NeurIPS 2023 Competition: Privacy Preserving Federated Learning Document VQA

TL;DR

The paper discusses the NeurIPS 2023 competition on privacy-preserving federated learning for document VQA, focusing on invoice processing, with solutions balancing privacy, communication efficiency, and utility.

Contribution

It introduces a new dataset and challenge for federated document VQA, fostering development of privacy-preserving methods and providing best practices for future federated learning competitions.

Findings

Participants developed methods to reduce communication costs.

Solutions incorporated differential privacy to protect sensitive data.

The competition established a new benchmark for privacy-preserving federated document analysis.

Abstract

The Privacy Preserving Federated Learning Document VQA (PFL-DocVQA) competition challenged the community to develop provably private and communication-efficient solutions in a federated setting for a real-life use case: invoice processing. The competition introduced a dataset of real invoice documents, along with associated questions and answers requiring information extraction and reasoning over the document images. Thereby, it brings together researchers and expertise from the document analysis, privacy, and federated learning communities. Participants fine-tuned a pre-trained, state-of-the-art Document Visual Question Answering model provided by the organizers for this new domain, mimicking a typical federated invoice processing setup. The base model is a multi-modal generative language model, and sensitive information could be exposed through either the visual or textual input modality. Participants proposed elegant solutions to reduce communication costs while maintaining a minimum utility threshold in track 1 and to protect all information from each document provider using differential privacy in track 2. The competition served as a new testbed for developing and testing private federated learning methods, simultaneously raising awareness about privacy within the document image analysis and recognition community. Ultimately, the competition analysis provides best practices and recommendations for successfully running privacy-focused federated learning challenges in the future.