On Monotonicity in AI Alignment

• URL: http://arxiv.org/abs/2506.08998v1
• Date: Tue, 10 Jun 2025 17:17:48 GMT
• Title: On Monotonicity in AI Alignment
• Authors: Gilles Bareilles, Julien Fageot, Lê-Nguyên Hoang, Peva Blanchard, Wassim Bouaziz, Sébastien Rouault, El-Mahdi El-Mhamdi,
• Abstract summary: This paper investigates the root causes of (non) monotonicity, for a general comparison-based preference learning framework.<n>Under mild assumptions, we prove that such methods still satisfy what we call local pairwise monotonicity.<n>We also provide a bouquet of formalizations of monotonicity, and identify sufficient conditions for their guarantee, thereby providing a toolbox to evaluate how prone learning models are to monotonicity violations.
• Score: 10.244128221542228
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Comparison-based preference learning has become central to the alignment of AI models with human preferences. However, these methods may behave counterintuitively. After empirically observing that, when accounting for a preference for response $y$ over $z$, the model may actually decrease the probability (and reward) of generating $y$ (an observation also made by others), this paper investigates the root causes of (non) monotonicity, for a general comparison-based preference learning framework that subsumes Direct Preference Optimization (DPO), Generalized Preference Optimization (GPO) and Generalized Bradley-Terry (GBT). Under mild assumptions, we prove that such methods still satisfy what we call local pairwise monotonicity. We also provide a bouquet of formalizations of monotonicity, and identify sufficient conditions for their guarantee, thereby providing a toolbox to evaluate how prone learning models are to monotonicity violations. These results clarify the limitations of current methods and provide guidance for developing more trustworthy preference learning algorithms.

Related papers

• Learning Parametric Distributions from Samples and Preferences [19.879505582147807]
  We show that preference-based M-estimators achieve a better variance than sample-only M-estimators.<n>We propose an estimator achieving an estimation error scaling of $mathcalO (1/n)$ -- a significant improvement over the $Theta (1/sqrtn)$ rate attainable with samples alone.
  arXiv  Detail & Related papers  (2025-05-29T15:33:43Z)
• Achieving $\widetilde{\mathcal{O}}(\sqrt{T})$ Regret in Average-Reward POMDPs with Known Observation Models [56.92178753201331]
  We tackle average-reward infinite-horizon POMDPs with an unknown transition model.<n>We present a novel and simple estimator that overcomes this barrier.
  arXiv  Detail & Related papers  (2025-01-30T22:29:41Z)
• Rethinking Bradley-Terry Models in Preference-Based Reward Modeling: Foundations, Theory, and Alternatives [14.401557416713315]
  We revisit the foundations of using Bradley-Terry (BT) models in reward modeling.<n>We argue that the BT model is not a necessary choice from the perspective of downstream optimization.<n>We propose a simple and straightforward upper-bound algorithm, compatible with off-the-shelf binary classifiers.
  arXiv  Detail & Related papers  (2024-11-07T18:57:03Z)
• Learning from negative feedback, or positive feedback or both [21.95277469346728]
  We introduce a novel approach that decouples learning from positive and negative feedback.<n>A key contribution is demonstrating stable learning from negative feedback alone.
  arXiv  Detail & Related papers  (2024-10-05T14:04:03Z)
• Beyond Bradley-Terry Models: A General Preference Model for Language Model Alignment [51.14207112118503]
  We introduce preference embedding, an approach that embeds responses into a latent space to capture preferences efficiently.<n>We also propose preference score-based General Preference Optimization (GPO), which generalizes reward-based reinforcement learning from human feedback.<n>Our method may enhance the alignment of foundation models with nuanced human values.
  arXiv  Detail & Related papers  (2024-10-03T04:22:55Z)
• Robust Preference Optimization through Reward Model Distillation [68.65844394615702]
  Direct Preference Optimization (DPO) is a popular offline alignment method that trains a policy directly on preference data.<n>We analyze this phenomenon and use distillation to get a better proxy for the true preference distribution over generation pairs.<n>Our results show that distilling from such a family of reward models leads to improved robustness to distribution shift in preference annotations.
  arXiv  Detail & Related papers  (2024-05-29T17:39:48Z)
• $i$REPO: $i$mplicit Reward Pairwise Difference based Empirical Preference Optimization [12.266207199002604]
  Large Language Models (LLM) can sometimes produce outputs that deviate from human expectations. We propose a novel framework named $i$REPO, which utilizes implicit Reward pairwise difference regression for Empirical Preference Optimization. We show that $i$REPO effectively achieves self-alignment using soft-label, self-generated responses and the logit of empirical AI annotators.
  arXiv  Detail & Related papers  (2024-05-24T05:42:11Z)
• Optimal Baseline Corrections for Off-Policy Contextual Bandits [61.740094604552475]
  We aim to learn decision policies that optimize an unbiased offline estimate of an online reward metric. We propose a single framework built on their equivalence in learning scenarios. Our framework enables us to characterize the variance-optimal unbiased estimator and provide a closed-form solution for it.
  arXiv  Detail & Related papers  (2024-05-09T12:52:22Z)
• Model-based Offline Imitation Learning with Non-expert Data [7.615595533111191]
  We propose a scalable model-based offline imitation learning algorithmic framework that leverages datasets collected by both suboptimal and optimal policies. We show that the proposed method textitalways outperforms Behavioral Cloning in the low data regime on simulated continuous control domains.
  arXiv  Detail & Related papers  (2022-06-11T13:08:08Z)
• Distributionally Robust Models with Parametric Likelihood Ratios [123.05074253513935]
  Three simple ideas allow us to train models with DRO using a broader class of parametric likelihood ratios. We find that models trained with the resulting parametric adversaries are consistently more robust to subpopulation shifts when compared to other DRO approaches.
  arXiv  Detail & Related papers  (2022-04-13T12:43:12Z)
• Approximate Bayesian inference from noisy likelihoods with Gaussian process emulated MCMC [0.24275655667345403]
  We model the log-likelihood function using a Gaussian process (GP) The main methodological innovation is to apply this model to emulate the progression that an exact Metropolis-Hastings (MH) sampler would take. The resulting approximate sampler is conceptually simple and sample-efficient.
  arXiv  Detail & Related papers  (2021-04-08T17:38:02Z)
• Control as Hybrid Inference [62.997667081978825]
  We present an implementation of CHI which naturally mediates the balance between iterative and amortised inference. We verify the scalability of our algorithm on a continuous control benchmark, demonstrating that it outperforms strong model-free and model-based baselines.
  arXiv  Detail & Related papers  (2020-07-11T19:44:09Z)
• Learning the Truth From Only One Side of the Story [58.65439277460011]
  We focus on generalized linear models and show that without adjusting for this sampling bias, the model may converge suboptimally or even fail to converge to the optimal solution. We propose an adaptive approach that comes with theoretical guarantees and show that it outperforms several existing methods empirically.
  arXiv  Detail & Related papers  (2020-06-08T18:20:28Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.