Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2512.19323 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909973509832704 |
|---|---|
| author | Lopez-Rubio, Ezequiel Decena-Gimenez, Macoris Luque-Baena, Rafael Marcos |
| author_facet | Lopez-Rubio, Ezequiel Decena-Gimenez, Macoris Luque-Baena, Rafael Marcos |
| contents | A key module in neural transformer-based deep architectures is positional encoding. This module enables a suitable way to encode positional information as input for transformer neural layers. This success has been rooted in the use of sinusoidal functions of various frequencies, in order to capture recurrent patterns of differing typical periods. In this work, an alternative set of periodic functions is proposed for positional encoding. These functions preserve some key properties of sinusoidal ones, while they depart from them in fundamental ways. Some tentative experiments are reported, where the original sinusoidal version is substantially outperformed. This strongly suggests that the alternative functions may have a wider use in other transformer architectures. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_19323 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Alternative positional encoding functions for neural transformers Lopez-Rubio, Ezequiel Decena-Gimenez, Macoris Luque-Baena, Rafael Marcos Machine Learning Artificial Intelligence 68T07 I.2 A key module in neural transformer-based deep architectures is positional encoding. This module enables a suitable way to encode positional information as input for transformer neural layers. This success has been rooted in the use of sinusoidal functions of various frequencies, in order to capture recurrent patterns of differing typical periods. In this work, an alternative set of periodic functions is proposed for positional encoding. These functions preserve some key properties of sinusoidal ones, while they depart from them in fundamental ways. Some tentative experiments are reported, where the original sinusoidal version is substantially outperformed. This strongly suggests that the alternative functions may have a wider use in other transformer architectures. |
| title | Alternative positional encoding functions for neural transformers |
| topic | Machine Learning Artificial Intelligence 68T07 I.2 |
| url | https://arxiv.org/abs/2512.19323 |