Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2603.05573 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917316425416704 |
|---|---|
| author | Heo, Gyuryang Ngotiaoco, Timothy Irie, Kazuki Gershman, Samuel J. Sabatini, Bernardo |
| author_facet | Heo, Gyuryang Ngotiaoco, Timothy Irie, Kazuki Gershman, Samuel J. Sabatini, Bernardo |
| contents | Scalable sequence models, such as Transformer variants and structured state-space models, often trade expressivity power for sequence-level parallelism, which enables efficient training. Here we examine the bounds on error and how error scales when models operate outside of their expressivity regimes using a Lie-algebraic control perspective. Our theory formulates a correspondence between the depth of a sequence model and the tower of Lie algebra extensions. Echoing recent theoretical studies, we characterize the Lie-algebraic class of constant-depth sequence models and their corresponding expressivity bounds. Furthermore, we analytically derive an approximation error bound and show that error diminishes exponentially as the depth increases, consistent with the strong empirical performance of these models. We validate our theoretical predictions using experiments on symbolic word and continuous-valued state-tracking problems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_05573 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Why Depth Matters in Parallelizable Sequence Models: A Lie Algebraic View Heo, Gyuryang Ngotiaoco, Timothy Irie, Kazuki Gershman, Samuel J. Sabatini, Bernardo Machine Learning Scalable sequence models, such as Transformer variants and structured state-space models, often trade expressivity power for sequence-level parallelism, which enables efficient training. Here we examine the bounds on error and how error scales when models operate outside of their expressivity regimes using a Lie-algebraic control perspective. Our theory formulates a correspondence between the depth of a sequence model and the tower of Lie algebra extensions. Echoing recent theoretical studies, we characterize the Lie-algebraic class of constant-depth sequence models and their corresponding expressivity bounds. Furthermore, we analytically derive an approximation error bound and show that error diminishes exponentially as the depth increases, consistent with the strong empirical performance of these models. We validate our theoretical predictions using experiments on symbolic word and continuous-valued state-tracking problems. |
| title | Why Depth Matters in Parallelizable Sequence Models: A Lie Algebraic View |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2603.05573 |