aiqbench analysis
This research introduces a rigorous framework for measuring model responsiveness, quantifying how often predictions shift under input variations caused by user interaction or noise. By providing statistical guarantees to validate these shifts across domains like recidivism prediction and content moderation, the work directly addresses safety concerns regarding models operating outside their training distributions. The methodology offers practical tools for assessing reliability in dynamic environments where strategic manipulation or routine feedback loops are present.
Responsiveness Verification: Will Predictions Change? How Much? How Often?
arXiv:2507.02169v2 Announce Type: replace Abstract: Machine learning models are often used in applications where their inputs change due to routine interactions, strategic manipulation, or noise. In such settings, models can undermine safety as these changes lead them to predict over regions of input space they have not seen. We propose to address these challenges by measuring responsiveness---the probability that a model output changes when its inputs change under an interaction model. We develop algorithms to estimate responsiveness for any machine learning model, and a framework to specify broad classes of interaction models. We pair these algorithms with statistical guarantees that support practical validation. We demonstrate how our tools can promote safety and reliability across domains by detecting preclusion in recidivism prediction, estimating the cost of gaming in content moderation, and testing the robustness of benchmarks for LLMs.
Title:Responsiveness Verification: Will Predictions Change? How Much? How Often?
Abstract:Machine learning models are often used in applications where their inputs change due to routine interactions, strategic manipulation, or noise. In such settings, models can undermine safety as these changes lead them to predict over regions of input space they have not seen. We propose to address these challenges by measuring responsiveness---the probability that a model output changes when its inputs change under an interaction model. We develop algorithms to estimate responsiveness for any machine learning model, and a framework to specify broad classes of interaction models. We pair these algorithms with statistical guarantees that support practical validation. We demonstrate how our tools can promote safety and reliability across domains by detecting preclusion in recidivism prediction, estimating the cost of gaming in content moderation, and testing the robustness of benchmarks for LLMs.
Machine Learning (cs.LG)
arXiv:2507.02169 [cs.LG]
(or arXiv:2507.02169v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2507.02169 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 233 words
