Comparing Chatbot Performance Enhanced with Persistent Homology

2026-06-29

AI Summary

This study addresses the challenge of limited scalability and data scarcity in privacy-sensitive domains such as mental health, where chatbots often rely on small proprietary datasets. The authors propose a novel approach that leverages persistent homology—a technique from topological data analysis—to enrich input representations without increasing data volume, computational overhead, or compromising privacy. By embedding original text vectors with persistent homology features and integrating localized training with multi-model evaluation, the method demonstrates significant performance improvements over baseline models across multiple metrics. This work achieves a zero-cost, privacy-preserving enhancement in model efficacy, offering a practical solution for deploying robust conversational agents in constrained, sensitive settings.

Abstract

Chatbots have become increasingly prevalent across various domains, offering automated assistance in many areas, especially mental health support. The training is done using extremely large datasets, which are sometimes not available in very specific domains. Moreover, it would sometimes be ideal to train the chatbot with personal information about the patients, which, of course, cannot be done on shared servers since it would violate patient confidentiality. Hence, being able to improve the performance of a chatbot, possibly trained locally and on a restricted dataset, without having to increase the dataset itself, would be extremely beneficial. In this work, we will enhance the input datasets using persistent homology (PH) vectorizations computed from the raw datasets themselves. Then we will compare, across several metrics, the performance of multiple chatbot models with or without the PH enhancement. Our experiments suggest that, while at times the PH enhancement is not particularly beneficial, it sometimes brings remarkable advantages for virtually no cost.

Problem

• chatbot performance
• persistent homology
• limited dataset
• patient confidentiality
• mental health support

Innovation

• persistent hom日消息ology
• topological data analysis
• chatbot enhancement
• data-efficient learning
• privacy-preserving AI