Abstract:Traditional Learning-To-Rank (LETOR) approaches, including pairwise methods like RankNet and LambdaMART, often fall short by solely focusing on pairwise comparisons, leading to sub-optimal global rankings. Conversely, deep learning based listwise methods, while aiming to optimise entire lists, require complex tuning and yield only marginal improvements over robust pairwise models. To overcome these limitations, we introduce Travelling Salesman Problem Rank (TSPRank), a hybrid pairwise-listwise ranking method. TSPRank reframes the ranking problem as a Travelling Salesman Problem (TSP), a well-known combinatorial optimisation challenge that has been extensively studied for its numerous solution algorithms and applications. This approach enables the modelling of pairwise relationships and leverages combinatorial optimisation to determine the listwise ranking. This approach can be directly integrated as an additional component into embeddings generated by existing backbone models to enhance ranking performance. Our extensive experiments across three backbone models on diverse tasks, including stock ranking, information retrieval, and historical events ordering, demonstrate that TSPRank significantly outperforms both pure pairwise and listwise methods. Our qualitative analysis reveals that TSPRank's main advantage over existing methods is its ability to harness global information better while ranking. TSPRank's robustness and superior performance across different domains highlight its potential as a versatile and effective LETOR solution. The code and preprocessed data are available at <a class="link-external link-https" href="https://github.com/waylonli/TSPRank-KDD2025" rel="external noopener nofollow">this https URL</a>.

FastLexRank: Efficient Lexical Ranking for Structuring Social Media Posts

Compact Indexing and Judicious Searching for Billion-Scale Microblog Retrieval.

LexRank: Graph-based Lexical Centrality as Salience in Text Summarization

Simple Attention-Based Representation Learning for Ranking Short Social Media Posts

Explicit and Implicit Semantic Ranking Framework

A Novel Approach For Generating Personalized Mention List On Micro-Blogging System

LiRank: Industrial Large Scale Ranking Models at LinkedIn

Messages Ranking in Social Network.

LexFindR: A fast, simple, and extensible R package for finding similar words in a lexicon

Learning to Rank Microblog Posts for Real-Time Ad-Hoc Search.

A Social Search Model for Large Scale Social Networks

Improving Link Prediction In Social Networks By User Comments And Sentiment Lexicon

Real Time Personalized Search on Social Networks

TSPRank: Bridging Pairwise and Listwise Methods with a Bilinear Travelling Salesman Model

SPLADE: Sparse Lexical and Expansion Model for First Stage Ranking

Leveraging Tweet Ranking In An Optimization Framework For Tweet Timeline Generation

Leveraging Tweet Ranking in an Optimization Frameworkfor Tweet Timeline Generation.

Efficient Neural Ranking using Forward Indexes and Lightweight Encoders

RATE: Overcoming Noise and Sparsity of Textual Features in Real-Time Location Estimation

An Early FIRST Reproduction and Improvements to Single-Token Decoding for Fast Listwise Reranking

Adaptive Content-Aware Influence Maximization via Online Learning to Rank