Page 39 - 《软件学报》2026年第5期

1918 软件学报 2026 年第 37 卷第 5 期

170: 103224. [doi: 10.1016/j.specom.2025.103224]
[19] Yang M, Kanda N, Wang XF, Chen JK, Wang PD, Xue J, Li JY, Yoshioka T. DiariST: Streaming speech translation with speaker
diarization. In: Proc. of the 2024 IEEE Int’l Conf. on Acoustics, Speech and Signal Processing (ICASSP). Seoul: IEEE, 2024.
10866–10870. [doi: 10.1109/ICASSP48485.2024.10446050]
[20] Piñeiro-Martín A, García-Mateo C, Docío-Fernández L, López-Pérez MDC, Rehm G. Weighted cross-entropy for low-resource languages
in multilingual speech recognition. In: Proc. of the 25th Annual Conf. of the Int’l Speech Communication Association. Kos, 2024.
1235–1239. [doi: 10.21437/Interspeech.2024-734]
[21] Khare A, Han EJ, Yang YG, Stolcke A. ASR-aware end-to-end neural diarization. In: Proc. of the 2022 IEEE Int’l Conf. on Acoustics,
Speech and Signal Processing (ICASSP). Singapore: IEEE, 2022. 8092–8096. [doi: 10.1109/ICASSP43922.2022.9746964]
[22] Kumar S, Madikeri S, Nigmatulina I, Villatoro-Tello E, Motlicek P, Pandia K, Dubagunta SP, Ganapathiraju A. Multitask speech
recognition and speaker change detection for unknown number of speakers. In: Proc. of the 2024 IEEE Int’l Conf. on Acoustics, Speech
and Signal Processing (ICASSP). Seoul: IEEE, 2024. 12592–12596. [doi: 10.1109/ICASSP48485.2024.10446130]
[23] Li G, Liu JZ, Dinkel H, Niu YD, Zhang JB, Luan J. Reinforcement learning outperforms supervised fine-tuning: A case study on audio
question answering. arXiv:2503.11197, 2025.
[24] Nagpal C, Venugopalan S, Tobin J, Ladewig M, Heller K, Tomanek K. Speech recognition with LLMs adapted to disordered speech
using reinforcement learning. In: Proc. of the 2025 IEEE Int’l Conf. on Acoustics, Speech and Signal Processing (ICASSP). Hyderabad:
IEEE, 2025. 1–5. [doi: 10.1109/ICASSP49660.2025.10888006]
[25] Rafailov R, Sharma A, Mitchell E, Ermon S, Manning CD, Finn C. Direct preference optimization: Your language model is secretly a
reward model. In: Proc. of the 37th Int’l Conf. on Neural Information Processing Systems. New Orleans: Curran Associates Inc., 2023.
53728–53741.
[26] Schulman J, Wolski F, Dhariwal P, Radford A, Klimov O. Proximal policy optimization algorithms. arXiv:1707.06347, 2017.
[27] Wang SH, Zhang SY, Zhang J, Hu RY, Li XY, Zhang TW, Li JW, Wu F, Wang GY, Hovy E. Reinforcement learning enhanced LLMs: A
survey. arXiv:2412.10400, 2025.
[28] Park T, Medennikov I, Dhawan K, Wang WQ, Huang H, Koluguri NR, Puvvada KC, Balam J, Ginsburg B. Sortformer: A novel approach
for permutation-resolved speaker supervision in speech-to-text systems. arXiv:2409.06656, 2025.
[29] Gao ZF, Zhang SL, McLoughlin I, Yan ZJ. Paraformer: Fast and accurate parallel Transformer for non-autoregressive end-to-end speech
recognition. arXiv:2206.08317, 2023.
[30] Mao YX, Wang Q, Chen C, Qu Y, Ji XY. Offline reinforcement learning with OOD state correction and OOD action suppression. In:
Proc. of the 38th Int’l Conf. on Neural Information Processing Systems. Vancouver: Curran Associates Inc., 2025. 93568–93601.
[31] Haider T, Roscher K, da Roza FS, Günnemann S. Out-of-distribution detection for reinforcement learning agents with probabilistic
dynamics models. In: Proc. of the 2023 Int’l Conf. on Autonomous Agents and Multiagent Systems. London: Int’l Foundation for
Autonomous Agents and Multiagent Systems, 2023. 851–859.
[32] Gulati A, Qin J, Chiu CC, Parmar N, Zhang Y, Yu JH, Han W, Wang SB, Zhang ZD, Wu YH, Pang RM. Conformer: Convolution-
augmented Transformer for speech recognition. In: Proc. of the 21st Annual Conf. of the Int’l Speech Communication Association.
Shanghai, 2020. 5036–5040.
[33] Zhong SS, Gao SH, Huang ZZ, Wen WS, Zitnik M, Zhou P. MoExtend: Tuning new experts for modality and task extension. In: Proc. of
the 62nd Annual Meeting of the Association for Computational Linguistics (Vol. 4: Student Research Workshop). Bangkok: ACL, 2024.
494–505.

附中文参考文献
[12] 朱必松, 毛启容, 高利剑, 沈雅馨. 基于时间分段和重组聚类的说话人日志方法. 计算机应用研究, 2024, 41(9): 2649–2654. [doi:
10.19734/j.issn.1001-3695.2024.01.0017]
[13] 毛青青, 贾洪杰, 朱必松. 面向说话人日志的多原型驱动图神经网络方法. 计算机应用研究, 2025, 42(6): 1778–1783. [doi: 10.19734/j.
issn.1001-3695.2024.11.0458]

作者简介
韦舒羽, 博士生, 主要研究领域为多模态大模型.
丘德来, 硕士, 主要研究领域为多模态大模型, 自然语言处理, 智能问答.
刘升平, 博士, CCF 专业会员, 主要研究领域为大语言模型, 自然语言处理, 知识图谱.
桑基韬, 博士, 教授, 博士生导师, CCF 高级会员, 主要研究领域为多模态智能, 可信与对齐, 推理模型, AI Agent.