No Language Left Behind: Scaling Human-Centered Machine Translation

2022/07/11 by Nllb Team, Marta R. Costa-jussà, NLLB Team +82 · 1 voice · 261 citations
Computer Science · #68T50 #Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #FOS: Computer and information sciences #I.2.7 #Natural Language Processing Techniques #Text Readability and Simplification #Topic Modeling #acm:68T50 #cs.AI #cs.CL #msc:68T50

paper · pdf · doi:10.48550/arxiv.2207.04672

190 pages

openalex publication_date 2022/07/11 · arxiv created 2022/08/25 · arxiv updated 2022/08/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/30

Abstract

Driven by the goal of eradicating language barriers on a global scale, machine translation has solidified itself as a key focus of artificial intelligence research today. However, such efforts have coalesced around a small subset of languages, leaving behind the vast majority of mostly low-resource languages. What does it take to break the 200 language barrier while ensuring safe, high quality results, all while keeping ethical considerations in mind? In No Language Left Behind, we took on this challenge by first contextualizing the need for low-resource language translation support through exploratory interviews with native speakers. Then, we created datasets and models aimed at narrowing the performance gap between low and high-resource languages. More specifically, we developed a conditional compute model based on Sparsely Gated Mixture of Experts that is trained on data obtained with novel and effective data mining techniques tailored for low-resource languages. We propose multiple architectural and training improvements to counteract overfitting while training on thousands of tasks. Critically, we evaluated the performance of over 40,000 different translation directions using a human-translated benchmark, Flores-200, and combined human evaluation with a novel toxicity benchmark covering all languages in Flores-200 to assess translation safety. Our model achieves an improvement of 44% BLEU relative to the previous state-of-the-art, laying important groundwork towards realizing a universal translation system. Finally, we open source all contributions described in this work, accessible at https://github.com/facebookresearch/fairseq/tree/nllb.

Cited by

• Remedy-R: Generative Reasoning for Machine Translation Evaluation without Error Annotations
• Assessing the Impact of Typological Features on Multilingual Machine Translation in the Age of Large Language Models
• The Commodification of AI Sovereignty: Lessons from the Fight for Sovereign Oil
• BenCSSmark: Making the Social Sciences Count in LLM Research
• Form and Meaning in Intrinsic Multilingual Evaluations
• Beg to Differ: Understanding Reasoning-Answer Misalignment Across Languages
• The Tokenizer Tax: Quantifying and Explaining the Cross-Lingual Cost of Subword Tokenization for Indian Languages
• Inspect India Evals: An Open Benchmarking Framework for Evaluating Large Language Models in the Indian Linguistic and Cultural Context
• MPR-CiteG: Enhancing RAG with Multi-Portfolio Retrieval and Citation-Grounded Generation
• TokSuite: Measuring the Impact of Tokenizer Choice on Language Model Behavior
• Corpus of Cross-lingual Dialogues with Minutes and Detection of Misunderstandings
• Fine-Tuned In-Context Learners for Efficient Adaptation
• When the Gold Standard Isn't Necessarily Standard: Challenges of Evaluating the Translation of User-Generated Content
• Governance-Aware Hybrid Fine-Tuning for Multilingual Large Language Models
• Hearing to Translate: The Effectiveness of Speech Modality Integration into LLMs
• A First-Order Logic-Based Alternative to Reward Models in RLHF
• Beyond surface form: A pipeline for semantic analysis in Alzheimer's Disease detection from spontaneous speech
• Advancing Bangla Machine Translation Through Informal Datasets
• MiniLingua: A Small Open-Source LLM for European Languages
• PrahokBART: A Pre-trained Sequence-to-Sequence Model for Khmer Natural Language Generation
• MultiScript30k: Leveraging Multilingual Embeddings to Extend Cross Script Parallel Data
• Fluent Alignment with Disfluent Judges: Post-training for Lower-resource Languages
• Mitigating Catastrophic Forgetting in Target Language Adaptation of LLMs via Source-Shielded Updates
• What Language is This? Ask Your Tokenizer
• AdiBhashaa: A Community-Curated Benchmark for Machine Translation into Indian Tribal Languages
• Teaching Old Tokenizers New Words: Efficient Tokenizer Adaptation for Pre-trained Models
• M3DR: Towards Universal Multilingual Multimodal Document Retrieval
• Adapting Large Language Models to Low-Resource Tibetan: A Two-Stage Continual and Supervised Fine-Tuning Study
• Swivuriso: The South African Next Voices Multilingual Speech Dataset
• OmniFusion: Simultaneous Multilingual Multimodal Translations via Modular Fusion
• Exploring Performance Variations in Finetuned Translators of Ultra-Low Resource Languages: Do Linguistic Differences Matter?
• "AGI" team at SHROOM-CAP: Data-Centric Approach to Multilingual Hallucination Detection using XLM-RoBERTa
• RegionMarker: A Region-Triggered Semantic Watermarking Framework for Embedding-as-a-Service Copyright Protection
• "Don't Teach Minerva": Guiding LLMs Through Complex Syntax for Faithful Latin Translation with RAG
• Confounding Factors in Relating Model Performance to Morphology
• Enhancing LLM Code Generation Capabilities through Test-Driven Development and Code Interpreter
• NSL-MT: Linguistically Informed Negative Samples for Efficient Machine Translation in Low-Resource Languages
• MTQ-Eval: Multilingual Text Quality Evaluation for Language Models
• Introducing A Bangla Sentence - Gloss Pair Dataset for Bangla Sign Language Translation and Research
• LLiMba: Sardinian on a Single GPU -- Adapting a 3B Language Model to a Vanishing Romance Language
• A Picture is Worth a Thousand (Correct) Captions: A Vision-Guided Judge-Corrector System for Multimodal Machine Translation
• Rethinking what Matters: Effective and Robust Multilingual Realignment for Low-Resource Languages
• Ibom NLP: A Step Toward Inclusive Natural Language Processing for Nigeria's Minority Languages
• Iterative Layer-wise Distillation for Efficient Compression of Large Language Models
• It Takes Two: A Dual Stage Approach for Terminology-Aware Translation
• Evaluating Machine Translation Datasets for Low-Web Data Languages: A Gendered Lens
• BanglaSTEM: A Parallel Corpus for Technical Domain Bangla-English Translation
• Automatic Machine Translation Detection Using a Surrogate Multilingual Translation Model
• HPLT 3.0: Very Large-Scale Multilingual Resources for LLMs and MT. Mono- and Bi-lingual Data, Multilingual Evaluation, and Pre-Trained Models
• Leveraging the Cross-Domain & Cross-Linguistic Corpus for Low Resource NMT: A Case Study On Bhili-Hindi-English Parallel Corpus
• TransAlign: Machine Translation Encoders are Strong Word Aligners, Too
• Evaluating Prompt Scope and Demonstration Similarity in Local LLM Machine Translation
• Language Models are not Equally Robust to Non-Canonical Tokenization across Languages
• Pretraining Strategies using Monolingual and Parallel Data for Low-Resource Machine Translation
• Do Large Language Models Grasp The Grammar? Evidence from Grammar-Book-Guided Probing in Luxembourgish
• Text Simplification with Sentence Embeddings
• Quality-Aware Translation Tagging in Multilingual RAG system
• Model-Aware Tokenizer Transfer
• Explaining and Mitigating Crosslingual Tokenizer Inequities
• Finding the Sweet Spot: Trading Quality, Cost, and Speed During Inference-Time LLM Reflection
• The Art of Asking: Multilingual Prompt Optimization for Synthetic Data
• Tibetan Language and AI: A Comprehensive Survey of Resources, Methods and Challenges
• See the Text: From Tokenization to Visual Reading
• SemiAdapt and SemiLoRA: Efficient Domain Adaptation for Transformer-based Low-Resource Language Translation with a Case Study on Irish
• Chain-of-Thought Reasoning Improves Context-Aware Translation with Large Language Models
• Evaluating Large Language Models on Urdu Idiom Translation
• AFRICAPTION: Establishing a New Paradigm for Image Captioning in African Languages
• KITE: A Benchmark for Evaluating Korean Instruction-Following Abilities in Large Language Models
• Semantic Prosody in Machine Translation: the English-Chinese Case of Passive Structures
• MERLIN: A Testbed for Multilingual Multimodal Entity Recognition and Linking
• M²PO: Multi-Perspective Multi-Pair Preference Optimization for Machine Translation
• Sparse Subnetwork Enhancement for Underrepresented Languages in Large Language Models
• ACADATA: Parallel Dataset of Academic Data for Machine Translation
• Tokenization Disparities as Infrastructure Bias: How Subword Systems Create Inequities in LLM Access and Efficiency
• MoBiLE: Efficient Mixture-of-Experts Inference on Consumer GPU with Mixture of Big Little Experts
• LLM Reasoning for Machine Translation: Synthetic Data Generation over Thinking Tokens
• AMiD: Knowledge Distillation for LLMs with α-mixture Assistant Distribution
• LLM Knowledge is Brittle: Truthfulness Representations Rely on Superficial Resemblance
• End-to-end Automatic Speech Recognition and Speech Translation: Integration of Speech Foundational Models and LLMs
• ChatGPT as a Translation Engine: A Case Study on Japanese-English
• Sunflower: A New Approach To Expanding Coverage of African Languages in Large Language Models
• Parallel Tokenizers: Rethinking Encoder Models' Vocabulary Design in Cross-Lingual Transfer of Low-Resource Languages
• The African Languages Lab: A Collaborative Approach to Advancing Low-Resource African NLP
• SD-MVSum: Script-Driven Multimodal Video Summarization Method and Datasets
• Multilingual Routing in Mixture-of-Experts
• Distillation of Large Language Models via Concrete Score Matching
• AlignX: Advancing Multilingual Large Language Models with Multilingual Representation Alignment
• Multilingual Vision-Language Models, A Survey
• SiniticMTError: A Machine Translation Dataset with Error Annotations for Sinitic Languages
• Low-Resource English-Tigrinya MT: Leveraging Multilingual Models, Custom Tokenizers, and Clean Evaluation Benchmarks
• Tokenization and Representation Biases in Multilingual Models on Dialectal NLP Tasks
• Faster, Smaller, and Smarter: Task-Aware Expert Merging for Online MoE Inference
• EnAnchored-X2X: English-Anchored Optimization for Many-to-Many Translation
• CorIL: Towards Enriching Indian Language to Indian Language Parallel Corpora and Machine Translation Systems
• Evaluating Language Translation Models by Playing Telephone
• DRISHTIKON: A Multimodal Multilingual Benchmark for Testing Language Models' Understanding on Indian Culture
• WolBanking77: Wolof Banking Speech Intent Classification Dataset
• Enhancing Cross-Lingual Transfer through Reversible Transliteration: A Huffman-Based Approach for Low-Resource Languages
• Specification-Aware Machine Translation and Evaluation for Purpose Alignment
• CUTE: A Multilingual Dataset for Enhancing Cross-Lingual Knowledge Transfer in Low-Resource Languages
• Whisper-UT: A Unified Translation Framework for Speech and Text
• TrueMoE: Dual-Routing Mixture of Discriminative Experts for Synthetic Image Detection
• ReCoVeR the Target Language: Language Steering without Sacrificing Task Performance
• Do LLMs Understand Wine Descriptors Across Cultures? A Benchmark for Cultural Adaptations of Wine Reviews
• A comparison of pipelines for the translation of a low resource language based on transformers
• MTEB-NL and E5-NL: Embedding Benchmark and Models for Dutch
• SENSE models: an open source solution for multilingual and multimodal semantic-based tasks
• Linguistic Neuron Overlap Patterns to Facilitate Cross-lingual Transfer on Low-resource Languages
• TigerCoder: A Novel Suite of LLMs for Code Generation in Bangla
• COCO-Urdu: A Large-Scale Urdu Image-Caption Dataset with Multimodal Quality Estimation
• MoVoC: Morphology-Aware Subword Construction for Geez Script Languages
• MultimodalHugs: Enabling Sign Language Processing in Hugging Face
• MERLIN: Multi-Stage Curriculum Alignment for Multilingual Encoder-LLM Integration in Cross-Lingual Reasoning
• Building Large-Scale English-Romanian Literary Translation Resources with Open Models
• From Scarcity to Efficiency: Investigating the Effects of Data Augmentation on African Machine Translation
• mmBERT: A Modern Multilingual Encoder with Annealed Language Learning
• Recognizing Every Voice: Towards Inclusive ASR for Rural Bhojpuri Women
• Exploring Subjective Tasks in Farsi: A Survey Analysis and Evaluation of Language Models
• PRIM: Towards Practical In-Image Multilingual Machine Translation
• Hunyuan-MT Technical Report
• MedCOD: Enhancing English-to-Spanish Medical Translation of Large Language Models Using Enriched Chain-of-Dictionary Framework
• Robustness Assessment and Enhancement of Text Watermarking for Google's SynthID
• HAP: Hybrid Adaptive Parallelism for Efficient Mixture-of-Experts Inference
• It's All About In-Context Learning! Teaching Extremely Low-Resource Languages to LLMs
• Long Chain-of-Thought Reasoning Across Languages
• Filling the Gap for Uzbek: Creating Translation Resources for Southern Uzbek
• In2x at WMT25 Translation Task
• Utilizing Multilingual Encoders to Improve Large Language Models for Low-Resource Languages
• Label Smoothing is a Pragmatic Information Bottleneck
• Optimal Corpus Aware Training for Neural Machine Translation
• LingBench++: A Linguistically-Informed Benchmark and Reasoning Framework for Multi-Step and Cross-Cultural Inference with LLMs
• MegaWika 2: A More Comprehensive Multilingual Collection of Articles and their Sources
• A French Version of the OLDI Seed Corpus
• TIBSTC-CoT: A Multi-Domain Instruction Dataset for Chain-of-Thought Reasoning in Language Models
• Language Arithmetics: Towards Systematic Language Neuron Identification and Manipulation
• RL from Teacher-Model Refinement: Gradual Imitation Learning for Machine Translation
• Multi-Hypothesis Distillation of Multilingual Neural Translation Models for Low-Resource Languages
• On The Role of Pretrained Language Models in General-Purpose Text Embeddings: A Survey
• Phoneme-Level Visual Speech Recognition via Point-Visual Fusion and Language Model Reconstruction
• A Case Against Implicit Standards: Homophone Normalization in Machine Translation for Languages that use the Ge'ez Script
• From Neurons to Semantics: Evaluating Cross-Linguistic Alignment Capabilities of Large Language Models via Neurons Alignment
• FLEXITOKENS: Flexible Tokenization for Evolving Language Models
• ILID: Native Script Language Identification for Indian Languages
• LLMs Cannot Reliably Judge (Yet?): A Comprehensive Assessment on the Robustness of LLM-as-a-Judge
• mSTEB: Massively Multilingual Evaluation of LLMs on Speech and Text Tasks
• Multiple Choice Learning of Low-Rank Adapters for Language Modeling
• TACTIC: Translation Agents with Cognitive-Theoretic Interactive Collaboration
• Invariant-based Robust Weights Watermark for Large Language Models
• Cross-Lingual Transfer for Machine Translation in Turkic Languages
• ViDove: A Translation Agent System with Multimodal Context and Memory-Augmented Reasoning
• NeoBabel: A Multilingual Open Tower for Visual Generation
• Nile-Chat: Egyptian Language Models for Arabic and Latin Scripts
• GRAFT: A Graph-based Flow-aware Agentic Framework for Document-level Machine Translation
• La Leaderboard: A Large Language Model Leaderboard for Spanish Varieties and Languages of Spain and Latin America
• The Cognate Data Bottleneck in Language Phylogenetics
• HyperCLOVA X THINK Technical Report
• What Do Indonesians Really Need from Language Technology? A Nationwide Survey
• KaLM-Embedding-V2: Superior Training Techniques and Data Inspire A Versatile Embedding Model
• skLEP: A Slovak General Language Understanding Benchmark
• Bridging Compositional and Distributional Semantics: A Survey on Latent Semantic Geometry via AutoEncoder
• Has Machine Translation Evaluation Achieved Human Parity? The Human Reference and the Limits of Progress
• Beyond the Sentence: A Survey on Context-Aware Machine Translation with Large Language Models
• TranslationCorrect: A Unified Framework for Machine Translation Post-Editing with Predictive Error Assistance
• Prompt, Translate, Fine-Tune, Re-Initialize, or Instruction-Tune? Adapting LLMs for In-Context Learning in Low-Resource Languages
• Multilingual Tokenization through the Lens of Indian Languages: Challenges and Insights
• Data Quality Issues in Multilingual Speech Datasets: The Need for Sociolinguistic Awareness and Proactive Language Planning
• End-to-End Speech Translation for Low-Resource Languages Using Weakly Labeled Data
• Utility-Driven Speculative Decoding for Mixture-of-Experts
• MAS-LitEval : Multi-Agent System for Literary Translation Quality Assessment
• Just Go Parallel: Improving the Multilingual Capabilities of Large Language Models
• A Culturally-diverse Multilingual Multimodal Video Benchmark & Model
• KoGEC : Korean Grammatical Error Correction with Pre-trained Translation Models
• FormosanBench: Benchmarking Low-Resource Austronesian Languages in the Era of Large Language Models
• Q2E: Query-to-Event Decomposition for Zero-Shot Multilingual Text-to-Video Retrieval
• Unveiling Language-Specific Features in Large Language Models via Sparse Autoencoders
• MELABenchv1: Benchmarking Large Language Models against Smaller Fine-Tuned Models for Low-Resource Maltese NLP
• It's Not a Walk in the Park! Challenges of Idiom Translation in Speech-to-text Systems
• Different Speech Translation Models Encode and Translate Speaker Gender Differently
• Speech-to-Speech Translation Pipelines for Conversations in Low-Resource Languages
• MMD-Flagger: Leveraging Maximum Mean Discrepancy to Detect Hallucinations
• Statement-Tuning Enables Efficient Cross-lingual Generalization in Encoder-only Models
• The State of Large Language Models for African Languages: Progress and Challenges
• Chandomitra: Towards Generating Structured Sanskrit Poetry from Natural Language Inputs
• TESU-LLM: Training Speech-LLMs Without Speech via Unified Encoder Alignment
• Massively Multilingual Adaptation of Large Language Models Using Bilingual Translation Data
• Exploring In-context Example Generation for Machine Translation
• Translate With Care: Addressing Gender Bias, Neutrality, and Reasoning in Large Language Model Translations
• Graph-Assisted Culturally Adaptable Idiomatic Translation for Indic Languages
• Found in Translation: Measuring Multilingual LLM Consistency as Simple as Translate then Evaluate
• Bemba Speech Translation: Exploring a Low-Resource African Language
• Uncertainty Unveiled: Can Exposure to More In-context Examples Mitigate Uncertainty for Large Language Models?
• Dub-S2ST: Textless Speech-to-Speech Translation for Seamless Dubbing
• KIT's Low-resource Speech Translation Systems for IWSLT2025: System Enhancement with Synthetic Data and Model Regularization
• MT³: Scaling MLLM-based Text Image Machine Translation via Multi-Task Reinforcement Learning
• An End-to-End Model for Logits-Based Large Language Models Watermarking
• MMATH: A Multilingual Benchmark for Mathematical Reasoning
• Building a Functional Machine Translation Corpus for Kpelle
• Semantic Alignment of AI Models: Concept Collapse, Checkpoint Dynamics, and Cross-Lingual Transfer
• NileChat: Towards Linguistically Diverse and Culturally Aware LLMs for Local Communities
• Mutarjim: Advancing Bidirectional Arabic-English Translation with a Small Language Model
• Intent Classification on Low-Resource Languages with Query Similarity Search
• On Multilingual Encoder Language Model Compression for Low-Resource Languages
• MPO: Multilingual Safety Alignment via Reward Gap Optimization
• SSR-Zero: Simple Self-Rewarding Reinforcement Learning for Machine Translation
• HOFT: Householder Orthogonal Fine-tuning
• Comparative analysis of subword tokenization approaches for Indian languages
• From Tens of Hours to Tens of Thousands: Scaling Back-Translation for Speech Recognition
• In-Domain African Languages Translation Using LLMs and Multi-armed Bandits
• Set-LLM: A Permutation-Invariant LLM
• Not All Models Suit Expert Offloading: On Local Routing Consistency of Mixture-of-Expert Models
• Phonetic Perturbations Reveal Tokenizer-Rooted Safety Gaps in LLMs
• MultiHal: Multilingual Dataset for Knowledge-Graph Grounded Evaluation of LLM Hallucinations
• FuxiMT: Sparsifying Large Language Models for Chinese-Centric Multilingual Machine Translation
• From Unaligned to Aligned: Scaling Multilingual LLMs with Multi-Way Parallel Corpora
• Cross-Lingual Representation Alignment Through Contrastive Image-Caption Tuning
• Occult: Optimizing Collaborative Communication across Experts for Accelerated Parallel MoE Training and Inference
• Combining the Best of Both Worlds: A Method for Hybrid NMT and LLM Translation
• Improving Multilingual Language Models by Aligning Representations through Steering
• The Effect of Language Diversity When Fine-Tuning Large Language Models for Translation
• ExpertSteer: Intervening in LLMs through Expert Knowledge
• LLM-Based Evaluation of Low-Resource Machine Translation: A Reference-less Dialect Guided Approach with a Refined Sylheti-English Benchmark
• Multilingual Prompt Engineering in Large Language Models: A Survey Across NLP Tasks
• Low-Resource Language Processing: An OCR-Driven Summarization and Translation Pipeline
• The Devil Is in the Word Alignment Details: On Translation-Based Cross-Lingual Transfer for Token Classification Tasks
• MorphMark: Flexible Adaptive Watermarking for Large Language Models
• Are We Paying Attention to Her? Investigating Gender Disambiguation and Attention in Machine Translation
• ALOHA: Empowering Multilingual Agent for University Orientation with Hierarchical Retrieval
• LangFIR: Discovering Sparse Language-Specific Features from Monolingual Data for Language Steering
• Language translation, and change of accent for speech-to-speech task using diffusion model
• ReLI: A Language-Agnostic Approach to Human-Robot Interaction
• Intern-S1-Pro: Scientific Multimodal Foundation Model at Trillion Scale
• TamperBench: Systematically Stress-Testing LLM Safety Under Fine-Tuning and Tampering
• Overcoming language barriers via machine translation with sparse Mixture-of-Experts fusion of large language models
• LENSLLM: Unveiling Fine-Tuning Dynamics for LLM Selection
• Fine-Tuning LLMs for Low-Resource Dialect Translation: The Case of Lebanese
• Geometry-Aware Localized Watermarking for Copyright Protection in Embedding-as-a-Service
• DMDTEval: An Evaluation and Analysis of LLMs on Disambiguation in Multi-domain Translation
• Enhancing LLM Language Adaption through Cross-lingual In-Context Pre-training
• Anthropogenic Regional Adaptation in Multimodal Vision-Language Model
• Artificial intelligence is creating a new global linguistic hierarchy
• Calibrating Translation Decoding with Quality Estimation on LLMs
• Spatial Speech Translation: Translating Across Space With Binaural Hearables
• Identity and investment in the age of generative AI
• Efficient Multilingual Neural Machine Translation via Corpus-Driven Vocabulary Pruning: An English-Arabic Case Study
• Strengthening Target-Language Features: SAE-Based Steering for Multilingual Inference
• Rethinking Multilingual Continual Pretraining: Data Mixing for Adapting LLMs Across Languages and Resources
• Mind the Cap: Output-Budget Regimes Change the Measured Multilingual Reasoning Gap
• Translate-R1: Cost-Aware Translation Tool Use via Reinforcement Learning
• Tokenization Matters: Improving Zero-Shot NER for Indic Languages
• Optimizing LLMs for Italian: Reducing Token Fertility and Enhancing Efficiency Through Vocabulary Adaptation
• The Bitter Lesson Learned from 2,000+ Multilingual Benchmarks
• Multilingual Scholarly Publishing and Artificial Intelligence Translation Tools: Weighing Social Justice and Climate Justice
• Trans-Zero: Self-Play Incentivizes Large Language Models for Multilingual Translation Without Parallel Data
• A Multimodal Recaptioning Framework to Account for Perceptual Diversity Across Languages in Vision-Language Modeling
• Multilingual Contextualization of Large Language Models for Document-Level Machine Translation
• GlotEval: A Test Suite for Massively Multilingual Evaluation of Large Language Models
• MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
• Lugha-Llama: Adapting Large Language Models for African Languages
• Two Intermediate Translations Are Better Than One: Fine-tuning LLMs for Document-level Translation Refinement
• HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference
• CARE: Multilingual Human Preference Learning for Cultural Awareness

Discussions

• NLLB paper has some evidence for this: arxiv.org/abs/2207.04672 [bsky, 1 points, 0 comments]

Related

• Building Machine Translation Systems for the Next Thousand Languages
• Is Small Language Model the Silver Bullet to Low-Resource Languages Machine Translation?
• Quality or Quantity? On Data Scale and Diversity in Adapting Large Language Models for Low-Resource Translation
• Participatory Research for Low-resourced Machine Translation: A Case Study in African Languages
• Bridging the Linguistic Divide: A Survey on Leveraging Large Language Models for Machine Translation