[Paper] Doc-to-Atom: Learning to Compile and Compose Memory Atoms

Published: (June 10, 2026 at 01:58 PM EDT)
2 min read
Source: arXiv

Source: arXiv - 2606.12400v1

Overview

Long input sequences are central to document understanding and multi-step reasoning in Large Language Models, yet the quadratic cost of attention makes inference both memory-intensive and slow. Context distillation mitigates this by compressing contextual information into model parameters, and recent work such as Doc-to-LoRA amortizes context distillation into a single forward pass that generates one LoRA adapter per document. However, producing a single monolithic adapter for all queries leads to irrelevant-query interference, limited compositional recall, and poor scalability to long-document reasoning. To address these challenges, we propose Doc-to-Atom (Doc2Atom), a compositional parametric memory framework that decomposes each document into semantically typed knowledge atoms. Each atom is compiled into an independent micro-LoRA adapter and a provenance retrieval key. At inference time, a lightweight query router selects and assembles only the relevant atoms into a query-specific adapter, which is then injected into a frozen base model. The entire system is trained end-to-end through a multi-objective distillation framework. Experiments on six diverse QA benchmarks demonstrate that Doc2Atom outperforms Doc-to-LoRA baselines while reducing the memory cost of document internalization.

Key Contributions

This paper presents research in the following areas:

  • cs.CL
  • cs.IR

Methodology

Please refer to the full paper for detailed methodology.

Practical Implications

This research contributes to the advancement of cs.CL.

Authors

  • Xingjian Diao
  • Wenbo Li
  • Yashas Malur Saidutta
  • Avinash Amballa
  • Lazar Valkov
  • Srinivas Chappidi

Paper Information

  • arXiv ID: 2606.12400v1
  • Categories: cs.CL, cs.IR
  • Published: June 10, 2026
  • PDF: Download PDF
0 views
Back to Blog

Related posts

Read more »