Multi-modal simultaneous machine translation fusion of image information

Simultaneous machine translation has recently gained traction thanks to significant quality improvements and the advent of streaming applications. Simultaneous translation systems need to find a trade-off between translation quality and response time, and with this purpose multiple latency measures have been proposed. However, latency evaluations for simultaneous translation are estimated at the sentence level, not taking into account the sequential nature of a streaming scenario. Indeed, these sentence-level latency measures are not well suited for continuous stream translation resulting in figures that are not coherent with the simultaneous translation policy of the system being assessed. This work proposes a stream-level adaptation of the current latency measures based on a re-segmentation approach applied to the output translation, that is successfully evaluated on streaming conditions for a reference IWSLT task.

Simultaneous translation involves translating a sentence before the speaker's utterance is completed in order to realize real-time understanding in multiple languages.This task is significantly more challenging than the general full sentence translation because of the shortage of input information during decoding.To alleviate this shortage, we propose multimodal simultaneous neural machine translation (MSNMT), which leverages visual information as an additional modality.Our experiments with the Multi30k dataset showed that MSNMT significantly outperforms its textonly counterpart in more timely translation situations with low latency.Furthermore, we verified the importance of visual information during decoding by performing an adversarial evaluation of MSNMT, where we studied how models behaved with incongruent input modality and analyzed the effect of different word order between source and target languages.

Simultaneous machine translation (SiMT) outputs translation while reading source sentence and hence requires a policy to decide whether to wait for the next source word (READ) or generate a target word (WRITE), the actions of which form a read/write path. Although the read/write path is essential to SiMT performance, no direct supervision is given to the path in the existing methods. In this paper, we propose a method of dual-path SiMT which introduces duality constraints to direct the read/write path. According to duality constraints, the read/write path in source-totarget and target-to-source SiMT models can be mapped to each other. As a result, the two SiMT models can be optimized jointly by forcing their read/write paths to satisfy the mapping. Experiments on En-Vi and De-En tasks show that our method can outperform strong baselines under all latency.

We present Sequential Policy Optimization for Simultaneous Machine Translation (SeqPO-SiMT), a new policy optimization framework that defines the simultaneous machine translation (SiMT) task as a sequential decision making problem, incorporating a tailored reward to enhance translation quality while reducing latency.In contrast to popular Reinforcement Learning from Human Feedback (RLHF) methods, such as PPO and DPO, which are typically applied in single-step tasks, SeqPO-SiMT effectively tackles the multi-step SiMT task.This intuitive framework allows the SiMT LLMs to simulate and refine the SiMT process using a tailored reward.We conduct experiments on six datasets from diverse domains for En Zh and Zh En SiMT tasks, demonstrating that SeqPO-SiMT consistently achieves significantly higher translation quality with lower latency.In particular, SeqPO-SiMT outperforms the supervised fine-tuning (SFT) model by 1.13 points in COMET, while reducing the Average Lagging by 6.17 in the NEWSTEST2021 En Zh dataset.While SiMT operates with far less context than offline translation, the SiMT results of SeqPO-SiMT on 7B LLM surprisingly rival the offline translation of high-performing LLMs, including Qwen-2.5-7B-Instruct and LLaMA-3-8B-Instruct.

Simultaneous speech translation aims to maintain translation quality while minimizing the delay between reading input and incrementally producing the output. We propose a new general-purpose prediction action which predicts future words in the input to improve quality and minimize delay in simultaneous translation. We train this agent using reinforcement learning with a novel reward function. Our agent with prediction has better translation quality and less delay compared to an agent-based simultaneous translation system without prediction.

Simultaneous machine translation, which aims at a realtime translation, is useful in many live scenarios but very challenging due to the trade-off between accuracy and latency. To achieve the balance for both, the model needs to wait for appropriate streaming text (READ policy) and then generates its translation (WRITE policy). However, WRITE policies of previous work either are specific to the method itself due to the end-to-end training or suffer from the input mismatch between training and decoding for the non-end-to-end training. Therefore, it is essential to learn a generic and better WRITE policy for simultaneous machine translation. Inspired by strategies utilized by human interpreters and "wait" policies, we propose a novel adaptive prefix-to-prefix training policy called LEAPT, which allows our machine translation model to learn how to translate source sentence prefixes and make use of the future context. Experiments show that our proposed methods greatly outperform competitive baselines and achieve promising results.

Multimodal machine translation (MMT) aims to improve neural machine translation (NMT) with additional visual information, but most existing MMT methods require paired input of source sentence and image, which makes them suffer from shortage of sentence-image pairs. In this paper, we propose a phrase-level retrieval-based method for MMT to get visual information for the source input from existing sentence-image data sets so that MMT can break the limitation of paired sentence-image input. Our method performs retrieval at the phrase level and hence learns visual information from pairs of source phrase and grounded region, which can mitigate data sparsity. Furthermore, our method employs the conditional variational auto-encoder to learn visual representations which can filter redundant visual information and only retain visual information related to the phrase. Experiments show that the proposed method significantly outperforms strong baselines on multiple MMT datasets, especially when the textual context is limited.

Speech translation systems commonly couple automatic speech recognition (ASR) and machine translation (MT) components. Hereby the automatic segmentation of the ASR output for the subsequent MT is critical for the overall performance. In simultaneous translation systems, which require a continuous output with a low latency, chunking of the ASR output into translatable segments is even more critical. This paper addresses the question how utterance chunking influences machine translation performance in an empirical study. In addition, the machine translation performance is also set in relation to the segment length produced by the chunking strategy, which is important for simultaneous translation. Therefore, we compare different relatively simple chunking/ segmentation strategies on speech recognition hypotheses as well as on reference transcripts.

To achieve high-quality translation with low latency, a Simultaneous Speech Translation (SimulST) system relies on a policy module to decide whether to translate immediately or wait for additional streaming input, along with a translation model capable of effectively handling partial speech input. Prior research has tackled these components separately, either using ``wait-k'' policies based on fixed-length segments or detected word boundaries, or dynamic policies based on different strategies (e.g., meaningful units), while employing offline models for prefix-to-prefix translation. In this paper, we propose Divergence-Guided Simultaneous Speech Translation (DiG-SST), a tightly integrated approach focusing on both translation quality and latency for streaming input. Specifically, we introduce a simple yet effective prefix-based strategy for training translation models with partial speech input, and develop an adaptive policy that makes read/write decisions for the translation model based on the expected divergence in translation distributions resulting from future input. Our experiments on multiple translation directions of the MuST-C benchmark demonstrate that our approach achieves a better trade-off between translation quality and latency compared to existing methods.

Simultaneous speech translation (SimulST) systems must balance translation quality with response time, making latency measurement crucial for evaluating their real-world performance.However, there has been a longstanding belief that current metrics yield unrealistically high latency measurements in unsegmented streaming settings.In this paper, we investigate this phenomenon, revealing its root cause in a fundamental misconception underlying existing latency evaluation approaches.We demonstrate that this issue affects not only streaming but also segment-level latency evaluation across different metrics.Furthermore, we propose a modification to correctly measure computationaware latency for SimulST systems, addressing the limitations present in existing metrics.

Current multimodal neural machine translation (MNMT) approaches primarily focus on ensuring consistency between visual annotations and the source language, often overlooking the broader aspect of multimodal coherence, including target–visual and bilingual–visual alignment. In this paper, we propose a novel approach that effectively leverages target–visual consistency (TVC) and bilingual–visual consistency (BiVC) to improve MNMT performance. Our method leverages visual annotations depicting concepts across bilingual parallel sentences to enhance multimodal coherence in translation. We exploit target–visual harmony by extracting contextual cues from visual annotations during auto-regressive decoding, incorporating vital future context to improve target sentence representation. Additionally, we introduce a consistency loss promoting semantic congruence between bilingual sentence pairs and their visual annotations, fostering a tighter integration of textual and visual modalities. Extensive experiments on diverse multimodal translation datasets empirically demonstrate our approach’s effectiveness. This visually aware, data-driven framework opens exciting opportunities for intelligent learning, adaptive control, and robust distributed optimization of multi-agent systems in uncertain, complex environments. By seamlessly fusing multimodal data and machine learning, our method paves the way for novel control paradigms capable of effectively handling the dynamics and constraints of real-world multi-agent applications.

This article concerns the trans-editing (simultaneous translation and editing) of coverage of the 2008 US presidential elections on BBC World Service websites. We investigate how English-language source texts were reworked in Arabic, Persian, Tamil, and Turkish, with a detailed analysis of the structuring, content, and rhetoric of a sample text in English and in these other languages. This analysis shows that, while the BBC’s corporate aim is to provide a univocal service across its multilingual output, this aim is in tension with widely differing journalistic norms, and differing assumptions about audience knowledge and needs, in each of the World Service’s language departments. The ‘melody’ remains essentially the same, but it is orchestrated differently by each department.

Multi-modal neural machine translation (MNMT) aims to translate sentences from the source language into the target language with the aid of corresponding images. Unfortunately, there is a considerable modality gap between the semantic-related images and texts in terms of data form and semantic expression. How to fully incorporate visual information into texts to enhance the performance of machine translation is one of the critical issues for MNMT. However, the initial visual and textual features are generally extracted with their modality-specific models; Consequently, there is a considerable representation gap between images and texts. Most previous MNMT works prefer only to adopt the feature-level fusion strategies to learn multi-modal representation, while the modality representation gap is often ignored. To this end, this paper proposes a progressive multi-modal Transformer (ProMul-Trans) with Modality Difference-Aware (MDA) to address the visual-to-textual fusion problem raised in MNMT. We first employ MDA to capture the modality-consistency information by taking visual and textual representations as inputs in each Transformer layer. Then a layer-level progressive fusion (Layer-ProFusion) strategy is adopted to progressively align visual and textual representations layer-by-layer to enhance machine translation performance. Experiment results on the Multi30k dataset are conducted, and the results show that the proposed approach outperforms the compared state-of-the-art (SOTA) methods on English <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\rightarrow$</tex-math></inline-formula> German (En <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\rightarrow$</tex-math></inline-formula> De), English <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\rightarrow$</tex-math></inline-formula> French (En <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\rightarrow$</tex-math></inline-formula> Fr) and English <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\rightarrow$</tex-math></inline-formula> Czech (En <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\rightarrow$</tex-math></inline-formula> Cs) tasks. We release the code at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/JunjieYe-MMT/HierProMul-Trans</uri> .

This paper proposed a supervised visual attention mechanism for multimodal neural machine translation (MNMT), trained with constraints based on manual alignments between words in a sentence and their corresponding regions of an image. The proposed visual attention mechanism captures the relationship between a word and an image region more precisely than a conventional visual attention mechanism trained through MNMT in an unsupervised manner. Our experiments on English-German and German-English translation tasks using the Multi30k dataset and on English-Japanese and Japanese-English translation tasks using the Flickr30k Entities JP dataset show that a Transformer-based MNMT model can be improved by incorporating our proposed supervised visual attention mechanism and that further improvements can be achieved by combining it with a supervised cross-lingual attention mechanism (up to +1.61 BLEU, +1.7 METEOR).

Visual Genome is a dataset connecting structured image information with English language. We present “Hindi Visual Genome”, a multi-modal dataset consisting of text and images suitable for English-Hindi multi-modal machine translation task and multi-modal research. We have selected short English segments (captions) from Visual Genome along with the associated images and automatically translated them to Hindi. A careful manual post-editing followed which took the associated images into account. Overall, we prepared a set of 31,525 segments (which we conveniently split into training, development and testdata), accompanied by a challenge test set of 1,400 segments. This challenge test set was created by searching for particularly ambiguous English words based on the embedding similarity and manually selecting those where the image helps to resolvethe ambiguity. Our dataset is the first manually revised dataset for multi-modal English-Hindi machine translation, freely available for non-commercial research purposes. Our Hindi version of Visual Genome also allows to create Hindi image labelers or other practical tools. Hindi Visual Genome also served in Workshop on Asian Translation (WAT) 2019 Multi-Modal Translation Task.