Machine Translation Digest for Aug 25 2025
Here is today's selection of cs.CL papers focused on advancements in multilingual and language-specific technologies. The papers explore diverse aspects such as metric development, readability-controlled paraphrasing, enhancement of vision-language models, and representation analysis in multilingual contexts. These studies highlight the ongoing efforts to improve accessibility and efficiency across different languages in machine translation and language technology.
COMET-poly: Machine Translation Metric Grounded in Other Candidates
Automated metrics for machine translation attempt to replicate human judgment. Unlike humans, who often assess a translation in the context of multiple alternatives, these metrics typically consider only the source sentence and a single translation. This discrepancy in the evaluation setup may negatively impact the performance of automated metrics. We propose two automated metrics that incorporate additional information beyond the single translation. COMET-polycand uses alternative translations of the same source sentence to compare and contrast with the translation at hand, thereby providing a more informed assessment of its quality. COMET-polyic, inspired by retrieval-based in-context learning, takes in translations of similar source texts along with their human-labeled quality scores to guide the evaluation. We find that including a single additional translation in COMET-polycand improves the segment-level metric performance (0.079 to 0.118 Kendall's tau-b correlation), with further gains when more translations are added. Incorporating retrieved examples in COMET-polyic yields similar improvements (0.079 to 0.116 Kendall's tau-b correlation). We release our models publicly.
German4All - A Dataset and Model for Readability-Controlled Paraphrasing in German
The ability to paraphrase texts across different complexity levels is essential for creating accessible texts that can be tailored toward diverse reader groups. Thus, we introduce German4All, the first large-scale German dataset of aligned readability-controlled, paragraph-level paraphrases. It spans five readability levels and comprises over 25,000 samples. The dataset is automatically synthesized using GPT-4 and rigorously evaluated through both human and LLM-based judgments. Using German4All, we train an open-source, readability-controlled paraphrasing model that achieves state-of-the-art performance in German text simplification, enabling more nuanced and reader-specific adaptations. We opensource both the dataset and the model to encourage further research on multi-level paraphrasing
Language-Specific Layer Matters: Efficient Multilingual Enhancement for Large Vision-Language Models
Large vision-language models (LVLMs) have demonstrated exceptional capabilities in understanding visual information with human languages but also exhibit an imbalance in multilingual capabilities. In this work, we delve into the multilingual working pattern of LVLMs and identify a salient correlation between the multilingual understanding ability of LVLMs and language-specific neuron activations in shallow layers. Building on this insight, we introduce PLAST, a training recipe that achieves efficient multilingual enhancement for LVLMs by Precise LAnguage-Specific layers fine-Tuning. PLAST first identifies layers involved in multilingual understanding by monitoring language-specific neuron activations. These layers are then precisely fine-tuned with question-translation pairs to achieve multilingual alignment. Our empirical results on MM-Bench and MMMB demonstrate that PLAST effectively improves the multilingual capabilities of LVLMs and achieves significant efficiency with only 14% of the parameters tuned. Further analysis reveals that PLAST can be generalized to low-resource and complex visual reasoning tasks, facilitating the language-specific visual information engagement in shallow layers.
Evaluating the Representation of Vowels in Wav2Vec Feature Extractor: A Layer-Wise Analysis Using MFCCs
Automatic Speech Recognition has advanced with self-supervised learning, enabling feature extraction directly from raw audio. In Wav2Vec, a CNN first transforms audio into feature vectors before the transformer processes them. This study examines CNN-extracted information for monophthong vowels using the TIMIT corpus. We compare MFCCs, MFCCs with formants, and CNN activations by training SVM classifiers for front-back vowel identification, assessing their classification accuracy to evaluate phonetic representation.
Information availability in different languages and various technological constraints related to multilinguism on the Internet
The usage of Internet has grown exponentially over the last two decades. The number of Internet users has grown from 16 Million to 1650 Million from 1995 to 2010. It has become a major repository of information catering almost every area. Since the Internet has its origin in USA which is English speaking country there is huge dominance of English on the World Wide Web. Although English is a globally acceptable language, still there is a huge population in the world which is not able to access the Internet due to language constraints. It has been estimated that only 20-25% of the world population speaks English as a native language. More and more people are accessing the Internet nowadays removing the cultural and linguistic barriers and hence there is a high growth in the number of non-English speaking users over the last few years on the Internet. Although many solutions have been provided to remove the linguistic barriers, still there is a huge gap to be filled. This paper attempts to analyze the need of information availability in different languages and the various technological constraints related to multi-linguism on the Internet.
| Curated by yukajii.com |