Parallel Algorithms Align with Neural Execution
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2023
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909059910729728 |
|---|---|
| author | Engelmayer, Valerie Georgiev, Dobrik Veličković, Petar |
| author_facet | Engelmayer, Valerie Georgiev, Dobrik Veličković, Petar |
| contents | Neural algorithmic reasoners are parallel processors. Teaching them sequential algorithms contradicts this nature, rendering a significant share of their computations redundant. Parallel algorithms however may exploit their full computational power, therefore requiring fewer layers to be executed. This drastically reduces training times, as we observe when comparing parallel implementations of searching, sorting and finding strongly connected components to their sequential counterparts on the CLRS framework. Additionally, parallel versions achieve (often strongly) superior predictive performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2307_04049 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Parallel Algorithms Align with Neural Execution Engelmayer, Valerie Georgiev, Dobrik Veličković, Petar Machine Learning Neural algorithmic reasoners are parallel processors. Teaching them sequential algorithms contradicts this nature, rendering a significant share of their computations redundant. Parallel algorithms however may exploit their full computational power, therefore requiring fewer layers to be executed. This drastically reduces training times, as we observe when comparing parallel implementations of searching, sorting and finding strongly connected components to their sequential counterparts on the CLRS framework. Additionally, parallel versions achieve (often strongly) superior predictive performance. |
| title | Parallel Algorithms Align with Neural Execution |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2307.04049 |