Machine translation

Enhancing e-commerce product title translation with retrieval-augmented generation and large language models

Bryan Zhang, Taichi Nakatani, Stephan Walter

CIKM 2024 Workshop on Data-Centric AI

2024

E-commerce stores enable multilingual product discovery which require accurate product title translation. Multilingual large language models (LLMs) have shown promising capacity to perform machine translation tasks, and it can also enhance and translate product titles cross-lingually in one step. However, product title translation often requires more than just language conversion because titles are short

Conversational AI

A shocking amount of the web is machine translated: Insights from multi-way parallelism

Brian Thompson, Mehak Dhaliwal, Peter Frisch, Tobias Domhan, Marcello Federico

ACL Findings 2024

2024

We show that content on the web is often translated into many languages, and the low quality of these multi-way translations indicates they were likely created using Machine Translation (MT). Multi-way parallel, machine generated content not only dominates the translations in lower resource languages; it also constitutes a large fraction of the total web content in those languages. We also find evidence

Conversational AI

Eliciting better multilingual structured reasoning from LLMs through code

Bryan Li, Tamer Alkhouli, Daniele Bonadiman, Nikolaos Pappas, Saab Mansour

ACL 2024, SoCal NLP Symposium 2024, COLING 2025 Workshop on Scaling Up Multilingual & Multi-Cultural Evaluation

2024

The development of large language models (LLM) has shown progress on reasoning, though studies have largely considered either English or simple reasoning tasks. To address this, we introduce a multilingual structured reasoning and explanation dataset, termed xSTREET, that covers four tasks across six languages. xSTREET exposes a gap in base LLM performance between English and non-English reasoning tasks

Conversational AI

Impacts of misspelled queries on translation and product search

Greg Hanneman, Natawut Monaikul, Taichi Nakatani

ACL 2024

2024

Machine translation is used in e-commerce to translate second-language queries into the primary language of the store, to be matched by the search system against the product catalog. However, many queries contain spelling mistakes. We first present an analysis of the spelling-robustness of a population of MT systems, quantifying how spelling variations affect MT output, the list of returned products, and

Conversational AI

Fine-tuned machine translation metrics struggle in unseen domains

Vilém Zouhar, Shuoyang Ding, Anna Currey, Tatyana Badeka, Jenyuan Wang, Brian Thompson

ACL 2024

2024

We introduce a new, extensive multidimensional quality metrics (MQM) annotated dataset covering 11 language pairs in the biomedical domain. We use this dataset to investigate whether machine translation (MT) metrics which are fine-tuned on human-generated MT quality judgements are robust to domain shifts between training and inference. We find that fine-tuned metrics exhibit a substantial performance drop

Conversational AI

The fine-tuning paradox: Boosting translation quality without sacrificing LLM abilities

David Stap, Eva Hasler, Bill Byrne, Christof Monz, Ke Tran

ACL 2024

2024

Fine-tuning large language models (LLMs) for machine translation has shown improvements in overall translation quality. However, it is unclear what is the impact of fine-tuning on desirable LLM behaviors that are not present in neural machine translation models, such as steerability, inherent document-level translation abilities, and the ability to produce less literal translations. We perform an extensive

Conversational AI

M3T: A new benchmark dataset for multi-modal document-level machine translation

Benjamin Hsu, Xiaoyu Liu, Huayang Li, Yoshinari Fujinuma, Maria Nădejde, Xing Niu, Yair Kittenplon, Ron Litman, Raghavendra Pappagari

NAACL 2024

2024

Document translation poses a challenge for Neural Machine Translation (NMT) systems. Most document-level NMT systems rely on meticulously curated sentence-level parallel data, assuming flawless extraction of text from documents along with their precise reading order. These systems also tend to disregard additional visual cues such as the document layout, deeming it irrelevant. However, real-world documents

Computer vision

Making machine translation more robust, consistent, and stable

Benjamin Hsu

May 19, 2023

Training on pseudo-labeled data limits the consequences of slight input variations and prevents updated models from backsliding on particular tasks.

Conversational AI

Leveraging latent topic information to improve product machine translation

Bryan Zhang, Stephan Walter, Liling Tan, Amita Misra

Machine Translation Summit 2023 (MTS)

2023

The expectations of e-commerce customers include the ability to shop online in their preferred language. Modern e-commerce platforms utilize machine translation to provide multilingual product information at scale. However, maintaining machine translation quality that keeps up with an ever-expanding product information remains an open challenge for industrial machine translation systems. Topical clustering

Machine learning

Brand consistency for multilingual e-commerce machine translation

Bryan Zhang, Stephan Walter, Saurabh Chetan Birari, Ozlem Eren

Machine Translation Summit 2023 (MTS)

2023

Brand translations need to be consistently localized in e-commerce stores. Emerging brands and their localized forms are constantly appearing in the dynamic e-commerce landscape. These variant brand forms and aliases pose a challenge to brand handling in MT. This study examines the enforcement of brand consistency in MT at scale on the e-commerce sites worldwide. We propose various practical and sustainable

Conversational AI

Machine translation

Work with us