Hierarchical Policy Optimization for Simultaneous Translation of Unbounded Speech

• URL: http://arxiv.org/abs/2604.21045v1
• Date: Wed, 22 Apr 2026 19:43:51 GMT
• Title: Hierarchical Policy Optimization for Simultaneous Translation of Unbounded Speech
• Abstract summary: Simultaneous speech translation (SST) generates translations while receiving partial speech input.<n>Recent advances show that large language models (LLMs) can substantially improve SST quality, but at the cost of high computational overhead.<n>We propose a Hierarchical Policy Optimization (HPO) approach that trains post-train models on imperfect SFT data.<n> Experiments on English to Chinese/German/Japanese demonstrate improvements of over +7 COMET score and +1.25 MetricX score at a latency of 1.5 seconds.
• Score: 50.45710815530982
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Simultaneous speech translation (SST) generates translations while receiving partial speech input. Recent advances show that large language models (LLMs) can substantially improve SST quality, but at the cost of high computational overhead. To reduce this cost, prior work reformulates SST as a multi-turn dialogue task, enabling full reuse of the LLM's key-value (KV) cache and eliminating redundant feature recomputation. However, this approach relies on supervised fine-tuning (SFT) data in dialogue form, for which few human annotations exist, and existing synthesis methods cannot guarantee data quality. In this work, we propose a Hierarchical Policy Optimization (HPO) approach that post-train models trained on imperfect SFT data. We introduce a hierarchical reward that balances translation quality and latency objectives. Experiments on English to Chinese/German/Japanese demonstrate improvements of over +7 COMET score and +1.25 MetricX score at a latency of 1.5 seconds. Comprehensive ablation studies further validate the effectiveness of different quality rewards, hierarchical reward formulations, and segmentation strategies. Code can be found here https://github.com/owaski/HPO

Related papers

• Adapting Text LLMs to Speech via Multimodal Depth Up-Scaling [52.02344262645619]
  We propose Multimodal Depth Upscaling, where new transformer layers are inserted into a frozen text LLM and only the added layers are trained on speech data.<n>Experiments with SmolLM2-360M and SmolLM2-1.7B on 48k hours of English Automatic Speech Recognition (ASR) data show that depth up-scaling achieves ASR comparable to full fine-tuning.<n>We further show that incorporating E-Branchformer, an architecture designed for speech recognition, as the inserted layers achieves ASR that matches or surpasses full fine-tuning on the larger model while reducing text degradation by over 75% with 60% fewer trainable parameters
  arXiv  Detail & Related papers  (2026-04-01T05:16:06Z)
• Simultaneous Speech-to-Speech Translation Without Aligned Data [52.467808474293605]
  Simultaneous speech translation requires translating source speech into a target language in real-time.<n>We propose Hibiki-Zero, which eliminates the need for word-level alignments entirely.<n>Hibiki-Zero achieves state-of-the-art performance in translation accuracy, latency, voice transfer, and naturalness across five X-to-English tasks.
  arXiv  Detail & Related papers  (2026-02-11T17:41:01Z)
• Get away with less: Need of source side data curation to build parallel corpus for low resource Machine Translation [3.3393607383304253]
  We develop framework that screens source sentences to form efficient parallel text.<n>We find that by training mostly on complex sentences from both existing and synthetic datasets, our method significantly improves translation quality.<n>This approach not only reduces MT systems training cost by reducing training data requirement, but also showcases LALITA's utility in data augmentation.
  arXiv  Detail & Related papers  (2026-01-13T15:05:19Z)
• Efficient and Adaptive Simultaneous Speech Translation with Fully Unidirectional Architecture [14.056534007451763]
  Simultaneous speech translation (SimulST) produces translations incrementally while processing partial speech input.<n>Existing LLM-based SimulST approaches incur significant computational overhead due to repeated encoding of bidirectional speech encoder.<n>We introduce Efficient and Adaptive Simultaneous Speech Translation (EASiST) with fully unidirectional architecture.
  arXiv  Detail & Related papers  (2025-04-16T06:46:15Z)
• InfiniSST: Simultaneous Translation of Unbounded Speech with Large Language Model [10.40867923457809]
  InfiniSST is a novel approach that formulates SST as a multi-turn dialogue task.<n>We construct translation trajectories and robust segments from MuST-C with multi-latency augmentation during training.<n> Experiments on MuST-C En-Es, En-De, and En-Zh demonstrate that InfiniSST reduces computation-aware latency by 0.5 to 1 second.
  arXiv  Detail & Related papers  (2025-03-04T19:51:29Z)
• Alleviating Distribution Shift in Synthetic Data for Machine Translation Quality Estimation [55.73341401764367]
  We introduce DCSQE, a novel framework for alleviating distribution shift in synthetic QE data.<n> DCSQE uses references, i.e., translation supervision signals, to guide both the generation and annotation processes.<n>Experiments demonstrate that DCSQE outperforms SOTA baselines in both supervised and unsupervised settings.
  arXiv  Detail & Related papers  (2025-02-27T10:11:53Z)
• FASST: Fast LLM-based Simultaneous Speech Translation [9.65638081954595]
  Simultaneous speech translation (SST) takes streaming speech input and generates text translation on the fly. We propose FASST, a fast large language model based method for streaming speech translation. Experiment results show that FASST achieves the best quality-latency trade-off.
  arXiv  Detail & Related papers  (2024-08-18T10:12:39Z)
• Prefix Text as a Yarn: Eliciting Non-English Alignment in Foundation Language Model [50.339632513018934]
  supervised fine-tuning (SFT) has been a straightforward approach for tailoring the output of foundation large language model (LLM) to specific preferences. We critically examine this hypothesis within the scope of cross-lingual generation tasks. We introduce a novel training-free alignment method named PreTTY, which employs minimal task-related prior tokens.
  arXiv  Detail & Related papers  (2024-04-25T17:19:36Z)
• Strategies for improving low resource speech to text translation relying on pre-trained ASR models [59.90106959717875]
  This paper presents techniques and findings for improving the performance of low-resource speech to text translation (ST) We conducted experiments on both simulated and real-low resource setups, on language pairs English - Portuguese, and Tamasheq - French respectively.
  arXiv  Detail & Related papers  (2023-05-31T21:58:07Z)
• CTC-based Non-autoregressive Speech Translation [51.37920141751813]
  We investigate the potential of connectionist temporal classification for non-autoregressive speech translation. We develop a model consisting of two encoders that are guided by CTC to predict the source and target texts. Experiments on the MuST-C benchmarks show that our NAST model achieves an average BLEU score of 29.5 with a speed-up of 5.67$times$.
  arXiv  Detail & Related papers  (2023-05-27T03:54:09Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.