Attention-guided Evidence Grounding for Spoken Question Answering
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917350946635776 |
|---|---|
| author | Yang, Ke Chen, Bolin Li, Yuejie Hua, Yueying Nie, Jianhao He, Yueping Li, Bowen Mao, Chengjun |
| author_facet | Yang, Ke Chen, Bolin Li, Yuejie Hua, Yueying Nie, Jianhao He, Yueping Li, Bowen Mao, Chengjun |
| contents | Spoken Question Answering (Spoken QA) presents a challenging cross-modal problem: effectively aligning acoustic queries with textual knowledge while avoiding the latency and error propagation inherent in cascaded ASR-based systems. In this paper, we introduce Attention-guided Evidence Grounding (AEG), a novel end-to-end framework that leverages the internal cross-modal attention of Speech Large Language Models (SpeechLLMs) to explicitly locate and ground key evidence in the model's latent space. To address the diffuse attention distribution in pre-trained models, we propose Learning to Focus on Evidence (LFE), a supervised fine-tuning paradigm that calibrates the model's attention mechanism to distinguish query-relevant segments from irrelevant context. Experiments on SQuAD, HotpotQA, and MuSiQue demonstrate that AEG reduces hallucinations and achieves strong efficiency gains, outperforming large-scale cascaded baselines (Whisper-Large-v3 + Reranker) while reducing inference latency by approximately 62%. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_16292 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Attention-guided Evidence Grounding for Spoken Question Answering Yang, Ke Chen, Bolin Li, Yuejie Hua, Yueying Nie, Jianhao He, Yueping Li, Bowen Mao, Chengjun Computation and Language Artificial Intelligence Spoken Question Answering (Spoken QA) presents a challenging cross-modal problem: effectively aligning acoustic queries with textual knowledge while avoiding the latency and error propagation inherent in cascaded ASR-based systems. In this paper, we introduce Attention-guided Evidence Grounding (AEG), a novel end-to-end framework that leverages the internal cross-modal attention of Speech Large Language Models (SpeechLLMs) to explicitly locate and ground key evidence in the model's latent space. To address the diffuse attention distribution in pre-trained models, we propose Learning to Focus on Evidence (LFE), a supervised fine-tuning paradigm that calibrates the model's attention mechanism to distinguish query-relevant segments from irrelevant context. Experiments on SQuAD, HotpotQA, and MuSiQue demonstrate that AEG reduces hallucinations and achieves strong efficiency gains, outperforming large-scale cascaded baselines (Whisper-Large-v3 + Reranker) while reducing inference latency by approximately 62%. |
| title | Attention-guided Evidence Grounding for Spoken Question Answering |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2603.16292 |