Token Pruning for Improving Graph-Generating State Space Model Performance

Read full text Summarise with VIRAT
Authors: Monish Beegamudre, Jack Zheng, Margaret Capetz
Cited by: 0
References: 7

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

State Space Models (SSMs) have recently emerged as efficient alternatives to Transformers for sequence modeling, yet extending them to two-dimensional tasks remains challenging.The Graph-Generating State Space Model (GG-SSM) addresses this challenge by constructing an adaptive graph, achieving competitive performance on vision benchmarks.However, state propagation over the resulting graph introduces substantial inference overhead, limiting scalability to high-resolution inputs.In this work, we introduce a leaf-guided computation pruning strategy that accelerates GG-SSM inference without modifying the underlying graph topology.Rather than removing nodes or edges, our approach selectively scales or bypasses secondary refinement computations associated with high-dissimilarity leaf nodes, while preserving the low-weight MST backbone.Experiments on multiple long-term time series forecasting benchmarks demonstrate consistent throughput improvements with controlled accuracy degradation across a range of pruning ratios.These results indicate that structure-aware computation pruning is an effective mechanism for improving the scalability of graph-based state space models.Our source code is publicly available at https://github.com/MonishB123/ token-pruning-for-ggssm

Fields of Study & Subjects

Advanced Neural Network Applications Time Series Analysis and Forecasting Visual Attention and Saliency Detection

Publication details

Journal / source
Conference of the European Chapter of the Association for Computational Linguistics
Pages
476–482
Publication date
1 January 2026
Language
EN