Hidden Incentives for Auto-Induced Distributional Shift
Decisions made by machine learning systems have increasing influence on the world, yet it is common for machine learning algorithms to assume that no such influence exists. An example is the use of the i.i.d. assumption in content recommendation. In fact, the (choice of) content displayed can change users' perceptions and preferences, or even drive them away, causing a shift in the distribution of users. We introduce the term auto-induced distributional shift (ADS) to describe the phenomenon of an algorithm causing a change in the distribution of its own inputs. Our goal is to ensure that machine learning systems do not leverage ADS to increase performance when doing so could be undesirable. We demonstrate that changes to the learning algorithm, such as the introduction of meta-learning, can cause hidden incentives for auto-induced distributional shift (HI-ADS) to be revealed. To address this issue, we introduce `unit tests' and a mitigation strategy for HI-ADS, as well as a toy environment for modelling real-world issues with HI-ADS in content recommendation, where we demonstrate that strong meta-learners achieve gains in performance via ADS. We show meta-learning and Q-learning both sometimes fail unit tests, but pass when using our mitigation strategy.
Code (0)
등록된 구현이 없습니다.
Tasks
BIG-bench Machine LearningMeta-LearningQ-LearningMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Hidden incentives for self-induced distributional shift
Decisions made by machine learning systems have increasing influence on the world. Yet it is common for machine learning algorithms to assume that no such influence exists. An example is the use of the i.i.d. assumption …
BIG-bench Machine LearningMeta-LearningRevealing the Incentive to Cause Distributional Shift
Decisions made by machine learning systems have increasing influence on the world, yet it is common for machine learning algorithms to assume that no such influence exists. An example is the use of the i.i.d. assumption …
Meta-LearningDistributionally Robust Imitation Learning: Layered Control Architecture for Certifiable Autonomy
Imitation learning (IL) enables autonomous behavior by learning from expert demonstrations. While more sample-efficient than comparative alternatives like reinforcement learning, IL is sensitive to compounding errors ind…
Reinforcement LearningThe Hidden Puppet Master: Predicting Human Belief Change in Manipulative LLM Dialogues
As users increasingly turn to LLMs for practical and personal advice, they become vulnerable to subtle steering toward hidden incentives misaligned with their own interests. While existing NLP research has benchmarked ma…
Distributional Shift-Aware Off-Policy Interval Estimation: A Unified Error Quantification Framework
We study high-confidence off-policy evaluation in the context of infinite-horizon Markov decision processes, where the objective is to establish a confidence interval (CI) for the target policy value using only offline d…
Off-policy evaluation