Language model pre-training, such as BERT, has achieved remarkable results in many NLP tasks. However, it is unclear why the pre-training-then-fine-tuning paradigm can improve performance and generalization capability across different tasks. In this paper, we propose to visualize loss landscapes and optimization trajectories of fine-tuning BERT on specific datasets. First, we find that pre-training reaches a good initial point across downstream tasks, which leads to wider optima and easier optimization compared with training from scratch. We also demonstrate that the fine-tuning procedure is robust to overfitting, even though BERT is highly over-parameterized for downstream tasks. Second, the visualization results indicate that fine-tuning BERT tends to generalize better because of the flat and wide optima, and the consistency between the training loss surface and the generalization error surface. Third, the lower layers of BERT are more invariant during fine-tuning, which suggests that the layers that are close to input learn more transferable representations of language.
Transformer architectures show significant promise for natural language processing. Given that a single pretrained model can be fine-tuned to perform well on many different tasks, these networks appear to extract generally useful linguistic features. A natural question is how such networks represent this information internally. This paper describes qualitative and quantitative investigations of one particularly effective model, BERT. At a high level, linguistic features seem to be represented in separate semantic and syntactic subspaces. We find evidence of a fine-grained geometric representat…
Microsoft Research Asia（微软亚洲研究院）；Peking University（北京大学）；University of Michigan（密歇根大学）
We study the problem of visualizing large-scale and high-dimensional data in a low-dimensional (typically 2D or 3D) space. Much success has been reported recently by techniques that first compute a similarity structure of the data points and then project them into a low-dimensional space with the structure preserved. These two steps suffer from considerable computational costs, preventing the state-of-the-art methods such as the t-SNE from scaling to large-scale and high-dimensional data (e.g., millions of data points and hundreds of dimensions). We propose the LargeVis, a technique that first…
Large pre-trained language models such as BERT have shown their effectiveness in various natural language processing tasks. However, the huge parameter size makes them difficult to be deployed in real-time applications that require quick inference with limited resources. Existing methods compress BERT into small models while such compression is task-independent, i.e., the same compressed BERT for all different downstream tasks. Motivated by the necessity and benefits of task-oriented BERT compression, we propose a novel compression method, AdaBERT, that leverages differentiable Neural Architec…
We compare self-supervised representation learning algorithms which either explicitly quantize the audio data or learn representations without quantization. We find the former to be more accurate since it builds a good vocabulary of the data through vq-wav2vec [1] to enable learning of effective representations in subsequent BERT training. Different to previous work, we directly fine-tune the pre-trained BERT models on transcribed speech using a Connectionist Temporal Classification (CTC) loss instead of feeding the representations into a task-specific model. We also propose a BERT-style model…
School of Data and Computer Science（数据与计算机学院）；Sun Yat-sen University（中山大学）；Guangzhou, China（中国广州（注：此为地址信息，非机构，根据规则应忽略，此处原文本可能存在排版问题，正确机构应为前两项））
Aspect based sentiment analysis aims to identify the sentimental tendency towards a given aspect in text. Fine-tuning of pretrained BERT performs excellent on this task and achieves state-of-the-art performances. Existing BERT-based works only utilize the last output layer of BERT and ignore the semantic knowledge in the intermediate layers. This paper explores the potential of utilizing BERT intermediate layers to enhance the performance of fine-tuning of BERT. To the best of our knowledge, no existing work has been done on this research. To show the generality, we also apply this approach to…
Stanford Center for Professional Development（斯坦福专业发展中心）
Question Answering with NLP has progressed through the evolution of advanced model architectures like BERT and BiDAF and earlier word, character, and context-based embeddings. As BERT has leapfrogged the accuracy of models, an element of the next frontier can be the introduction of deep networks and an effective way to train them. In this context, I explored the effectiveness of deep networks focussing on the model encoder layer of BiDAF. BiDAF with its heterogeneous layers provides the opportunity not only to explore the effectiveness of deep networks but also to evaluate whether the refineme…
Pretrained contextual representation models (Peters et al., 2018; Devlin et al., 2018) have pushed forward the state-of-the-art on many NLP tasks. A new release of BERT (Devlin, 2018) includes a model simultaneously pretrained on 104 languages with impressive performance for zero-shot cross-lingual transfer on a natural language inference task. This paper explores the broader cross-lingual potential of mBERT (multilingual) as a zero shot language transfer model on 5 NLP tasks covering a total of 39 languages from various language families: NLI, document classification, NER, POS tagging, and de…
ASAPP Inc.（ASAPP公司）；Stanford University（斯坦福大学）；Penn State University（宾夕法尼亚州立大学）；Cornell University（康奈尔大学）
This paper is a study of fine-tuning of BERT contextual representations, with focus on commonly observed instabilities in few-sample scenarios. We identify several factors that cause this instability: the common use of a non-standard optimization method with biased gradient estimation; the limited applicability of significant parts of the BERT network for down-stream tasks; and the prevalent practice of using a pre-determined, and small number of training iterations. We empirically test the impact of these factors, and identify alternative practices that resolve the commonly observed instabili…
School of Foreign Languages and Cultures, Xiamen University（厦门大学外文学院）；DeepAI Tech Co., Ltd（深湃信息科技（深圳）有限公司）
The automatic classification is a process of automatically assigning text documents to predefined categories. An accurate automatic patent classifier is crucial to patent inventors and patent examiners in terms of intellectual property protection, patent management, and patent information retrieval. We present BERT-CNN, a hierarchical patent classifier based on pre-trained language model by training the national patent application documents collected from the State Information Center, China. The experimental results show that BERT-CNN achieves 84.3% accuracy, which is far better than the two c…
Contextualized word embeddings have demonstrated state-of-the-art performance in various natural language processing tasks including those that concern historical semantic change. However, language models such as BERT was trained primarily on contemporary corpus data. To investigate whether training on historical corpus data improves diachronic semantic analysis, we present a pre-trained BERT-based language model, HistBERT, trained on the balanced Corpus of Historical American English. We examine the effectiveness of our approach by comparing the performance of the original BERT and that of Hi…
Center for Artificial Intelligence Research (CAiRE)（人工智能研究中心）；The Hong Kong University of Science and Technology（香港科技大学）
Nowadays, offensive content in social media has become a serious problem, and automatically detecting offensive language is an essential task. In this paper, we build an offensive language detection system, which combines multi-task learning with BERT-based models. Using a pre-trained language model such as BERT, we can effectively learn the representations for noisy text in social media. Besides, to boost the performance of offensive language detection, we leverage the supervision signals from other related tasks. In the OffensEval-2020 competition, our model achieves 91.51% F1 score in Engli…
User reviews have an essential role in the success of the developed mobile apps. User reviews in the textual form are unstructured data, creating a very high complexity when processed for sentiment analysis. Previous approaches that have been used often ignore the context of reviews. In addition, the relatively small data makes the model overfitting. A new approach, BERT, has been introduced as a transfer learning model with a pre-trained model that has previously been trained to have a better context representation. This study examines the effectiveness of fine-tuning BERT for sentiment analy…
Active learning has been shown to be an effective way to alleviate some of the effort required in utilising large collections of unlabelled data for machine learning tasks without needing to fully label them. The representation mechanism used to represent text documents when performing active learning, however, has a significant influence on how effective the process will be. While simple vector representations such as bag-of-words and embedding-based representations based on techniques such as word2vec have been shown to be an effective way to represent documents during active learning, the e…
Indian Institute of Technology Madras（印度理工学院马德拉斯分校）；L3Cube
We present L3Cube-MahaCorpus a Marathi monolingual data set scraped from different internet sources. We expand the existing Marathi monolingual corpus with 24.8M sentences and 289M tokens. We further present, MahaBERT, MahaAlBERT, and MahaRoBerta all BERT-based masked language models, and MahaFT, the fast text word embeddings both trained on full Marathi corpus with 752M tokens. We show the effectiveness of these resources on downstream Marathi sentiment analysis, text classification, and named entity recognition (NER) tasks. We also release MahaGPT, a generative Marathi GPT model trained on M…
We introduce HUBERT which combines the structured-representational power of Tensor-Product Representations (TPRs) and BERT, a pre-trained bidirectional Transformer language model. We show that there is shared structure between different NLP datasets that HUBERT, but not BERT, is able to learn and leverage. We validate the effectiveness of our model on the GLUE benchmark and HANS dataset. Our experiment results show that untangling data-specific semantics from general language structure is key for better transfer among NLP tasks.
CSIRO Data61（澳大利亚联邦科学与工业研究组织数据61）；University of Sydney（悉尼大学）；Harrison ai（哈里森人工智能）
Recent studies on domain-specific BERT models show that effectiveness on downstream tasks can be improved when models are pretrained on in-domain data. Often, the pretraining data used in these models are selected based on their subject matter, e.g., biology or computer science. Given the range of applications using social media text, and its unique language variety, we pretrain two models on tweets and forum text respectively, and empirically demonstrate the effectiveness of these two resources. In addition, we investigate how similarity measures can be used to nominate in-domain pretraining …
We aim to highlight an interesting trend to contribute to the ongoing debate around advances within legal Natural Language Processing. Recently, the focus for most legal text classification tasks has shifted towards large pre-trained deep learning models such as BERT. In this paper, we show that a more traditional approach based on Support Vector Machine classifiers reaches surprisingly competitive performance with BERT-based models on the classification tasks in the LexGLUE benchmark. We also highlight that error reduction obtained by using specialised BERT-based models over baselines is noti…
Peking University（北京大学）；Microsoft Research（微软研究院）；Microsoft Research Asia（微软亚洲研究院）
Pre-trained contextual representations (e.g., BERT) have become the foundation to achieve state-of-the-art results on many NLP tasks. However, large-scale pre-training is computationally expensive. ELECTRA, an early attempt to accelerate pre-training, trains a discriminative model that predicts whether each input token was replaced by a generator. Our studies reveal that ELECTRA's success is mainly due to its reduced complexity of the pre-training task: the binary classification (replaced token detection) is more efficient to learn than the generation task (masked language modeling). However, …
Harbin Institute of Technology（哈尔滨工业大学）；iFLYTEK Research（科大讯飞研究院）；iFLYTEK AI Research (Hebei)（科大讯飞人工智能研究院（河北））
Bidirectional Encoder Representations from Transformers (BERT) has shown marvelous improvements across various NLP tasks, and its consecutive variants have been proposed to further improve the performance of the pre-trained language models. In this paper, we aim to first introduce the whole word masking (wwm) strategy for Chinese BERT, along with a series of Chinese pre-trained language models. Then we also propose a simple but effective model called MacBERT, which improves upon RoBERTa in several ways. Especially, we propose a new masking strategy called MLM as correction (Mac). To demonstrat…
Tribhuvan University（特里布万大学）；Institute of Engineering, Tribhuvan University（特里布万大学工程学院）
Sentiment classification is an important process in understanding people's perception towards a product, service, or topic. Many natural language processing models have been proposed to solve the sentiment classification problem. However, most of them have focused on binary sentiment classification. In this paper, we use a promising deep learning model called BERT to solve the fine-grained sentiment classification task. Experiments show that our model outperforms other popular models for this task without sophisticated architecture. We also demonstrate the effectiveness of transfer learning in…
Université de Montréal（蒙特利尔大学）；School of Information, Renmin University of China（中国人民大学信息学院）；Mila - Québec Artificial Intelligence Institute（米拉-魁北克人工智能研究所）
Sentence ordering aims to arrange the sentences of a given text in the correct order. Recent work frames it as a ranking problem and applies deep neural networks to it. In this work, we propose a new method, named BERT4SO, by fine-tuning BERT for sentence ordering. We concatenate all sentences and compute their representations by using multiple special tokens and carefully designed segment (interval) embeddings. The tokens across multiple sentences can attend to each other which greatly enhances their interactions. We also propose a margin-based listwise ranking loss based on ListMLE to facili…
The effectiveness of the BERT model on multiple linguistic tasks has been well documented. On the other hand, its potentials for narrow and specific domains such as Legal, have not been fully explored. In this paper, we examine how BERT can be used in the Arabic legal domain and try customizing this language model for several downstream tasks using several different domain-relevant training and testing datasets to train BERT from scratch. We introduce the AraLegal-BERT, a bidirectional encoder Transformer-based model that have been thoroughly tested and carefully optimized with the goal to amp…
Nomura Research Institute, Ltd.（野村综合研究所株式会社）；Retrieva, Inc.（Retrieva公司）；Tokyo Metropolitan University（东京都立大学）
In recent years, pre-trained models have been extensively studied, and several downstream tasks have benefited from their utilization. In this study, we verify the effectiveness of two methods that incorporate a BERT-based pre-trained model developed by Cui et al. (2020) into an encoder-decoder model on Chinese grammatical error correction tasks. We also analyze the error type and conclude that sentence-level errors are yet to be addressed.
Enhancing machine capabilities to answer questions has been a topic of considerable focus in recent years of NLP research. Language models like Embeddings from Language Models (ELMo)[1] and Bidirectional Encoder Representations from Transformers (BERT) [2] have been very successful in developing general purpose language models that can be optimized for a large number of downstream language tasks. In this work, we focused on augmenting the pre-trained BERT language model with different output neural net architectures and compared their performance on question answering task posed by the Stanfor…
School of Biomedical Informatics, The University of Texas Health Science Center at Houston（德克萨斯大学休斯顿健康科学中心生物医学信息学院）
Developing high-performance entity normalization algorithms that can alleviate the term variation problem is of great interest to the biomedical community. Although deep learning-based methods have been successfully applied to biomedical entity normalization, they often depend on traditional context-independent word embeddings. Bidirectional Encoder Representations from Transformers (BERT), BERT for Biomedical Text Mining (BioBERT) and BERT for Clinical Text Mining (ClinicalBERT) were recently introduced to pre-train contextualized word representation models using bidirectional Transformers, a…
Leveraging large amounts of unlabeled data using Transformer-like architectures, like BERT, has gained popularity in recent times owing to their effectiveness in learning general representations that can then be further fine-tuned for downstream tasks to much success. However, training these models can be costly both from an economic and environmental standpoint. In this work, we investigate how to effectively use unlabeled data: by exploring the task-specific semi-supervised approach, Cross-View Training (CVT) and comparing it with task-agnostic BERT in multiple settings that include domain a…
Adversarial training (AT) is one of the most reliable methods for defending against adversarial attacks in machine learning. Variants of this method have been used as regularization mechanisms to achieve SOTA results on NLP benchmarks, and they have been found to be useful for transfer learning and continual learning. We search for the reasons for the effectiveness of AT by contrasting vanilla and adversarially fine-tuned BERT models. We identify partial preservation of BERT's syntactic abilities during fine-tuning as the key to the success of AT. We observe that adversarially fine-tuned model…
Peking University（北京大学）；Microsoft Research Asia（微软亚洲研究院）；Institute of Information Engineering, Chinese Academy of Sciences（中国科学院信息工程研究所）
Pre-trained language models like BERT achieve superior performances in various NLP tasks without explicit consideration of syntactic information. Meanwhile, syntactic information has been proved to be crucial for the success of NLP applications. However, how to incorporate the syntax trees effectively and efficiently into pre-trained Transformers is still unsettled. In this paper, we address this problem by proposing a novel framework named Syntax-BERT. This framework works in a plug-and-play mode and is applicable to an arbitrary pre-trained checkpoint based on Transformer architecture. Exper…
Pre-trained language model word representation, such as BERT, have been extremely successful in several Natural Language Processing tasks significantly improving on the state-of-the-art. This can largely be attributed to their ability to better capture semantic information contained within a sentence. Several tasks, however, can benefit from information available at a corpus level, such as Term Frequency-Inverse Document Frequency (TF-IDF). In this work we test the effectiveness of integrating this information with BERT on the task of identifying abuse on social media and show that integrating…
Huazhong University of Science and Technology（华中科技大学）；Huawei Technologies Co., Ltd.（华为技术有限公司）
The multilingual pre-trained language models (e.g, mBERT, XLM and XLM-R) have shown impressive performance on cross-lingual natural language understanding tasks. However, these models are computationally intensive and difficult to be deployed on resource-restricted devices. In this paper, we propose a simple yet effective distillation method (LightMBERT) for transferring the cross-lingual generalization ability of the multilingual BERT to a small student model. The experiment results empirically demonstrate the efficiency and effectiveness of LightMBERT, which is significantly better than the …
Acronym disambiguation (AD) task aims to find the correct expansions of an ambiguous ancronym in a given sentence. Although it is convenient to use acronyms, sometimes they could be difficult to understand. Identifying the appropriate expansions of an acronym is a practical task in natural language processing. Since few works have been done for AD in scientific field, we propose a binary classification model incorporating BERT and several training strategies including dynamic negative sample selection, task adaptive pretraining, adversarial training and pseudo labeling in this paper. Experimen…
Bosch Center for Artificial Intelligence（博世人工智能中心）；Carnegie Mellon University（卡内基梅隆大学）
We study the utility of the lexical translation model (IBM Model 1) for English text retrieval, in particular, its neural variants that are trained end-to-end. We use the neural Model1 as an aggregator layer applied to context-free or contextualized query/document embeddings. This new approach to design a neural ranking system has benefits for effectiveness, efficiency, and interpretability. Specifically, we show that adding an interpretable neural Model 1 layer on top of BERT-based contextualized embeddings (1) does not decrease accuracy and/or efficiency; and (2) may overcome the limitation …
We propose a practical scheme to train a single multilingual sequence labeling model that yields state of the art results and is small and fast enough to run on a single CPU. Starting from a public multilingual BERT checkpoint, our final model is 6x smaller and 27x faster, and has higher accuracy than a state-of-the-art multilingual baseline. We show that our model especially outperforms on low-resource languages, and works on codemixed input text without being explicitly trained on codemixed examples. We showcase the effectiveness of our method by reporting on part-of-speech tagging and morph…
Pretrained language models such as BERT, GPT have shown great effectiveness in language understanding. The auxiliary predictive tasks in existing pretraining approaches are mostly defined on tokens, thus may not be able to capture sentence-level semantics very well. To address this issue, we propose CERT: Contrastive self-supervised Encoder Representations from Transformers, which pretrains language representation models using contrastive self-supervised learning at the sentence level. CERT creates augmentations of original sentences using back-translation. Then it finetunes a pretrained langu…
University of Sheffield（谢菲尔德大学）；University of Durham（杜伦大学）
Neural language representation models such as BERT, pre-trained on large-scale unstructured corpora lack explicit grounding to real-world commonsense knowledge and are often unable to remember facts required for reasoning and inference. Natural Language Inference (NLI) is a challenging reasoning task that relies on common human understanding of language and real-world commonsense knowledge. We introduce a new model for NLI called External Knowledge Enhanced BERT (ExBERT), to enrich the contextual representation with real-world commonsense knowledge from external knowledge sources and enhance B…
This paper studies the performances and behaviors of BERT in ranking tasks. We explore several different ways to leverage the pre-trained BERT and fine-tune it on two ranking tasks: MS MARCO passage reranking and TREC Web Track ad hoc document ranking. Experimental results on MS MARCO demonstrate the strong effectiveness of BERT in question-answering focused passage ranking tasks, as well as the fact that BERT is a strong interaction-based seq2seq matching model. Experimental results on TREC show the gaps between the BERT pre-trained on surrounding contexts and the needs of ad hoc document ran…
Large pre-trained sentence encoders like BERT start a new chapter in natural language processing. A common practice to apply pre-trained BERT to sequence classification tasks (e.g., classification of sentences or sentence pairs) is by feeding the embedding of [CLS] token (in the last layer) to a task-specific classification layer, and then fine tune the model parameters of BERT and classifier jointly. In this paper, we conduct systematic analysis over several sequence classification datasets to examine the embedding values of [CLS] token before the fine tuning phase, and present the biased emb…
Recently, a simple combination of passage retrieval using off-the-shelf IR techniques and a BERT reader was found to be very effective for question answering directly on Wikipedia, yielding a large improvement over the previous state of the art on a standard benchmark dataset. In this paper, we present a data augmentation technique using distant supervision that exploits positive as well as negative examples. We apply a stage-wise approach to fine tuning BERT on multiple datasets, starting with data that is "furthest" from the test data and ending with the "closest". Experimental results show …
The University of Hong Kong（香港大学）；Shanghai University of Finance and Economics（上海财经大学）；Huawei Noah’s Ark Lab（华为诺亚方舟实验室）
By introducing a small set of additional parameters, a probe learns to solve specific linguistic tasks (e.g., dependency parsing) in a supervised manner using feature representations (e.g., contextualized embeddings). The effectiveness of such probing tasks is taken as evidence that the pre-trained model encodes linguistic knowledge. However, this approach of evaluating a language model is undermined by the uncertainty of the amount of knowledge that is learned by the probe itself. Complementary to those works, we propose a parameter-free probing technique for analyzing pre-trained language mo…
University of Information Technology（信息技术大学）；Vietnam National University, Ho Chi Minh City（胡志明市越南国家大学）
Text classification is a typical natural language processing or computational linguistics task with various interesting applications. As the number of users on social media platforms increases, data acceleration promotes emerging studies on Social Media Text Classification (SMTC) or social media text mining on these valuable resources. In contrast to English, Vietnamese, one of the low-resource languages, is still not concentrated on and exploited thoroughly. Inspired by the success of the GLUE, we introduce the Social Media Text Classification Evaluation (SMTCE) benchmark, as a collection of …
The pre-training models such as BERT have achieved great results in various natural language processing problems. However, a large number of parameters need significant amounts of memory and the consumption of inference time, which makes it difficult to deploy them on edge devices. In this work, we propose a knowledge distillation method LRC-BERT based on contrastive learning to fit the output of the intermediate layer from the angular distance aspect, which is not considered by the existing distillation methods. Furthermore, we introduce a gradient perturbation-based training architecture in …
Pre-trained language models, such as BERT, have achieved significant accuracy gain in many natural language processing tasks. Despite its effectiveness, the huge number of parameters makes training a BERT model computationally very challenging. In this paper, we propose an efficient multi-stage layerwise training (MSLT) approach to reduce the training time of BERT. We decompose the whole training process into several stages. The training is started from a small model with only a few encoder layers and we gradually increase the depth of the model by adding new encoder layers. At each stage, we …
BERT-LID: Leveraging BERT to Improve Spoken Language Identification
Beijing National Research Center for Information Science and Technology（北京信息科学与技术国家研究中心）；Tsinghua University（清华大学）；Computational Media Innovation Centre（计算媒体创新中心）；Victoria University of Wellington（惠灵顿维多利亚大学）；TAL Education（好未来教育集团）
The Hong Kong University of Science and Technology（香港科技大学）；Alibaba Group（阿里巴巴集团）
Named entity recognition (NER) models generally perform poorly when large training datasets are unavailable for low-resource domains. Recently, pre-training a large-scale language model has become a promising direction for coping with the data scarcity issue. However, the underlying discrepancies between the language modeling and NER task could limit the models' performance, and pre-training for the NER task has rarely been studied since the collected NER datasets are generally small or large but with low quality. In this paper, we construct a massive NER corpus with a relatively high quality,…
Nagoya University（名古屋大学）；Academia Sinica（中央研究院）；National Taiwan University of Science and Technology（台湾科技大学）
We propose a simple method for automatic speech recognition (ASR) by fine-tuning BERT, which is a language model (LM) trained on large-scale unlabeled text data and can generate rich contextual representations. Our assumption is that given a history context sequence, a powerful LM can narrow the range of possible choices and the speech signal can be used as a simple clue. Hence, comparing to conventional ASR systems that train a powerful acoustic model (AM) from scratch, we believe that speech recognition is possible by simply fine-tuning a BERT model. As an initial study, we demonstrate the e…
Recent work on enhancing BERT-based language representation models with knowledge graphs (KGs) and knowledge bases (KBs) has yielded promising results on multiple NLP tasks. State-of-the-art approaches typically integrate the original input sentences with KG triples and feed the combined representation into a BERT model. However, as the sequence length of a BERT model is limited, such a framework supports little knowledge other than the original input sentences and is thus forced to discard some knowledge. This problem is especially severe for downstream tasks for which the input is a long par…
