Papers by Vishrav Chaudhary
Facebook AI’s WAT19 Myanmar-English Translation Task Submission (D19-52)
Copied to clipboard
Peng-Jen Chen, Jiajun Shen, Matthew Le, Vishrav Chaudhary, Ahmed El-Kishky, Guillaume Wenzek, Myle Ott, Marc’Aurelio Ranzato
| Challenge: | Using back-translation, we can improve generalization by using noisy channel re-ranking and ensembling. |
| Approach: | They propose to use BPE-based transformer models to leverage monolingual data to improve generalization and use noisy channel re-ranking and ensembling to improve results. |
| Outcome: | The proposed system improves on the baseline system trained exclusively on the provided small parallel dataset, and the human evaluation and BLEU score are higher. |
WikiMatrix: Mining 135M Parallel Sentences in 1620 Language Pairs from Wikipedia (2021.eacl-main)
Copied to clipboard
| Challenge: | a new approach to extract parallel sentences from Wikipedia articles is proposed . the approach is based on multilingual sentence embeddings, but does not limit it to English . |
| Approach: | They propose to automatically extract parallel sentences from Wikipedia articles in 96 languages . they train neural MT baseline systems on the mined data and evaluate them on the TED corpus . |
| Outcome: | The proposed approach extracts parallel sentences from Wikipedia articles in 96 languages . the extracted sentences achieve strong BLEU scores for many language pairs . |
Current Advances in LLM Reasoning (2026.acl-tutorials)
Copied to clipboard
| Challenge: | This tutorial examines comprehensive evaluation strategies to assess the reasoning abilities of large language models (LLMs) advanced inference time methods and post-training methods that aim to make LLMs think more like humans are discussed in this tutorial. |
| Approach: | This tutorial explores comprehensive evaluation strategies to assess the reasoning abilities of large language models (LLMs) and discusses two types of methods to improve models’ reasoning: advanced inference time methods, structured and self-improvement inference methods, and post-training methods, such as RLHF, DPO, and GRPO. |
| Outcome: | This tutorial examines evaluation strategies to assess the reasoning abilities of large language models and discusses two types of methods to improve models’ reasoning. |
The FLORES Evaluation Datasets for Low-Resource Machine Translation: Nepali–English and Sinhala–English (D19-1)
Copied to clipboard
Francisco Guzmán, Peng-Jen Chen, Myle Ott, Juan Pino, Guillaume Lample, Philipp Koehn, Vishrav Chaudhary, Marc’Aurelio Ranzato
| Challenge: | a vast majority of language pairs in the world are considered low-resource because they have little parallel data available. |
| Approach: | They propose to use a dataset to evaluate methods trained on low-resource language pairs . they report baseline performance using supervised, weakly supervised and semi-supervised settings . |
| Outcome: | The proposed evaluation datasets show that current state-of-the-art methods perform poorly on this benchmark, posing a challenge to the research community working on low-resource MT. |
CCAligned: A Massive Collection of Cross-Lingual Web-Document Pairs (2020.emnlp-main)
Copied to clipboard
| Challenge: | Cross-lingual document alignment aims to identify pairs of documents in two distinct languages that are of comparable content or translations of each other. |
| Approach: | They exploit the signals embedded in URLs to label web documents at scale with an average precision of 94.5% across different language pairs. |
| Outcome: | The proposed method can label documents at 94.5% across languages with high precision . the proposed method is useful for low-resource languages with limited resources . |
Classification-based Quality Estimation: Small and Efficient Models for Real-world Applications (2021.emnlp-main)
Copied to clipboard
| Challenge: | Sentence-level Quality estimation (QE) is traditionally a regression task . but large multilingual contextualized language models are expensive and infeasible for real-world applications. |
| Approach: | They evaluate several model compression techniques for QE and find they are inefficient . they argue that a full model parameterization is required to achieve SoTA results . |
| Outcome: | The proposed models are poorly expressive in a regression task, the authors argue . they show that reframing QE as a classification problem and evaluating models would improve their performance in real-world applications. |
MLQE-PE: A Multilingual Quality Estimation and Post-Editing Dataset (2022.lrec-1)
Copied to clipboard
Marina Fomicheva, Shuo Sun, Erick Fonseca, Chrysoula Zerva, Frédéric Blain, Vishrav Chaudhary, Francisco Guzmán, Nina Lopatina, Lucia Specia, André F. T. Martins
| Challenge: | Existing datasets for machine translation quality estimation and post-editing have several shortcomings. |
| Approach: | They propose a dataset for machine translation quality estimation and automatic post-editing . they report the performance of baseline systems trained on the MLQE-PE dataset . |
| Outcome: | The proposed dataset contains human labels for up to 10,000 translations per language pair. |
Alternative Input Signals Ease Transfer in Multilingual Machine Translation (2022.acl-long)
Copied to clipboard
| Challenge: | Recent work in multilingual machine translation (MMT) has focused on the potential of positive transfer between languages. |
| Approach: | They propose to augment training data with alternative signals that unify different writing systems, such as phonetic, romanized, and transliterated input. |
| Outcome: | The proposed model outperforms strong ensemble baselines on Indic and Turkic languages by 1.3 BLEU points on both languages. |
The Flores-101 Evaluation Benchmark for Low-Resource and Multilingual Machine Translation (2022.tacl-1)
Copied to clipboard
Naman Goyal, Cynthia Gao, Vishrav Chaudhary, Peng-Jen Chen, Guillaume Wenzek, Da Ju, Sanjana Krishnan, Marc’Aurelio Ranzato, Francisco Guzmán, Angela Fan
| Challenge: | a lack of good evaluation benchmarks hinders progress in low-resource and multilingual machine translation . despite advances in translation quality for a handful of languages, many low-source languages are not even supported by most popular translation engines. |
| Approach: | They propose a high-quality evaluation benchmark for machine translation using 3001 sentences from Wikipedia . they aim to improve evaluation of models on long tail of low-resource languages . |
| Outcome: | The proposed evaluation benchmarks are based on 3001 sentences extracted from Wikipedia . the results show that the models can be used to evaluate multilingual systems . |
Few-shot Learning with Multilingual Generative Language Models (2022.emnlp-main)
Copied to clipboard
Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O’Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li
| Challenge: | Large-scale generative language models such as GPT-3 are competitive few-shot learners. |
| Approach: | They train multilingual generative language models on a corpus covering a diverse set of languages and study their few- and zero-shot learning capabilities. |
| Outcome: | The proposed model outperforms GPT-3 on 171 out of 182 directions with 32 training examples and surpasses the official supervised baseline in 45 directions. |
Self-training Improves Pre-training for Natural Language Understanding (2021.naacl-main)
Copied to clipboard
Jingfei Du, Edouard Grave, Beliz Gunel, Vishrav Chaudhary, Onur Celebi, Michael Auli, Veselin Stoyanov, Alexis Conneau
| Challenge: | Unsupervised pretraining has led to improvements in natural language understanding . a data augmentation method can be used to generate labels for unlabeled examples . |
| Approach: | They propose a semi-supervised method which uses unlabeled data to retrieve sentences from a database of billions of unlabed sentences crawled from the web. |
| Outcome: | The proposed method improves on standard text classification benchmarks by 2.6% and knowledge distillation by few shots. |
AmericasNLI: Evaluating Zero-shot Natural Language Understanding of Pretrained Multilingual Models in Truly Low-resource Languages (2022.acl-long)
Copied to clipboard
Abteen Ebrahimi, Manuel Mager, Arturo Oncevay, Vishrav Chaudhary, Luis Chiruzzo, Angela Fan, John Ortega, Ricardo Ramos, Annette Rios, Ivan Vladimir Meza Ruiz, Gustavo Giménez-Lugo, Elisabeth Mager, Graham Neubig, Alexis Palmer, Rolando Coto-Solano, Thang Vu, Katharina Kann
| Challenge: | Pretrained multilingual models can perform cross-lingual transfer in a zero-shot setting, even for unseen languages. |
| Approach: | They propose to extend XNLI to 10 indigenous languages of the Americas and test multiple zero-shot and translation-based approaches. |
| Outcome: | The proposed model can perform cross-lingual transfer in a zero-shot setting even for languages unseen during pretraining. |
Unsupervised Cross-lingual Representation Learning at Scale (2020.acl-main)
Copied to clipboard
Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, Veselin Stoyanov
| Challenge: | Pretraining multilingual language models at scale leads to performance gains for cross-lingual transfer tasks. |
| Approach: | They present a transformer-based multilingual masked language model pre-trained on 100 languages . they show that pretraining multilingual models at scale leads to significant performance gains . |
| Outcome: | The proposed model outperforms multilingual BERT (mBERT) on cross-lingual benchmarks. |
OCR Improves Machine Translation for Low-Resource Languages (2022.findings-acl)
Copied to clipboard
| Challenge: | Despite many recent successes, Machine Translation still lacks support or fails to achieve good performance for most low-resource languages. |
| Approach: | They propose a benchmark to evaluate OCR systems on low-resource languages and low- resource scripts. |
| Outcome: | The proposed benchmark evaluates state-of-the-art OCR systems on low-resource languages and low-rural scripts. |
An Exploratory Study on Multilingual Quality Estimation (2020.aacl-main)
Copied to clipboard
Shuo Sun, Marina Fomicheva, Frédéric Blain, Vishrav Chaudhary, Ahmed El-Kishky, Adithya Renduchintala, Francisco Guzmán, Lucia Specia
| Challenge: | Existing approaches to predict the quality of machine translation use language-specific models, but they lack labelled data for each language pair. |
| Approach: | They propose to use scores from translation models to estimate quality of machine translations by predicting the quality of a translation at test time. |
| Outcome: | The proposed models outperform single-language models in less balanced quality label distributions and low-resource settings. |
Multilingual Translation from Denoising Pre-Training (2021.findings-acl)
Copied to clipboard
Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan
| Challenge: | Recent work shows potential of training one model for multilingual machine translation . but little has been explored on the potential to combine denoising pretraining with multilingual translation in a single model. |
| Approach: | They propose to combine denoising pretraining with multilingual machine translation in a single model. |
| Outcome: | The proposed model improves over models trained from scratch and bilingually for translation into English. |
Adapting High-resource NMT Models to Translate Low-resource Related Languages without Parallel Data (2021.acl-long)
Copied to clipboard
Wei-Jen Ko, Ahmed El-Kishky, Adithya Renduchintala, Vishrav Chaudhary, Naman Goyal, Francisco Guzmán, Pascale Fung, Philipp Koehn, Mona Diab
| Challenge: | linguistic overlap between low-resource languages and high-resourced languages is a major obstacle for training high-quality machine translation systems. |
| Approach: | They exploit linguistic overlap to facilitate translation to and from low-resource languages . they use monolingual data and parallel data in related high-resourced languages based on their method . |
| Outcome: | The proposed method significantly improves translation into low-resource language compared to baselines on 7 languages from three different language families. |
A Glitch in the Matrix? Locating and Detecting Language Model Grounding with Fakepedia (2024.acl-long)
Copied to clipboard
Giovanni Monea, Maxime Peyrard, Martin Josifoski, Vishrav Chaudhary, Jason Eisner, Emre Kiciman, Hamid Palangi, Barun Patra, Robert West
| Challenge: | Large language models (LLMs) have an impressive ability to draw on novel information supplied in their context, yet the mechanisms underlying contextual grounding remain unknown. |
| Approach: | They propose a method to study grounding abilities using a counterfactual dataset constructed to clash with a model's parametric knowledge using Fakepedia. |
| Outcome: | The proposed method evaluates grounding abilities when the internal parametric knowledge clashes with the contextual information. |
Data Selection Curriculum for Neural Machine Translation (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Neural Machine Translation models are typically trained on heterogeneous data that are concatenated and randomly shuffled. |
| Approach: | They propose a two-stage curriculum training framework where a NMT model is fine-tuned on subsets of data, selected by deterministic scoring and online scoring. |
| Outcome: | The proposed framework improves on six language pairs comprising low- and high-resource languages and shows up to +2.2 BLEU improvement and faster convergence. |
A Practical Analysis of Human Alignment with *PO (2025.findings-naacl)
Copied to clipboard
| Challenge: | Prior research focused on identifying the best-performing method to varying hyperparameters . prior research focused primarily on a grid search, which can be impractical for general practitioners . |
| Approach: | They propose a preference optimization method that is more stable across hyperparameters and reduces the average response length. |
| Outcome: | The proposed method increases likelihood of achieving better results through various metrics, such as KL divergence and response length. |
Language Model Decoding as Likelihood–Utility Alignment (2023.findings-eacl)
Copied to clipboard
Martin Josifoski, Maxime Peyrard, Frano Rajič, Jiheng Wei, Debjit Paul, Valentin Hartmann, Barun Patra, Vishrav Chaudhary, Emre Kiciman, Boi Faltings
| Challenge: | Existing studies only compare decoding algorithms in narrow scenarios, and their findings do not generalize across tasks. |
| Approach: | They propose a taxonomy of misalignment mitigation strategies to provide a unifying view of decoding as a tool for alignment. |
| Outcome: | The proposed taxonomy combines likelihood and utility assumptions to provide general statements about decoding as a tool for alignment across tasks. |
Unsupervised Quality Estimation for Neural Machine Translation (2020.tacl-1)
Copied to clipboard
Marina Fomicheva, Shuo Sun, Lisa Yankovskaya, Frédéric Blain, Francisco Guzmán, Mark Fishel, Nikolaos Aletras, Vishrav Chaudhary, Lucia Specia
| Challenge: | Existing approaches require large amounts of expert annotated data, computation, and time for training. |
| Approach: | They propose an unsupervised approach to QE where no training is required . they use a dataset that enables work on both black-box and glass-box approaches . |
| Outcome: | The proposed approach rivals state-of-the-art supervised QE models in terms of correlation with human judgments of quality. |
Scaling Laws for Multilingual Language Models (2025.findings-acl)
Copied to clipboard
Yifei He, Alon Benhaim, Barun Patra, Praneetha Vaddamanu, Sanchit Ahuja, Parul Chopra, Vishrav Chaudhary, Han Zhao, Xia Song
| Challenge: | Existing scaling laws for language models are limited to a limited number of languages, but they can be applied to arbitrary number of different languages. |
| Approach: | They propose a scaling law for general-purpose decoder-only language models trained on multilingual data that shifts focus from individual languages to language families. |
| Outcome: | The proposed scaling law can be applied to models trained on multilingual data . it can be used to predict performance across multiple languages and models . |
CCNet: Extracting High Quality Monolingual Datasets from Web Crawl Data (2020.lrec-1)
Copied to clipboard
Guillaume Wenzek, Marie-Anne Lachaux, Alexis Conneau, Vishrav Chaudhary, Francisco Guzmán, Armand Joulin, Edouard Grave
| Challenge: | Pre-training text representations have led to significant improvements in many areas of natural language processing. |
| Approach: | They propose a pipeline to extract monolingual datasets from Common Crawl . pipeline follows data processing introduced in fastText that deduplicates documents . |
| Outcome: | The proposed pipeline performs standard document deduplication and language identification similar to the pipeline introduced in fastText and a filtering step to select documents close to high quality corpora like Wikipedia. |
DUBLIN: Visual Document Understanding By Language-Image Network (2023.emnlp-industry)
Copied to clipboard
Kriti Aggarwal, Aditi Khandelwal, Kumar Tanmay, Owais Khan Mohammed, Qiang Liu, Monojit Choudhury, Hardik Chauhan, Subhojit Som, Vishrav Chaudhary, Saurabh Tiwary
| Challenge: | DUBLIN is a pixel-based visual document understanding model that does not rely on OCR. |
| Approach: | They propose a pixel-based visual document understanding model that does not rely on OCR. |
| Outcome: | The proposed model performs on extractive tasks such as DocVQA, InfoVQA and AI2D, and strong performance on abstraction datasets such as VisualMRC and text captioning. |
Beyond English-Centric Bitexts for Better Multilingual Language Representation Learning (2023.acl-long)
Copied to clipboard
Barun Patra, Saksham Singhal, Shaohan Huang, Zewen Chi, Li Dong, Furu Wei, Vishrav Chaudhary, Xia Song
| Challenge: | XY-LENT: X-Y bitext enhanced Language ENcodings achieves state-of-the-art performance over 5 cross-lingual tasks within all model size bands. |
| Approach: | They propose a method for building multilingual representation models that are competitive with existing models and more parameter efficient. |
| Outcome: | The proposed model outperforms XLM-R XXL and is 5x and 6x smaller respectively. |
A Length-Extrapolatable Transformer (2023.acl-long)
Copied to clipboard
Yutao Sun, Li Dong, Barun Patra, Shuming Ma, Shaohan Huang, Alon Benhaim, Vishrav Chaudhary, Xia Song, Furu Wei
| Challenge: | Existing Transformers can only deal with the in-distribution size of inputs. |
| Approach: | They propose a relative position embedding to explicitly maximize attention resolution . they also use blockwise causal attention during inference for better resolution a . |
| Outcome: | The proposed model achieves strong performance in interpolation and extrapolation settings. |
Everything you need to know about Multilingual LLMs: Towards fair, performant and reliable models for languages of the world (2023.acl-tutorials)
Copied to clipboard
| Challenge: | Responsible AI issues such as fairness, bias and toxicity will be discussed in this tutorial . |
| Approach: | This tutorial will describe various aspects of scaling up language technologies to many of the world’s languages by describing the latest research in Massively Multilingual Language Models (MMLMs). |
| Outcome: | This tutorial will cover various aspects of scaling up language technologies to many of the world's languages by describing the latest research in multilingual models. |
Performance and Risk Trade-offs for Multi-word Text Prediction at Scale (2023.findings-eacl)
Copied to clipboard
Aniket Vashishtha, S Sai Prasad, Payal Bajaj, Vishrav Chaudhary, Kate Cook, Sandipan Dandapat, Sunayana Sitaram, Monojit Choudhury
| Challenge: | Large Language Models (LLMs) generate ethically inappropriate texts even for seemingly innocuous contexts. |
| Approach: | They propose to use large language models to detect and filter toxic content in text prediction tasks by evaluating their toxicity detection approaches against a manually crafted CheckList of harms. |
| Outcome: | The proposed methods are compared against a checklist of harms targeted at different groups and different levels of severity in English. |
Quality Estimation without Human-labeled Data (2021.eacl-main)
Copied to clipboard
Yi-Lin Tuan, Ahmed El-Kishky, Adithya Renduchintala, Vishrav Chaudhary, Francisco Guzmán, Lucia Specia
| Challenge: | Quality estimation aims to measure the quality of translated content without access to a reference translation. |
| Approach: | They propose a method that uses synthetic training data to train supervised quality estimation models. |
| Outcome: | The proposed model outperforms models trained on human-annotated data for sentence and word-level prediction. |