activity
20242026
most citedTowards Lossless Token Pruning in Late-Interaction Retrieval Models

2 citations · 3 across the 9 of their papers we have counts for

collaborators

9 papers

math.PR2026

Sharp asymptotics for the tree-completion time in cylindrical Hastings--Levitov

Xiao-Ming Fu, Tianyang Sun, Yuxuan Zong

Let be the cylindrical Hastings--Levitov aggregation process with parameter on a cylinder of width with particles of fixed size , and let be…

cs.IR2026

STORM: Stepwise Token Optimization with Reward-Guided Beam Search

Arthur Satouf, Giulio D'Erasmo, Yuxuan Zong +3

Modern retrieval increasingly relies on dense and learned-sparse neural models that are effective but require encoding the entire corpus into a specialized index, rebuilt whenever…

cs.IR2026

From Tokens to Concepts: Leveraging SAE for SPLADE

Yuxuan Zong, Mathias Vast, Basile Van Cooten +2

Learned Sparse IR models, such as SPLADE, offer an excellent efficiency-effectiveness tradeoff. However, they rely on the underlying backbone vocabulary, which might hinder perform…

cs.IR2026

A Voronoi Cell Formulation for Principled Token Pruning in Late-Interaction Retrieval Models

Yash Kankanampati, Yuxuan Zong, Nadi Tomeh +2

Late-interaction models such as ColBERT offer competitive performance across various retrieval tasks but require storing a dense embedding for each document token, leading to a sub…

cs.IR2025

QueStER: Query Specification for Generative keyword-based Retrieval

Arthur Satouf, Yuxuan Zong, Habiboulaye Amadou-Boubacar +2

Generative retrieval (GR) differs from the traditional index-then-retrieve pipeline by storing relevance in model parameters and generating retrieval cues directly from the query,…

math.PR2025

One-arm domination time in Cylindrical Hastings-Levitov

Guanyi Chen, Eviatar B. Procaccia, Yuxuan Zong

The cylindrical Hastings-Levitov admits a single infinite connected tree (arm). For a cylinder of width and particles of size , {we consider the first time $\upsilon_{N…