NextLevelBERT: Masked Language Modeling with Higher-Level Representations for Long Documents
This addresses the problem of handling long sequences in AI for applications like book analysis, though it is incremental as it builds on existing masked language modeling techniques.
The paper tackles the challenge of processing long documents in language models by proposing NextLevelBERT, a masked language model that operates on higher-level semantic representations instead of tokens, and finds it outperforms larger embedding models on tasks like semantic textual similarity, long document classification, and multiple-choice question answering when fine detail is not required.
While (large) language models have significantly improved over the last years, they still struggle to sensibly process long sequences found, e.g., in books, due to the quadratic scaling of the underlying attention mechanism. To address this, we propose NextLevelBERT, a Masked Language Model operating not on tokens, but on higher-level semantic representations in the form of text embeddings. We pretrain NextLevelBERT to predict the vector representation of entire masked text chunks and evaluate the effectiveness of the resulting document vectors on three types of tasks: 1) Semantic Textual Similarity via zero-shot document embeddings, 2) Long document classification, 3) Multiple-choice question answering. We find that next-level Masked Language Modeling is an effective technique to tackle long-document use cases and can outperfor much larger embedding models as long as the required level of detail of semantic information is not too fine. Our models and code are publicly available online.