AIRE-Prune: Asymptotic Impulse-Response Energy for State Pruning in State Space Models
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910007324311552 |
|---|---|
| author | Padhy, Apurba Prasad Camacho, Fernando Mukhopadhyay, Saibal |
| author_facet | Padhy, Apurba Prasad Camacho, Fernando Mukhopadhyay, Saibal |
| contents | State space models (SSMs) often sacrifice capacity, search space, or stability to offset the memory and compute costs of large state dimensions. We introduce a structured post-training pruning method for SSMs -- AIRE-Prune (Asymptotic Impulse-Response Energy for State PRUN(E)) -- that reduces each layer's state dimension by directly minimizing long-run output-energy distortion. AIRE-Prune assigns every state a closed-form asymptotic impulse-response energy-based score, i.e., the total impulse-response energy it contributes over an infinite horizon (time), and normalizes these scores layer-wise to enable global cross-layer comparison and selection. This extends modal truncation from single systems to deep stacks and aligns pruning with asymptotic response energy rather than worst-case gain. Across diverse sequence benchmarks, AIRE-Prune reveals substantial redundancy in SISO and MIMO SSMs with average pruning of 60.8%, with average accuracy drop of 0.29% without retraining, while significantly lowering compute. Code: https://github.com/falcon-arrow/AIRE-Prune. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_00534 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | AIRE-Prune: Asymptotic Impulse-Response Energy for State Pruning in State Space Models Padhy, Apurba Prasad Camacho, Fernando Mukhopadhyay, Saibal Machine Learning Systems and Control State space models (SSMs) often sacrifice capacity, search space, or stability to offset the memory and compute costs of large state dimensions. We introduce a structured post-training pruning method for SSMs -- AIRE-Prune (Asymptotic Impulse-Response Energy for State PRUN(E)) -- that reduces each layer's state dimension by directly minimizing long-run output-energy distortion. AIRE-Prune assigns every state a closed-form asymptotic impulse-response energy-based score, i.e., the total impulse-response energy it contributes over an infinite horizon (time), and normalizes these scores layer-wise to enable global cross-layer comparison and selection. This extends modal truncation from single systems to deep stacks and aligns pruning with asymptotic response energy rather than worst-case gain. Across diverse sequence benchmarks, AIRE-Prune reveals substantial redundancy in SISO and MIMO SSMs with average pruning of 60.8%, with average accuracy drop of 0.29% without retraining, while significantly lowering compute. Code: https://github.com/falcon-arrow/AIRE-Prune. |
| title | AIRE-Prune: Asymptotic Impulse-Response Energy for State Pruning in State Space Models |
| topic | Machine Learning Systems and Control |
| url | https://arxiv.org/abs/2602.00534 |