recommendation and ranking

Estimating nonparametric ranking-based discrete choice models from transaction data—accounting for both single- and multiple-purchase behaviors—requires handling an exponential number of consumer types, leading to prohibitive computational complexity. This work proposes the first dynamic programming–based column generation framework that efficiently enumerates relevant consumer types in such models. The approach features a novel subproblem that generalizes the linear ordering problem and incorporates acceleration techniques to enhance optimization efficiency. The method accommodates various model extensions and demonstrates substantial improvements over existing approaches on both synthetic and real-world datasets, achieving significantly faster computation while maintaining high estimation accuracy. Furthermore, it exhibits strong performance in downstream assortment optimization tasks.

This work addresses the challenge of effectively integrating user–item and item–item collaborative filtering to enhance Top-N recommendation performance while maintaining computational efficiency. The authors propose a weighted similarity ensemble method based on shared embeddings, which, for the first time, unifies both recommendation pathways within a single framework. By sharing user and item embeddings across strategies, the approach simplifies model architecture and eliminates the need for separate hyperparameter tuning for each pathway, thereby substantially reducing deployment complexity. Experimental results demonstrate that the proposed method achieves competitive recommendation accuracy across multiple datasets and exhibits robust performance in scenarios favoring different collaborative filtering paradigms.

This paper studies preference learning from choice feedback over a dynamic item set, aiming to identify either the optimal item or the complete preference ranking with minimal samples and high confidence. We propose two algorithms—Nested Elimination (NE) and Nested Partitioning (NP)—that provide the first non-asymptotic, instance-dependent sample complexity bounds for arbitrary strict preference structures; NE achieves asymptotic optimality in the information-theoretic worst case, while NP attains constant-factor optimality. Our approach integrates multi-dimensional random walk modeling, divide-and-conquer strategies, and an information-theoretic analysis framework. Rigorous theoretical analysis is complemented by extensive experiments on both synthetic and real-world datasets, demonstrating superior efficiency and robustness. The core contribution is the establishment of the first algorithmic framework for dynamic preference learning that simultaneously ensures practical applicability and theoretical optimality.

Existing performance ranking methods in entity evaluation struggle to simultaneously satisfy application-specific preferences and theoretical rigor. Method: This paper establishes the first axiomatic, verifiable general theory framework for performance ranking. Grounded in probability theory and order theory, it formally defines core concepts—including performance objects, satisfaction, and importance—and introduces a performance order satisfying axioms such as ranking consistency, along with constructive procedures for deriving such orders. It further proposes a novel parameterized family of universal ranking scores that unifies classical metrics (e.g., accuracy, recall, F1-score) and rigorously proves that several widely used metrics—including precision—violate the ranking consistency axiom. Contribution/Results: The framework provides the first mathematically rigorous yet practically flexible foundation for performance evaluation in computer vision and machine learning, explicitly characterizing the validity boundaries and intrinsic limitations of reliable ranking metrics.

This paper investigates whether weighted matrix factorization (WMF) improves recommendation performance in implicit-feedback settings. Through systematic analysis of the coupling effects among weighting schemes, model capacity, and regularization, we find that unweighted training achieves performance comparable to—or even surpassing—that of state-of-the-art weighted methods under high-capacity models, challenging the conventional assumption that weighting universally enhances performance. To address the computational difficulty of exact optimization for classical weighted objectives (e.g., WALS), we propose a novel, efficient, and provably exact optimization algorithm—the first to minimize such non-convex, non-smooth weighted objectives without approximation. Extensive experiments across diverse MF architectures and weighting strategies on multiple benchmark datasets confirm our findings: weighting yields gains only under low-capacity models or strong regularization, whereas simplified unweighted training is more robust and efficient for large-scale models. Our core contributions are (i) identifying precise boundary conditions under which weighting is beneficial, (ii) providing theoretical justification, and (iii) delivering a practical, exact optimization tool.

Existing generative retrieval methods in recommendation systems rely solely on next-token prediction, which struggles to capture the hierarchical structure of user preferences and the deep interactions between items and behavioral sequences. This work proposes RankGR, the first approach to integrate listwise direct preference optimization (Listwise DPO) into generative retrieval. RankGR employs a two-stage collaborative mechanism: an Initial Assessment Phase (IAP) for coarse candidate filtering, followed by a Refinement Scoring Phase (RSP) that leverages a lightweight scoring module for fine-grained ranking. This design balances modeling depth with inference efficiency, enabling high-concurrency real-time deployment. Experiments demonstrate that RankGR significantly improves offline metrics across multiple academic and industrial datasets and delivers substantial online gains in Taobao’s “Guess You Like” scenario, stably supporting nearly 10,000 queries per second.

This work addresses the challenges of personalized ranking when users lack knowledge of data attributes or struggle to articulate their preferences explicitly. Existing approaches are limited by reliance on a single candidate item selection strategy, which constrains flexibility and user control. To overcome this, the authors propose a visual analytics framework that integrates model-driven active learning with human-driven item selection, establishing—for the first time—a unified interactive item selection space. This space supports six complementary strategies for expressing list-level preferences and enables iterative learning to produce interpretable rankings. A formative user study (N=10) demonstrates the approach’s effectiveness and reveals trade-offs among accuracy, diversity, novelty, transparency, perceived control, and user satisfaction across different selection strategies.

This work addresses the misalignment in existing financial forecasting models, which are typically evaluated using Rank IC yet trained with regression or ranking losses inconsistent with this metric. To bridge this gap, we propose LambdaRankIC, the first method to directly optimize Rank IC in an end-to-end differentiable learning-to-rank framework. By deriving a closed-form expression for the lambda gradients induced by pairwise rank swaps, we efficiently maximize an upper bound of Rank IC within the LambdaRank paradigm and implement a custom objective function based on XGBoost. Extensive experiments demonstrate that LambdaRankIC significantly outperforms current regression- and NDCG-oriented approaches across both synthetic and real-world financial datasets, achieving state-of-the-art performance in terms of Rank IC, ICIR, monthly returns, and Sharpe ratio.

Recommender systems are tasked to infer users'evolving preferences and rank items aligned with their intents, which calls for in-depth reasoning beyond pattern-based scoring. Recent efforts start to leverage large language models (LLMs) for recommendation, but how to effectively optimize the model for improved recommendation utility is still under explored. In this work, we propose Reasoning to Rank, an end-to-end training framework that internalizes recommendation utility optimization into the learning of step-by-step reasoning in LLMs. To avoid position bias in LLM reasoning and enable direct optimization of the reasoning process, our framework performs reasoning at the user-item level and employs reinforcement learning for end-to-end training of the LLM. Experiments on three Amazon datasets and a large-scale industrial dataset showed consistent gains over strong conventional and LLM-based solutions. Extensive in-depth analyses validate the necessity of the key components in the proposed framework and shed lights on the future developments of this line of work.