Although pre-trained contextualized language models such as BERT achieve significant performance on various downstream tasks, current language representation still only focuses on linguistic objective at a specific granularity, which may not applicable when multiple levels of linguistic units are involved at the same time. Thus this work introduces and explores the universal representation learning, i.e., embeddings of different levels of linguistic unit in a uniform vector space. We present a universal representation model, BURT (BERT-inspired Universal Representation from learning meaningful segmenT), to encode different levels of linguistic unit into the same vector space. Specifically, we extract and mask meaningful segments based on point-wise mutual information (PMI) to incorporate different granular objectives into the pre-training stage. We conduct experiments on datasets for English and Chinese including the GLUE and CLUE benchmarks, where our model surpasses its baselines and alternatives on a wide range of downstream tasks. We present our approach of constructing analogy datasets in terms of words, phrases and sentences and experiment with multiple representation models to examine geometric properties of the learned vector space through a task-independent evaluation. Finally, we verify the effectiveness of our unified pre-training strategy in two real-world text matching scenarios. As a result, our model significantly outperforms existing information retrieval (IR) methods and yields universal representations that can be directly applied to retrieval-based question-answering and natural language generation tasks.
Pre-trained contextualized language models such as BERT have shown great effectiveness in a wide range of downstream Natural Language Processing (NLP) tasks. However, the effective representations offered by the models target at each token inside a sequence rather than each sequence and the fine-tuning step involves the input of both sequences at one time, leading to unsatisfying representations of various sequences with different granularities. Especially, as sentence-level representations taken as the full training context in these models, there comes inferior performance on lower-level ling…
University of Chinese Academy of Sciences（中国科学院大学）
With the development and business adoption of knowledge graph, there is an increasing demand for extracting entities and relations of knowledge graphs from unstructured domain documents. This makes the automatic knowledge extraction for domain text quite meaningful. This paper proposes a knowledge extraction method based on BERT, which is used to extract knowledge points from unstructured specific domain texts (such as insurance clauses in the insurance industry) automatically to save manpower of knowledge graph construction. Different from the commonly used methods which are based on rules, t…
Automatic question generation aims at the generation of questions from a context, with the corresponding answers being sub-spans of the given passage. Whereas, most of the methods mostly rely on heuristic rules to generate questions, more recently also neural network approaches have been proposed. In this work, we propose a variant of the self-attention Transformer network architectures model to generate meaningful and diverse questions. To this end, we propose an easy to use model consisting of the conjunction of the Transformer decoder GPT-2 model with Transformer encoder BERT for the downst…
Tehran Institute for Advanced Studies（德黑兰高级研究所）；Khatam University（哈塔姆大学）；Iran University of Science and Technology（伊朗科技大学）
Several studies have been carried out on revealing linguistic features captured by BERT. This is usually achieved by training a diagnostic classifier on the representations obtained from different layers of BERT. The subsequent classification accuracy is then interpreted as the ability of the model in encoding the corresponding linguistic property. Despite providing insights, these studies have left out the potential role of token representations. In this paper, we provide a more in-depth analysis on the representation space of BERT in search for distinct and meaningful subspaces that can expl…
Texas A&M University（得克萨斯农工大学）；Facebook AI Research（Facebook人工智能研究院）
Existing search engines use keyword matching or tf-idf based matching to map the query to the web-documents and rank them. They also consider other factors such as page rank, hubs-and-authority scores, knowledge graphs to make the results more meaningful. However, the existing search engines fail to capture the meaning of query when it becomes large and complex. BERT, introduced by Google in 2018, provides embeddings for words as well as sentences. In this paper, I have developed a semantics-oriented search engine using neural networks and BERT embeddings that can search for query and rank the…
Arizona State University（亚利桑那州立大学）；The University of Texas Health Science Center（德克萨斯大学健康科学中心）
Background: Identifying relationships between clinical events and temporal expressions is a key challenge in meaningfully analyzing clinical text for use in advanced AI applications. While previous studies exist, the state-of-the-art performance has significant room for improvement. Methods: We studied several variants of BERT (Bidirectional Encoder Representations using Transformers) some involving clinical domain customization and the others involving improved architecture and/or training strategies. We evaluated these methods using a direct temporal relations dataset which is a semantically…
Most adversarial attack methods that are designed to deceive a text classifier change the text classifier's prediction by modifying a few words or characters. Few try to attack classifiers by rewriting a whole sentence, due to the difficulties inherent in sentence-level rephrasing as well as the problem of setting the criteria for legitimate rewriting. In this paper, we explore the problem of creating adversarial examples with sentence-level rewriting. We design a new sampling method, named ParaphraseSampler, to efficiently rewrite the original sentence in multiple ways. Then we propose a new …
Named entity recognition (NER) is a widely applicable natural language processing task and building block of question answering, topic modeling, information retrieval, etc. In the medical domain, NER plays a crucial role by extracting meaningful chunks from clinical notes and reports, which are then fed to downstream tasks like assertion status detection, entity resolution, relation extraction, and de-identification. Reimplementing a Bi-LSTM-CNN-Char deep learning architecture on top of Apache Spark, we present a single trainable NER model that obtains new state-of-the-art results on seven pub…
University of Electronics Science and Technology of China（电子科技大学）
BERT achieves remarkable results in text classification tasks, it is yet not fully exploited, since only the last layer is used as a representation output for downstream classifiers. The most recent studies on the nature of linguistic features learned by BERT, suggest that different layers focus on different kinds of linguistic features. We propose a CNN-Enhanced Transformer-Encoder model which is trained on top of fixed BERT $[CLS]$ representations from all layers, employing Convolutional Neural Networks to generate QKV feature maps inside the Transformer-Encoder, instead of linear projection…
School of Computer Science, Tel Aviv University（特拉维夫大学计算机科学学院）；Allen Institute for AI（艾伦人工智能研究所）
Large pre-trained language models have been shown to encode large amounts of world and commonsense knowledge in their parameters, leading to substantial interest in methods for extracting that knowledge. In past work, knowledge was extracted by taking manually-authored queries and gathering paraphrases for them using a separate pipeline. In this work, we propose a method for automatically rewriting queries into "BERTese", a paraphrase query that is directly optimized towards better knowledge extraction. To encourage meaningful rewrites, we add auxiliary loss functions that encourage the query …
University of Houston（休斯顿大学）；Salesforce Research（Salesforce研究院）
Byte-pair encoding (BPE) is a ubiquitous algorithm in the subword tokenization process of language models as it provides multiple benefits. However, this process is solely based on pre-training data statistics, making it hard for the tokenizer to handle infrequent spellings. On the other hand, though robust to misspellings, pure character-level models often lead to unreasonably long sequences and make it harder for the model to learn meaningful words. To alleviate these challenges, we propose a character-based subword module (char2subword) that learns the subword embedding table in pre-trained…
Northwestern Polytechnical University（西北工业大学）；University of Georgia（佐治亚大学）；Athens Academy（雅典学院）
Linking computational natural language processing (NLP) models and neural responses to language in the human brain on the one hand facilitates the effort towards disentangling the neural representations underpinning language perception, on the other hand provides neurolinguistics evidence to evaluate and improve NLP models. Mappings of an NLP model's representations of and the brain activities evoked by linguistic input are typically deployed to reveal this symbiosis. However, two critical problems limit its advancement: 1) The model's representations (artificial neurons, ANs) rely on layer-le…
Unsupervised dialogue structure learning is an important and meaningful task in natural language processing. The extracted dialogue structure and process can help analyze human dialogue, and play a vital role in the design and evaluation of dialogue systems. The traditional dialogue system requires experts to manually design the dialogue structure, which is very costly. But through unsupervised dialogue structure learning, dialogue structure can be automatically obtained, reducing the cost of developers constructing dialogue process. The learned dialogue structure can be used to promote the di…
Twitter is a well-known microblogging social site where users express their views and opinions in real-time. As a result, tweets tend to contain valuable information. With the advancements of deep learning in the domain of natural language processing, extracting meaningful information from tweets has become a growing interest among natural language researchers. Applying existing language representation models to extract information from Twitter does not often produce good results. Moreover, there is no existing language representation models for text analysis specific to the social media domai…
Deep pre-trained contextualized encoders like BERT (Delvin et al., 2019) demonstrate remarkable performance on a range of downstream tasks. A recent line of research in probing investigates the linguistic knowledge implicitly learned by these models during pre-training. While most work in probing operates on the task level, linguistic tasks are rarely uniform and can be represented in a variety of formalisms. Any linguistics-based probing study thereby inevitably commits to the formalism used to annotate the underlying data. Can the choice of formalism affect probing results? To investigate, w…
Neural language models exhibit impressive performance on a variety of tasks, but their internal reasoning may be difficult to understand. Prior art aims to uncover meaningful properties within model representations via probes, but it is unclear how faithfully such probes portray information that the models actually use. To overcome such limitations, we propose a technique, inspired by causal analysis, for generating counterfactual embeddings within models. In experiments testing our technique, we produce evidence that suggests some BERT-based models use a tree-distance-like representation of s…
Splintering Nonconcatenative Languages for Better Tokenization
Department of Computer Science（计算机科学系）；Data Science Research Center（数据科学研究中心）；Bar-Ilan University（巴伊兰大学）；Ben-Gurion University of the Negev（贝内-约瑟夫大学（内盖夫分校））；δ DICTA, Jerusalem, Israel（δ DICTA，耶路撒冷，以色列）
Language Models (LMs) have been ubiquitously leveraged in various tasks including spoken language understanding (SLU). Spoken language requires careful understanding of speaker interactions, dialog states and speech induced multimodal behaviors to generate a meaningful representation of the conversation. In this work, we propose to dissect SLU into three representative properties:conversational (disfluency, pause, overtalk), channel (speaker-type, turn-tasks) and ASR (insertion, deletion,substitution). We probe BERT based language models (BERT, RoBERTa) trained on spoken transcripts to investi…
We study the problem of model extraction in natural language processing, in which an adversary with only query access to a victim model attempts to reconstruct a local copy of that model. Assuming that both the adversary and victim model fine-tune a large pretrained language model such as BERT (Devlin et al. 2019), we show that the adversary does not need any real training data to successfully mount the attack. In fact, the attacker need not even use grammatical or semantically meaningful queries: we show that random sequences of words coupled with task-specific heuristics form effective queri…
All Roads Lead to Rome? Exploring the Invariance of Transformers' Representations
Tsinghua University（清华大学）；Fudan University（复旦大学）；Max Planck Institute for Intelligent Systems（马克斯·普朗克智能系统研究所）；ETH Zürich（苏黎世联邦理工学院）；Bar-Ilan University（巴伊兰大学）
University of Zurich（苏黎世大学）；Harbin Institute of Technology（哈尔滨工业大学）
This paper presents the participation of the MiniTrue team in the FinSim-3 shared task on learning semantic similarities for the financial domain in English language. Our approach combines contextual embeddings learned by transformer-based language models with network structures embeddings extracted on external knowledge sources, to create more meaningful representations of financial domain entities and terms. For this, two BERT based language models and a knowledge graph embedding model are used. Besides, we propose a voting function to joint three basic models for the final inference. Experi…
Masked speech modeling (MSM) methods such as wav2vec2 or w2v-BERT learn representations over speech frames which are randomly masked within an utterance. While these methods improve performance of Automatic Speech Recognition (ASR) systems, they have one major limitation. They treat all unsupervised speech samples with equal weight, which hinders learning as not all samples have relevant information to learn meaningful representations. In this work, we address this limitation. We propose ask2mask (ATM), a novel approach to focus on specific samples during MSM pre-training. ATM employs an exter…
In this paper, we propose Stacked DeBERT, short for Stacked Denoising Bidirectional Encoder Representations from Transformers. This novel model improves robustness in incomplete data, when compared to existing systems, by designing a novel encoding scheme in BERT, a powerful language representation model solely based on attention mechanisms. Incomplete data in natural language processing refer to text with missing or incorrect words, and its presence can hinder the performance of current models that were not implemented to withstand such noises, but must still perform well even under duress. T…
