Beyond Seen Data: Improving KBQA Generalization Through Schema-Guided Logical Form Generation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915487904956416 |
|---|---|
| author | Gao, Shengxiang Lau, Jey Han Qi, Jianzhong |
| author_facet | Gao, Shengxiang Lau, Jey Han Qi, Jianzhong |
| contents | Knowledge base question answering (KBQA) aims to answer user questions in natural language using rich human knowledge stored in large KBs. As current KBQA methods struggle with unseen knowledge base elements at test time,we introduce SG-KBQA: a novel model that injects schema contexts into entity retrieval and logical form generation to tackle this issue. It uses the richer semantics and awareness of the knowledge base structure provided by schema contexts to enhance generalizability. We show that SG-KBQA achieves strong generalizability, outperforming state-of-the-art models on two commonly used benchmark datasets across a variety of test settings. Our source code is available at https://github.com/gaosx2000/SG_KBQA. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_12737 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Beyond Seen Data: Improving KBQA Generalization Through Schema-Guided Logical Form Generation Gao, Shengxiang Lau, Jey Han Qi, Jianzhong Computation and Language Artificial Intelligence Knowledge base question answering (KBQA) aims to answer user questions in natural language using rich human knowledge stored in large KBs. As current KBQA methods struggle with unseen knowledge base elements at test time,we introduce SG-KBQA: a novel model that injects schema contexts into entity retrieval and logical form generation to tackle this issue. It uses the richer semantics and awareness of the knowledge base structure provided by schema contexts to enhance generalizability. We show that SG-KBQA achieves strong generalizability, outperforming state-of-the-art models on two commonly used benchmark datasets across a variety of test settings. Our source code is available at https://github.com/gaosx2000/SG_KBQA. |
| title | Beyond Seen Data: Improving KBQA Generalization Through Schema-Guided Logical Form Generation |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2502.12737 |