Over-generation Cannot Be Rewarded: Length-adaptive Average Lagging For Simultaneous Speech Translation

Abstract

Simultaneous speech translation (SimulST) systems aim at generating their output with the lowest possible latency, which is normally computed in terms of Average Lagging (AL). In this paper we highlight that, despite its widespread adoption, AL provides underestimated scores for systems that generate longer predictions compared to the corresponding references. We also show that this problem has practical relevance, as recent SimulST systems have indeed a tendency to over-generate. As a solution, we propose LAAL (Length-Adaptive Average Lagging), a modified version of the metric that takes into account the over-generation phenomenon and allows for unbiased evaluation of both under-/over-generating systems.

Stats

• citations8
• S2 citations—
• github stars0
• HF likes0
• heat score7.16
• arxiv keypapi2022over

Related papers

• CA*: Addressing Evaluation Pitfalls In Computation-aware Latency For Simultaneous Speech Translation (2024)0.00
• Better Late Than Never: Meta-evaluation Of Latency Metrics For Simultaneous Speech-to-text Translation (2025)1.81
• Efficient And Adaptive Simultaneous Speech Translation With Fully Unidirectional Architecture (2025)2.26
• Visualization: The Missing Factor In Simultaneous Speech Translation (2021)0.00
• Learning When To Speak: Latency And Quality Trade-offs For Simultaneous Speech-to-speech Translation With Offline Models (2023)0.00
• From Start To Finish: Latency Reduction Strategies For Incremental Speech Synthesis In Simultaneous Speech-to-speech Translation (2021)2.26
• Towards Achieving Human Parity On End-to-end Simultaneous Speech Translation Via LLM Agent (2024)0.00
• Simuls2s-llm: Unlocking Simultaneous Inference Of Speech Llms For Speech-to-speech Translation (2025)3.58