findings

Conditioning on context reduces perplexity from 36.6 to 27.4.

Retaining 85% of gains with partial context during training.

Swapping pretrained sentence LMs without retraining encoders is feasible.

Abstract

We propose a framework to modularize the training of neural language models that use diverse forms of sentence-external context (including metadata) by eliminating the need to jointly train sentence-external and within-sentence encoders. Our approach, contextual universal embeddings (CUE), trains LMs on one set of context, such as date and author, and adapts to novel metadata types, such as article title, or previous sentence. The model consists of a pretrained neural sentence LM, a BERT-based context encoder, and a masked transformer decoder that estimates LM probabilities using sentence-internal and sentence-external information. When context or metadata are unavailable, our model learns to combine contextual and sentence-internal information using noisy oracle unigram embeddings as a proxy. Real contextual information can be introduced later and used to adapt a small number of parameters that map contextual data into the decoder's embedding space. We validate the CUE framework on a NYTimes text corpus with multiple metadata types, for which the LM perplexity can be lowered from 36.6 to 27.4 by conditioning on context. Bootstrapping a contextual LM with only a subset of the context/metadata during training retains 85\% of the achievable gain. Training the model initially with proxy context retains 67% of the perplexity gain after adapting to real context. Furthermore, we can swap one type of pretrained sentence LM for another without retraining the context encoders, by only adapting the decoder model. Overall, we obtain a modular framework that allows incremental, scalable training of context-enhanced LMs.