Enabling Unseen Language QA Without Training LLMs: XBridge’s Plug‑in Multilingual Extension

XBridge combines a pre‑trained English‑centric LLM with an external multilingual NMT model via optimal‑transport alignment and a three‑stage training scheme, allowing zero‑training of the LLM while achieving high‑quality question answering and generation for low‑resource and unseen languages, narrowing the performance gap with high‑resource languages.

LLMNMTXBridge

0 likes · 8 min read

Enabling Unseen Language QA Without Training LLMs: XBridge’s Plug‑in Multilingual Extension

NewBeeNLP

Jul 31, 2024 · Artificial Intelligence

How Continual Pre‑Training Boosts Llama‑3’s Chinese and Scientific Reasoning

This report presents a continual pre‑training approach that significantly enhances Llama‑3 (8B)’s Chinese language proficiency and scientific reasoning by using a carefully mixed corpus of existing and synthetic data, detailing the bilingual adaptation and synthetic‑enhancement stages, data‑mixing and curriculum strategies, and demonstrating strong results across multilingual and scientific benchmarks without sacrificing original capabilities.

BenchmarkingLLMLlama-3

0 likes · 9 min read

How Continual Pre‑Training Boosts Llama‑3’s Chinese and Scientific Reasoning

DataFunTalk

Jun 28, 2023 · Artificial Intelligence

Building and Applying a Multi‑Language Product Knowledge Graph at Shopee

This presentation details Shopee's approach to constructing a multilingual product knowledge graph, covering ontology modeling, data acquisition, fusion techniques, and practical applications, while discussing challenges, model architectures, and future directions for large‑scale e‑commerce AI systems.

e‑commerceknowledge graphmachine learning

0 likes · 20 min read

Building and Applying a Multi‑Language Product Knowledge Graph at Shopee

iQIYI Technical Product Team

Jul 9, 2021 · Artificial Intelligence

iQIYI Multi‑Language Subtitle Machine Translation: Practice, Model Exploration, and Deployment

iQIYI’s multi‑language subtitle machine‑translation system combines a one‑to‑many transformer, context‑fusion encoding, four custom attention masks, masked language modeling, global decoding loss, reconstruction and error‑correction modules, plus pronoun, idiom and name‑handling tricks, achieving higher quality than third‑party services and even surpassing human translation for several languages.

Error CorrectionOne-to-Many ModelSubtitle Translation

0 likes · 17 min read

iQIYI Multi‑Language Subtitle Machine Translation: Practice, Model Exploration, and Deployment

DataFunTalk

Feb 20, 2021 · Artificial Intelligence

Industrial-Scale Machine Translation at Bytedance: Applications, Demos, and Research Advances

This article presents Bytedance's industrial machine‑translation platform, describing its global deployment, diverse product demos, underlying sequence‑to‑sequence models, BERT‑enhanced training strategies, prune‑tune sparsity techniques, multilingual pre‑training, document translation, and a high‑performance inference engine.

BERTmachine translationmultilingual NLP

0 likes · 19 min read

Industrial-Scale Machine Translation at Bytedance: Applications, Demos, and Research Advances