Fisher Random Walk: Automatic Debiasing Contextual Preference Inference for Large Language Model Evaluation

• URL: http://arxiv.org/abs/2509.05852v1
• Date: Sat, 06 Sep 2025 22:29:17 GMT
• Title: Fisher Random Walk: Automatic Debiasing Contextual Preference Inference for Large Language Model Evaluation
• Authors: Yichi Zhang, Alexander Belloni, Ethan X. Fang, Junwei Lu, Xiaoan Xu,
• Abstract summary: We develop a semiparametric efficient estimator that automates the debiased estimation.<n>We show that the efficiency is achieved when the weights are derived from a novel strategy called Fisher random walk.
• Score: 46.643610591694376
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Motivated by the need for rigorous and scalable evaluation of large language models, we study contextual preference inference for pairwise comparison functionals of context-dependent preference score functions across domains. Focusing on the contextual Bradley-Terry-Luce model, we develop a semiparametric efficient estimator that automates the debiased estimation through aggregating weighted residual balancing terms across the comparison graph. We show that the efficiency is achieved when the weights are derived from a novel strategy called Fisher random walk. We also propose a computationally feasible method to compute the weights by a potential representation of nuisance weight functions. We show our inference procedure is valid for general score function estimators accommodating the practitioners' need to implement flexible deep learning methods. We extend the procedure to multiple hypothesis testing using a Gaussian multiplier bootstrap that controls familywise error and to distributional shift via a cross-fitted importance-sampling adjustment for target-domain inference. Numerical studies, including language model evaluations under diverse contexts, corroborate the accuracy, efficiency, and practical utility of our method.

Related papers

• Efficient Covariance Estimation for Sparsified Functional Data [51.69796254617083]
  proposed Random-knots (Random-knots-Spatial) and B-spline (Bspline-Spatial) estimators of the covariance function are computationally efficient.<n>Asymptotic pointwise of the covariance are obtained for sparsified individual trajectories under some regularity conditions.
  arXiv  Detail & Related papers  (2025-11-23T00:50:33Z)
• Entropy-regularized Gradient Estimators for Approximate Bayesian Inference [2.44755919161855]
  This paper addresses the estimation of the Bayesian posterior to generate diverse samples by approximating the gradient flow of the Kullback-Leibler divergence.<n>It presents empirical evaluations on classification tasks to assess the method's performance and discuss its effectiveness for Model-Based Reinforcement Learning.
  arXiv  Detail & Related papers  (2025-03-15T02:30:46Z)
• Model-free Methods for Event History Analysis and Efficient Adjustment (PhD Thesis) [55.2480439325792]
  This thesis is a series of independent contributions to statistics unified by a model-free perspective.<n>The first chapter elaborates on how a model-free perspective can be used to formulate flexible methods that leverage prediction techniques from machine learning.<n>The second chapter studies the concept of local independence, which describes whether the evolution of one process is directly influenced by another.
  arXiv  Detail & Related papers  (2025-02-11T19:24:09Z)
• In2Core: Leveraging Influence Functions for Coreset Selection in Instruction Finetuning of Large Language Models [37.45103473809928]
  We propose the In2Core algorithm, which selects a coreset by analyzing the correlation between training and evaluation samples with a trained model. By applying our algorithm to instruction fine-tuning data of LLMs, we can achieve similar performance with just 50% of the training data.
  arXiv  Detail & Related papers  (2024-08-07T05:48:05Z)
• Online Estimation with Rolling Validation: Adaptive Nonparametric Estimation with Streaming Data [13.069717985067937]
  We propose a weighted rolling validation procedure, an online variant of leave-one-out cross-validation, that costs minimal extra for many typical gradient descent estimators.<n>Our analysis is straightforward, relying mainly on some general statistical assumptions.
  arXiv  Detail & Related papers  (2023-10-18T17:52:57Z)
• Boosting Fair Classifier Generalization through Adaptive Priority Reweighing [59.801444556074394]
  A performance-promising fair algorithm with better generalizability is needed. This paper proposes a novel adaptive reweighing method to eliminate the impact of the distribution shifts between training and test data on model generalizability.
  arXiv  Detail & Related papers  (2023-09-15T13:04:55Z)
• Value-Distributional Model-Based Reinforcement Learning [59.758009422067]
  Quantifying uncertainty about a policy's long-term performance is important to solve sequential decision-making tasks. We study the problem from a model-based Bayesian reinforcement learning perspective. We propose Epistemic Quantile-Regression (EQR), a model-based algorithm that learns a value distribution function.
  arXiv  Detail & Related papers  (2023-08-12T14:59:19Z)
• Scalable Personalised Item Ranking through Parametric Density Estimation [53.44830012414444]
  Learning from implicit feedback is challenging because of the difficult nature of the one-class problem. Most conventional methods use a pairwise ranking approach and negative samplers to cope with the one-class problem. We propose a learning-to-rank approach, which achieves convergence speed comparable to the pointwise counterpart.
  arXiv  Detail & Related papers  (2021-05-11T03:38:16Z)
• Scalable Approximate Inference and Some Applications [2.6541211006790983]
  In this thesis, we propose a new framework for approximate inference. Our proposed four algorithms are motivated by the recent computational progress of Stein's method. Results on simulated and real datasets indicate the statistical efficiency and wide applicability of our algorithm.
  arXiv  Detail & Related papers  (2020-03-07T04:33:27Z)
• Monotonic Cardinality Estimation of Similarity Selection: A Deep Learning Approach [22.958342743597044]
  We investigate the possibilities of utilizing deep learning for cardinality estimation of similarity selection. We propose a novel and generic method that can be applied to any data type and distance function.
  arXiv  Detail & Related papers  (2020-02-15T20:22:51Z)

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.