ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2405.13216
19
1

Equipping Transformer with Random-Access Reading for Long-Context Understanding

21 May 2024
Chenghao Yang
Zi Yang
Nan Hua
ArXivPDFHTML
Abstract

Long-context modeling presents a significant challenge for transformer-based large language models (LLMs) due to the quadratic complexity of the self-attention mechanism and issues with length extrapolation caused by pretraining exclusively on short inputs. Existing methods address computational complexity through techniques such as text chunking, the kernel approach, and structured attention, and tackle length extrapolation problems through positional encoding, continued pretraining, and data engineering. These approaches typically require sequential access\textbf{sequential access}sequential access to the document, necessitating reading from the first to the last token. We contend that for goal-oriented reading of long documents, such sequential access is not necessary, and a proficiently trained model can learn to omit hundreds of less pertinent tokens. Inspired by human reading behaviors and existing empirical observations, we propose random access\textbf{random access}random access, a novel reading strategy that enables transformers to efficiently process long documents without examining every token. Experimental results from pretraining, fine-tuning, and inference phases validate the efficacy of our method.

View on arXiv
Comments on this paper