ExPosST: Explicit Position Allocation for LLM-Based Simultaneous Machine Translation

ACL ARR 2025 July Submission978 Authors

Abstract: Large Language Models (LLMs) have achieved remarkable performance in simultaneous machine translation (SimulMT) via attention mask and positional reordering strategies. However, these approaches have strict constraints on positional encoding methods, such as ALiBi, which limit their general application. In this work, we introduce ExPosST, a simple and general framework to apply decoder-only LLMs to SimulMT tasks. ExPosST explicitly allocates the position range in the source and translation tokens, allowing decoding with KV cache under all positional methods. Experiments on multiple models show that ExPosST has comparable performance with state-of-the-art approaches in LLMs using ALiBi, while outperforming them in mainstream RoPE-based LLMs.

Paper Type: Short

Research Area: Machine Translation

Research Area Keywords: efficient inference for MT

Contribution Types: NLP engineering experiment

Languages Studied: English, French, Italian, Dutch, Romanian, German

Reassignment Request Area Chair: This is not a resubmission

Reassignment Request Reviewers: This is not a resubmission

A1 Limitations Section: This paper has a limitations section.

A2 Potential Risks: N/A

B Use Or Create Scientific Artifacts: Yes

B1 Cite Creators Of Artifacts: Yes

B1 Elaboration: Section 4.1

B2 Discuss The License For Artifacts: N/A

B3 Artifact Use Consistent With Intended Use: N/A

B4 Data Contains Personally Identifying Info Or Offensive Content: N/A

B5 Documentation Of Artifacts: N/A

B6 Statistics For Data: N/A

C Computational Experiments: Yes

C1 Model Size And Budget: Yes

C1 Elaboration: Section 4.1

C2 Experimental Setup And Hyperparameters: Yes

C2 Elaboration: Appendix A

C3 Descriptive Statistics: No

C3 Elaboration: Due to limited computational resources, all experiments were conducted as single run.

C4 Parameters For Packages: Yes

C4 Elaboration: Section 4.1

D Human Subjects Including Annotators: No

D1 Instructions Given To Participants: N/A

D2 Recruitment And Payment: N/A

D3 Data Consent: N/A

D4 Ethics Review Board Approval: N/A

D5 Characteristics Of Annotators: N/A

E Ai Assistants In Research Or Writing: Yes

E1 Information About Use Of Ai Assistants: No

E1 Elaboration: We only utilized AI Assistants for refining sentences in the paper, not for content generation.

Author Submission Checklist: yes

Submission Number: 978