Federated Learning Partner Selection Using Client Similarity

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing federated learning systems face challenges in appropriately selecting partners based on client-related information, particularly when clients are working on different tasks, leading to uncertainty in model accuracy.

Innovation Solution

An information processing apparatus and method that evaluates the degree of similarity between client-related information to facilitate the selection of appropriate partners for federated learning, using client-related information generated based on information registered by clients, and outputs screen data for partner selection.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If partner selection is based only on client-related information, then the selection process is simple, but the accuracy of partner selection deteriorates

Engineering Contradiction:

Improvepartner selection processVSAvoidpartner selection accuracy

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The partner selection process is segmented into multiple stages: initial filtering based on client-related information (business type, task type, number of data items) and subsequent evaluation based on data set feature similarity. This segmentation allows the system to maintain operational simplicity while improving selection accuracy through progressive refinement.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system introduces an intermediary evaluation mechanism that calculates similarity between data set features of the own client and candidate partners. This intermediary similarity evaluation acts as a bridge between simple client information matching and accurate partner selection, enabling the system to assess compatibility without directly comparing sensitive client data.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Adaptability or versatility

If federated learning is conducted among clients working on different tasks, then the system's versatility is improved, but the reliability of model accuracy deteriorates

Engineering Contradiction:

Improvefederated learning compatibilityVSAvoidmodel accuracy

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The system changes the evaluation parameters from task-type matching to data set feature similarity. By comparing feature values extracted from data sets (such as data distribution characteristics, feature dimensions, and statistical properties) rather than task classifications, the system can identify suitable partners across different tasks based on underlying data compatibility, thereby maintaining model accuracy while enabling cross-task federated learning.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentEP4693110A1Information processing apparatus, information processing system, information processing method, and carrier medium

Publication Date: 2026.02.11 RICOH CO LTD

Figure 1
Figure 2
Figure 3

AI summary

An information processing apparatus (10) to communicate, via a network, with another information processing apparatus (20) participating in federated learning includes an acquisition unit (103) to acquire client-related information generated based on information registered by a client from the another information processing apparatus (20), an evaluation unit (105) to evaluate a degree of similarity between the client-related information on the another information processing apparatus (20) and the client-related information on one or more partner candidates for the federated learning, and an output unit (106) to output screen data for displaying a screen allowing selection of one or more partners for the federated learning, based on the degree of similarity.