We present the IndicNLP corpus, a large-scale, general-domain corpus containing 2.7 billion words for 10 Indian languages from two language families. We share pre-trained word embeddings trained on these corpora. We create news article category classification datasets for 9 languages to evaluate the embeddings. We show that the IndicNLP embeddings significantly outperform publicly available pre-trained embedding on multiple evaluation tasks. We hope that the availability of the corpus will accelerate Indic NLP research. The resources are available at https://github.com/ai4bharat-indicnlp/indicnlp_corpus.
The paper presents the submission of the team indicnlp@kgp to the EACL 2021 shared task "Offensive Language Identification in Dravidian Languages." The task aimed to classify different offensive content types in 3 code-mixed Dravidian language datasets. The work leverages existing state of the art approaches in text classification by incorporating additional data and transfer learning on pre-trained models. Our final submission is an ensemble of an AWD-LSTM based model along with 2 different transformer model architectures based on BERT and RoBERTa. We achieved weighted-average F1 scores of 0.…
Harbin Institute of Technology（哈尔滨工业大学）；Baidu Inc.（百度公司）
In this paper, we provide a bilingual parallel human-to-human recommendation dialog dataset (DuRecDial 2.0) to enable researchers to explore a challenging task of multilingual and cross-lingual conversational recommendation. The difference between DuRecDial 2.0 and existing conversational recommendation datasets is that the data item (Profile, Goal, Knowledge, Context, Response) in DuRecDial 2.0 is annotated in two languages, both English and Chinese, while other datasets are built with the setting of a single language. We collect 8.2k dialogs aligned across English and Chinese languages (16.5…
Centre for Translation Studies, University of Leeds（利兹大学翻译研究中心）
Metrics for measuring the comparability of corpora or texts need to be developed and evaluated systematically. Applications based on a corpus, such as training Statistical MT systems in specialised narrow domains, require finding a reasonable balance between the size of the corpus and its consistency, with controlled and benchmarked levels of comparability for any newly added sections. In this article we propose a method that can meta-evaluate comparability metrics by calculating monolingual comparability scores separately on the 'source' and 'target' sides of parallel corpora. The range of sc…
Transformer language models (TLMs) are critical for most NLP tasks, but they are difficult to create for low-resource languages because of how much pretraining data they require. In this work, we investigate two techniques for training monolingual TLMs in a low-resource setting: greatly reducing TLM size, and complementing the masked language modeling objective with two linguistically rich supervised tasks (part-of-speech tagging and dependency parsing). Results from 7 diverse languages indicate that our model, MicroBERT, is able to produce marked improvements in downstream task evaluations re…
Iran University of Science and Technology（伊朗科技大学）；Tehran Institute for Advanced Studies, Khatam University（哈塔姆大学德黑兰高等研究院）
Several studies have explored various advantages of multilingual pre-trained models (such as multilingual BERT) in capturing shared linguistic knowledge. However, less attention has been paid to their limitations. In this paper, we investigate the multilingual BERT for two known issues of the monolingual models: anisotropic embedding space and outlier dimensions. We show that, unlike its monolingual counterpart, the multilingual BERT model exhibits no outlier dimension in its representations while it has a highly anisotropic space. There are a few dimensions in the monolingual BERT with high c…
Indian Institute of Technology Madras（印度理工学院马德拉斯分校）；L3Cube Pune（L3Cube浦那）
The monolingual Hindi BERT models currently available on the model hub do not perform better than the multi-lingual models on downstream tasks. We present L3Cube-HindBERT, a Hindi BERT model pre-trained on Hindi monolingual corpus. Further, since Indic languages, Hindi and Marathi share the Devanagari script, we train a single model for both languages. We release DevBERT, a Devanagari BERT model trained on both Marathi and Hindi monolingual datasets. We evaluate these models on downstream Hindi and Marathi text classification and named entity recognition tasks. The HindBERT and DevBERT-based m…
Kanari AI（卡纳里人工智能公司）；Johns Hopkins University（约翰斯·霍普金斯大学）；Qatar Computing Research Institute, HBKU（哈马德·本·哈利法大学卡塔尔计算研究所）
The success in designing Code-Switching (CS) ASR often depends on the availability of the transcribed CS resources. Such dependency harms the development of ASR in low-resourced languages such as Bengali and Hindi. In this paper, we exploit the transfer learning approach to design End-to-End (E2E) CS ASR systems for the two low-resourced language pairs using different monolingual speech data and a small set of noisy CS data. We trained the CS-ASR, following two steps: (i) building a robust bilingual ASR system using a convolution-augmented transformer (Conformer) based acoustic model and n-gra…
Language Technology Lab, University of Cambridge（剑桥大学语言技术实验室）；Faculty of Industrial Engineering and Management, Technion, IIT（以色列理工学院工业工程与管理学院）
Performance in cross-lingual NLP tasks is impacted by the (dis)similarity of languages at hand: e.g., previous work has suggested there is a connection between the expected success of bilingual lexicon induction (BLI) and the assumption of (approximate) isomorphism between monolingual embedding spaces. In this work we present a large-scale study focused on the correlations between monolingual embedding space similarity and task performance, covering thousands of language pairs and four different tasks: BLI, parsing, POS tagging and MT. We hypothesize that statistics of the spectrum of each mon…
Do Multilingual Large Language Models Mitigate Stereotype Bias?
Conversational AI and Social Analytics (CAISA) Lab, University of Bonn（波恩大学）；Fraunhofer Institute for Intelligent Analysis and Information Systems (IAIS)（弗劳恩霍夫智能分析与信息系统研究所）；Lamarr Institute for Machine Learning and Artificial Intelligence（拉马尔机器学习和人工智能研究所）
Training multilingual automatic speech recognition (ASR) systems is challenging because acoustic and lexical information is typically language specific. Training multilingual system for Indic languages is even more tougher due to lack of open source datasets and results on different approaches. We compare the performance of end to end multilingual speech recognition system to the performance of monolingual models conditioned on language identification (LID). The decoding information from a multilingual model is used for language identification and then combined with monolingual models to get a…
Pune Institute of Computer Technology（浦那计算机技术学院）；Indian Institute of Technology Madras（印度理工学院马德拉斯分校）
The Marathi language is one of the prominent languages used in India. It is predominantly spoken by the people of Maharashtra. Over the past decade, the usage of language on online platforms has tremendously increased. However, research on Natural Language Processing (NLP) approaches for Marathi text has not received much attention. Marathi is a morphologically rich language and uses a variant of the Devanagari script in the written form. This works aims to provide a comprehensive overview of available resources and models for Marathi text classification. We evaluate CNN, LSTM, ULMFiT, and BER…
University of Cambridge（剑桥大学）；Université de Montréal（蒙特利尔大学）；Harvey Mudd College（哈维·穆德学院）
Neural language models learn word representations, or embeddings, that capture rich linguistic and conceptual information. Here we investigate the embeddings learned by neural machine translation models, a recently-developed class of neural language model. We show that embeddings from translation models outperform those learned by monolingual models at tasks that require knowledge of both conceptual similarity and lexical-syntactic role. We further show that these effects hold when translating from both English to French and English to German, and argue that the desirable properties of transla…
Pune Institute of Computer Technology（浦那计算机技术学院）；L3Cube Pune（浦那L3Cube）；Indian Institute of Technology Madras（印度理工学院马德拉斯分校）
Transformers are the most eminent architectures used for a vast range of Natural Language Processing tasks. These models are pre-trained over a large text corpus and are meant to serve state-of-the-art results over tasks like text classification. In this work, we conduct a comparative study between monolingual and multilingual BERT models. We focus on the Marathi language and evaluate the models on the datasets for hate speech detection, sentiment analysis and simple text classification in Marathi. We use standard multilingual models such as mBERT, indicBERT and xlm-RoBERTa and compare with Ma…
Neural networks trained on natural language processing tasks capture syntax even though it is not provided as a supervision signal. This indicates that syntactic analysis is essential to the understating of language in artificial intelligence systems. This overview paper covers approaches of evaluating the amount of syntactic information included in the representations of words for different neural network architectures. We mainly summarize re-search on English monolingual data on language modeling tasks and multilingual data for neural machine translation systems and multilingual language mod…
We present a CLSRIL-23, a self supervised learning based audio pre-trained model which learns cross lingual speech representations from raw audio across 23 Indic languages. It is built on top of wav2vec 2.0 which is solved by training a contrastive task over masked latent speech representations and jointly learns the quantization of latents shared across all languages. We compare the language wise loss during pretraining to compare effects of monolingual and multilingual pretraining. Performance on some downstream fine-tuning tasks for speech recognition is also compared and our experiments sh…
Shanghai Jiao Tong University（上海交通大学）；University of Southern California（南加州大学）
In this paper, we report our discovery on named entity distribution in a general word embedding space, which helps an open definition on multilingual named entity definition rather than previous closed and constraint definition on named entities through a named entity dictionary, which is usually derived from human labor and replies on schedule update. Our initial visualization of monolingual word embeddings indicates named entities tend to gather together despite of named entity types and language difference, which enable us to model all named entities using a specific geometric structure ins…
Recent work indicated that pretrained language models (PLMs) such as BERT and RoBERTa can be transformed into effective sentence and word encoders even via simple self-supervised techniques. Inspired by this line of work, in this paper we propose a fully unsupervised approach to improving word-in-context (WiC) representations in PLMs, achieved via a simple and efficient WiC-targeted fine-tuning procedure: MirrorWiC. The proposed method leverages only raw texts sampled from Wikipedia, assuming no sense-annotated data, and learns context-aware word representations within a standard contrastive l…
University of Colorado Boulder（科罗拉多大学博尔德分校）；University of Michigan Ann Arbor（密歇根大学安娜堡分校）；San Diego State University（圣地亚哥州立大学）
This paper describes our system for the SemEval2022 task of matching dictionary glosses to word embeddings. We focus on the Reverse Dictionary Track of the competition, which maps multilingual glosses to reconstructed vector representations. More specifically, models convert the input of sentences to three types of embeddings: SGNS, Char, and Electra. We propose several experiments for applying neural network cells, general multilingual and multitask structures, and language-agnostic tricks to the task. We also provide comparisons over different types of word embeddings and ablation studies to…
IIT Madras（印度理工学院马德拉斯分校）；Columbia University（哥伦比亚大学）；National Institute of Information and Communications Technology（信息通信研究机构）；Microsoft（微软）；University of Edinburgh（爱丁堡大学）；AI4Bharat
Natural Language Generation (NLG) for non-English languages is hampered by the scarcity of datasets in these languages. In this paper, we present the IndicNLG Benchmark, a collection of datasets for benchmarking NLG for 11 Indic languages. We focus on five diverse tasks, namely, biography generation using Wikipedia infoboxes, news headline generation, sentence summarization, paraphrase generation and, question generation. We describe the created datasets and use them to benchmark the performance of several monolingual and multilingual baselines that leverage pre-trained sequence-to-sequence mo…
The University of Tokyo（东京大学）；Institute of Industrial Science（产业科学研究所）；National Institute of Informatics（国立信息学研究所）
Neural network methods exhibit strong performance only in a few resource-rich domains. Practitioners, therefore, employ domain adaptation from resource-rich domains that are, in most cases, distant from the target domain. Domain adaptation between distant domains (e.g., movie subtitles and research papers), however, cannot be performed effectively due to mismatches in vocabulary; it will encounter many domain-specific words (e.g., "angstrom") and words whose meanings shift across domains(e.g., "conductor"). In this study, aiming to solve these vocabulary mismatches in domain adaptation for neu…
Language Exposure Matters: Distribution and Order in Multilingual BabyLM Pre-training
School of Computation, Information and Technology, TU Munich（慕尼黑工业大学计算、信息与技术学院）；Munich Center for Machine Learning（慕尼黑机器学习中心）；Munich Data Science Institute（慕尼黑数据科学研究所）
We explore the impact of leveraging the relatedness of languages that belong to the same family in NLP models using multilingual fine-tuning. We hypothesize and validate that multilingual fine-tuning of pre-trained language models can yield better performance on downstream NLP applications, compared to models fine-tuned on individual languages. A first of its kind detailed study is presented to track performance change as languages are added to a base language in a graded and greedy (in the sense of best boost of performance) manner; which reveals that careful selection of subset of related la…
Abusive content detection in spoken text can be addressed by performing Automatic Speech Recognition (ASR) and leveraging advancements in natural language processing. However, ASR models introduce latency and often perform sub-optimally for profane words as they are underrepresented in training corpora and not spoken clearly or completely. Exploration of this problem entirely in the audio domain has largely been limited by the lack of audio datasets. Building on these challenges, we propose ADIMA, a novel, linguistically diverse, ethically sourced, expert annotated and well-balanced multilingu…
