在 Jetson Nano 上运行的质量感知自校正语音翻译系统

View PDF HTML (experimental)

Abstract:We present a fully offline speech-to-speech translation pipeline that runs on a Jetson Nano (4 GB) and corrects its own weak translations without retraining. A Whisper-tiny ASR feeds an Opus-MT translator; multilingual BERT cosine similarity acts as a Quality Estimation (QE) gate, triggering a secondary-pass when confidence falls below a pre-defined threshold $\tau$. We compare three correction methods: QE reranking (M1), Minimum Bayes-Risk decoding (M2), and constrained beam search (M3). On 1,012 FLORES-200 sentences (English-Spanish), M2 at $\tau=0.90$ produces statistically significant improvements over greedy decoding on BLEU (+0.67, p<0.001), ChrF (+0.51, p<0.001), and COMET (+0.0020 at N=3, p=0.002); M1 yields no significant gains, and M3 is significantly worse than baseline (p>0.99). Our central finding is that QE functions effectively as a gate but poorly as a ranker: removing the QE model from candidate selection (M1$\to$M2) does not hurt quality and frees 680 MB from the critical path. Using a gain-to-edit ratio adapted from the post-editing-effort literature, we further show that smaller candidate pools (N=3) yield more surgical corrections with better semantic adequacy, while larger pools (N=10) maximise lexical reward. We release the system and demonstrate live translation across six language pairs.

Comments:	v2: added code repository link; corrected figures, table labels and minor text. 7 pages, 2 figures, 4 tables. Full paper submitted to the Convergence 2026 proceedings; poster presented at Convergence 2026, University of Surrey. Code: this https URL
Subjects:	Computation and Language (cs.CL)
Cite as:	arXiv:2610.07545 [cs.CL]
	(or arXiv:2610.07545v2 [cs.CL] for this version)
	https://doi.org/10.48550/arXiv.2610.07545 arXiv-issued DOI via DataCite

Submission history

From: Zubair Ajmal Farooq [view email]
[v1] Tue, 6 Oct 2026 00:21:08 UTC (388 KB)
[v2] Wed, 14 Oct 2026 08:03:16 UTC (134 KB)