| Proj No. | B3088-251 |
| Title | Streaming Speech-to-Speech Conversational Agent with Semantic Continuity |
| Summary | This project aims to develop a real-time speech-to-speech conversational agent designed for role-play training in complex, multi-turn interactions. Unlike traditional models that respond utterance by utterance, the system maintains semantic continuity across conversation turns while responding in real-time. It integrates speech recognition, context tracking, and structured dialogue management to enable realistic, natural interactions. The goal of the prototype will support training scenarios such as consultations, with capabilities for post-hoc analysis and user feedback, contributing to advancements in real-time spoken dialogue systems and practical deployment of large language models. |
| Supervisor | A/P Lin Zhiping (Loc:S2 > S2 B2A > S2 B2A 14, Ext: +65 67906857) |
| Co-Supervisor | - |
| RI Co-Supervisor | - |
| Lab | Machine Learning and Data Analytics Lab (Loc: S2.1-B4-01) |
| Single/Group: | Single |
| Area: | Intelligent Systems and Control Engineering |
| ISP/RI/SMP/SCP?: | ISP: Dr Mak Lee Onn Deputy Director Q3 HTX mak_lee_onn@htx.gov.sg |