IncDSI: Incrementally Updatable Document Retrieval

TL;DR

IncDSI introduces a real-time, incremental update method for differentiable search index models, enabling fast addition of new documents without full retraining, thus facilitating dynamic and up-to-date document retrieval systems.

Contribution

The paper presents IncDSI, a novel approach that allows real-time document addition to differentiable search index models through minimal parameter adjustments, avoiding costly retraining.

Findings

IncDSI achieves 20-50ms per document addition.

The method maintains competitive retrieval performance.

It enables real-time updates in document retrieval systems.

Abstract

Differentiable Search Index is a recently proposed paradigm for document retrieval, that encodes information about a corpus of documents within the parameters of a neural network and directly maps queries to corresponding documents. These models have achieved state-of-the-art performances for document retrieval across many benchmarks. These kinds of models have a significant limitation: it is not easy to add new documents after a model is trained. We propose IncDSI, a method to add documents in real time (about 20-50ms per document), without retraining the model on the entire dataset (or even parts thereof). Instead we formulate the addition of documents as a constrained optimization problem that makes minimal changes to the network parameters. Although orders of magnitude faster, our approach is competitive with re-training the model on the whole dataset and enables the development of document retrieval systems that can be updated with new information in real-time. Our code for IncDSI is available at https://github.com/varshakishore/IncDSI.