Explanation Quality Assessment as Ranking with Listwise Rewards

• URL: http://arxiv.org/abs/2604.24176v1
• Date: Mon, 27 Apr 2026 08:35:30 GMT
• Title: Explanation Quality Assessment as Ranking with Listwise Rewards
• Abstract summary: We reformulate explanation quality assessment as a ranking problem rather than a generation problem.<n>We construct per-instance candidate sets with graded quality levels and train reward models to discriminate among multiple candidate explanations.<n>When used as rewards in policy optimization, ranking-based scores enable stable convergence in settings where regression-based rewards fail entirely.
• Score: 21.107094956399262
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We reformulate explanation quality assessment as a ranking problem rather than a generation problem. Instead of optimizing models to produce a single "best" explanation token-by-token, we train reward models to discriminate among multiple candidate explanations and learn their relative quality. Concretely, we construct per-instance candidate sets with graded quality levels and train listwise and pairwise ranking models (ListNet, LambdaRank, RankNet) to preserve ordinal structure and avoid score compression typical of pointwise regression or binary preference objectives. We observe three findings: First, ranking losses consistently outperform regression on score separation across all domains tested. Second, the optimal ranking loss depends on data characteristics: listwise objectives excel with well-separated quality tiers, while pairwise methods are more robust to noisy natural annotations. Third, when trained on carefully curated and well-structured data, small encoder models can match models that are orders of magnitude larger, suggesting that data quality matters more than model scale. Finally, when used as rewards in policy optimization, ranking-based scores enable stable convergence in settings where regression-based rewards fail entirely. Code and data are available at: https://github.com/Tankiit/PPO_Learning_to_rank

Related papers

• CAIRO: Decoupling Order from Scale in Regression [13.755937210012883]
  We propose a framework that decouples regression into two distinct stages.<n>In the first stage, we learn a scoring function by minimizing a scale-invariant ranking loss.<n>In the second, we recover the target scale via isotonic regression.
  arXiv  Detail & Related papers  (2026-02-16T03:50:05Z)
• RewardRank: Optimizing True Learning-to-Rank Utility [28.662272762911325]
  We introduce RewardRank, a data-driven learning-to-rank framework for counterfactual utility.<n>Our results show that learning-to-rank can be reformulated as direct optimization of counterfactual utility.
  arXiv  Detail & Related papers  (2025-08-19T18:08:35Z)
• ReasonRank: Empowering Passage Ranking with Strong Reasoning Ability [83.16850534680505]
  We propose an automated reasoning-intensive training data synthesis framework.<n>A self-consistency data filtering mechanism is designed to ensure the data quality.<n>Our trained reasoning-intensive reranker textbfReasonRank achieves state-of-the-art (SOTA) performance 40.6 on the BRIGHT leaderboard.
  arXiv  Detail & Related papers  (2025-08-09T17:26:18Z)
• Rankformer: A Graph Transformer for Recommendation based on Ranking Objective [27.953113185360174]
  We propose Rankformer, a ranking-inspired recommendation model.<n>The architecture is inspired by the gradient of the ranking objective, embodying a unique (graph) transformer architecture.<n>Extensive experimental results demonstrate that Rankformer outperforms state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-03-21T07:53:06Z)
• From Pairwise to Ranking: Climbing the Ladder to Ideal Collaborative Filtering with Pseudo-Ranking [13.01752267289297]
  An ideal collaborative filtering model should learn from users' full rankings over all items to make optimal top-K recommendations.<n>Most CF models rely on pairwise loss functions to approximate full rankings, resulting in an immense performance gap.<n>We propose a pseudo-ranking paradigm (PRP) that addresses the lack of ranking information by introducing pseudo-rankings supervised by an original noise injection mechanism.
  arXiv  Detail & Related papers  (2024-12-24T05:01:16Z)
• Permutative Preference Alignment from Listwise Ranking of Human Judgments [40.23480751285947]
  We develop an end-to-end alignment algorithm by approximating NDCG with a differentiable surrogate loss.<n>We show that NDCG-based approaches improve ranking accuracy more effectively than B-T-based methods.
  arXiv  Detail & Related papers  (2024-10-06T03:49:28Z)
• Robust Outlier Rejection for 3D Registration with Variational Bayes [70.98659381852787]
  We develop a novel variational non-local network-based outlier rejection framework for robust alignment. We propose a voting-based inlier searching strategy to cluster the high-quality hypothetical inliers for transformation estimation.
  arXiv  Detail & Related papers  (2023-04-04T03:48:56Z)
• Rank-LIME: Local Model-Agnostic Feature Attribution for Learning to Rank [16.780058676633914]
  Rank-LIME is a model-agnostic, local, post-hoc linear feature attribution method for the task of learning to rank. We employ novel correlation-based perturbations, differentiable ranking loss functions and introduce new metrics to evaluate ranking based additive feature attribution models.
  arXiv  Detail & Related papers  (2022-12-24T12:14:32Z)
• Recommendation Systems with Distribution-Free Reliability Guarantees [83.80644194980042]
  We show how to return a set of items rigorously guaranteed to contain mostly good items. Our procedure endows any ranking model with rigorous finite-sample control of the false discovery rate. We evaluate our methods on the Yahoo! Learning to Rank and MSMarco datasets.
  arXiv  Detail & Related papers  (2022-07-04T17:49:25Z)
• GNNRank: Learning Global Rankings from Pairwise Comparisons via Directed Graph Neural Networks [68.61934077627085]
  We introduce GNNRank, a modeling framework compatible with any GNN capable of learning digraph embeddings. We show that our methods attain competitive and often superior performance compared with existing approaches.
  arXiv  Detail & Related papers  (2022-02-01T04:19:50Z)
• Beyond Triplet Loss: Meta Prototypical N-tuple Loss for Person Re-identification [118.72423376789062]
  We introduce a multi-class classification loss, i.e., N-tuple loss, to jointly consider multiple (N) instances for per-query optimization. With the multi-class classification incorporated, our model achieves the state-of-the-art performance on the benchmark person ReID datasets.
  arXiv  Detail & Related papers  (2020-06-08T23:34:08Z)
• Addressing Class-Imbalance Problem in Personalized Ranking [47.11372043636176]
  We propose an efficient emphunderlineVital underlineNegative underlineSampler (VINS) to alleviate the class-imbalance issue for pairwise ranking model. VINS is a bias sampler with reject probability that will tend to accept a negative candidate with a larger degree weight than the given positive item.
  arXiv  Detail & Related papers  (2020-05-19T08:11:26Z)
• The World is Not Binary: Learning to Rank with Grayscale Data for Dialogue Response Selection [55.390442067381755]
  We show that grayscale data can be automatically constructed without human effort. Our method employs off-the-shelf response retrieval models and response generation models as automatic grayscale data generators. Experiments on three benchmark datasets and four state-of-the-art matching models show that the proposed approach brings significant and consistent performance improvements.
  arXiv  Detail & Related papers  (2020-04-06T06:34:54Z)

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.