A new multi-stage sequential model with selective classifiers is proposed to characterize agent actions and design sequences that incentivize genuine improvement rather than gaming in strategic classification.
Strategic Classification from Revealed Preferences
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
abstract
We study an online linear classification problem, in which the data is generated by strategic agents who manipulate their features in an effort to change the classification outcome. In rounds, the learner deploys a classifier, and an adversarially chosen agent arrives, possibly manipulating her features to optimally respond to the learner. The learner has no knowledge of the agents' utility functions or "real" features, which may vary widely across agents. Instead, the learner is only able to observe their "revealed preferences" --- i.e. the actual manipulated feature vectors they provide. For a broad family of agent cost functions, we give a computationally efficient learning algorithm that is able to obtain diminishing "Stackelberg regret" --- a form of policy regret that guarantees that the learner is obtaining loss nearly as small as that of the best classifier in hindsight, even allowing for the fact that agents will best-respond differently to the optimal classifier.
fields
cs.LG 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
Presents a game-theoretic model with group actions for data augmentation in LLM adversarial evaluation, demonstrating local generalization from fine-tuning on three model families and redefining benchmarks as orbits under group actions.
citing papers explorer
-
Sequential Strategic Classification with Multi-Stage Selective Classifiers
A new multi-stage sequential model with selective classifiers is proposed to characterize agent actions and design sequences that incentivize genuine improvement rather than gaming in strategic classification.
-
The Evaluation Game: Beyond Static LLM Benchmarking
Presents a game-theoretic model with group actions for data augmentation in LLM adversarial evaluation, demonstrating local generalization from fine-tuning on three model families and redefining benchmarks as orbits under group actions.