Incremental Sequence Classification with Temporal Consistency
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915573038841856 |
|---|---|
| author | Maystre, Lucas Barello, Gabriel Berariu, Tudor Cambray, Aleix Dolga, Rares Gonzalez, Alvaro Ortega Nica, Andrei Barber, David |
| author_facet | Maystre, Lucas Barello, Gabriel Berariu, Tudor Cambray, Aleix Dolga, Rares Gonzalez, Alvaro Ortega Nica, Andrei Barber, David |
| contents | We address the problem of incremental sequence classification, where predictions are updated as new elements in the sequence are revealed. Drawing on temporal-difference learning from reinforcement learning, we identify a temporal-consistency condition that successive predictions should satisfy. We leverage this condition to develop a novel loss function for training incremental sequence classifiers. Through a concrete example, we demonstrate that optimizing this loss can offer substantial gains in data efficiency. We apply our method to text classification tasks and show that it improves predictive accuracy over competing approaches on several benchmark datasets. We further evaluate our approach on the task of verifying large language model generations for correctness in grade-school math problems. Our results show that models trained with our method are better able to distinguish promising generations from unpromising ones after observing only a few tokens. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_16548 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Incremental Sequence Classification with Temporal Consistency Maystre, Lucas Barello, Gabriel Berariu, Tudor Cambray, Aleix Dolga, Rares Gonzalez, Alvaro Ortega Nica, Andrei Barber, David Machine Learning We address the problem of incremental sequence classification, where predictions are updated as new elements in the sequence are revealed. Drawing on temporal-difference learning from reinforcement learning, we identify a temporal-consistency condition that successive predictions should satisfy. We leverage this condition to develop a novel loss function for training incremental sequence classifiers. Through a concrete example, we demonstrate that optimizing this loss can offer substantial gains in data efficiency. We apply our method to text classification tasks and show that it improves predictive accuracy over competing approaches on several benchmark datasets. We further evaluate our approach on the task of verifying large language model generations for correctness in grade-school math problems. Our results show that models trained with our method are better able to distinguish promising generations from unpromising ones after observing only a few tokens. |
| title | Incremental Sequence Classification with Temporal Consistency |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2505.16548 |