Recent advances in NLP demonstrate the effectiveness of training large-scale language models and transferring them to downstream tasks. Can fine-tuning these models on tasks other than language modeling further improve performance? In this paper, we conduct an extensive study of the transferability between 33 NLP tasks across three broad classes of problems (text classification, question answering, and sequence labeling). Our results show that transfer learning is more beneficial than previously thought, especially when target task data is scarce, and can improve performance even when the source task is small or differs substantially from the target task (e.g., part-of-speech tagging transfers well to the DROP QA dataset). We also develop task embeddings that can be used to predict the most transferable source tasks for a given target task, and we validate their effectiveness in experiments controlled for source and target data size. Overall, our experiments reveal that factors such as source data size, task and domain similarity, and task complexity all play a role in determining transferability.
Research Center for Social Computing and Information Retrieval（社会计算与信息检索研究中心）；Harbin Institute of Technology（哈尔滨工业大学）
Many natural language processing (NLP) tasks can be generalized into segmentation problem. In this paper, we combine semi-CRF with neural network to solve NLP segmentation tasks. Our model represents a segment both by composing the input units and embedding the entire segment. We thoroughly study different composition functions and different segment embeddings. We conduct extensive experiments on two typical segmentation tasks: named entity recognition (NER) and Chinese word segmentation (CWS). Experimental results show that our neural semi-CRF model benefits from representing the entire segme…
We present fastHan, an open-source toolkit for four basic tasks in Chinese natural language processing: Chinese word segmentation (CWS), Part-of-Speech (POS) tagging, named entity recognition (NER), and dependency parsing. The backbone of fastHan is a multi-task model based on a pruned BERT, which uses the first 8 layers in BERT. We also provide a 4-layer base model compressed from the 8-layer model. The joint-model is trained and evaluated on 13 corpora of four tasks, yielding near state-of-the-art (SOTA) performance in dependency parsing and NER, achieving SOTA performance in CWS and POS. Be…
IT University of Copenhagen（哥本哈根信息技术大学）；LMU Munich（慕尼黑大学）；Munich Center for Machine Learning（慕尼黑机器学习中心）
With the increase in availability of large pre-trained language models (LMs) in Natural Language Processing (NLP), it becomes critical to assess their fit for a specific target task a priori - as fine-tuning the entire space of available LMs is computationally prohibitive and unsustainable. However, encoder transferability estimation has received little to no attention in NLP. In this paper, we propose to generate quantitative evidence to predict which LM, out of a pool of models, will perform best on a target task without having to fine-tune all candidates. We provide a comprehensive study on…
A Parameter-Efficient Transfer Learning Approach through Multitask Prompt Distillation and Decomposition for Clinical NLP
Department of Health Outcomes and Biomedical Informatics, College of Medicine, University of Florida（佛罗里达大学医学院健康结果与生物医学信息学系）；Cancer Informatics Shared Resource, University of Florida Health Cancer Center（佛罗里达大学健康癌症中心癌症信息学共享资源）
ETH Zurich（苏黎世联邦理工学院）；University of California, San Diego（加州大学圣迭戈分校）
Intermediate-task transfer can benefit a wide range of NLP tasks with properly selected source datasets. However, it is computationally infeasible to experiment with all intermediate transfer combinations, making choosing a useful source task a challenging problem. In this paper, we anticipate that task-specific parameters updated in parameter-efficient tuning methods are likely to encode task-specific information. Therefore, such parameters can be predictive for inter-task transferability. Thus, we propose to exploit these efficiently tuned parameters as off-the-shelf task embeddings for the …
University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Microsoft Research（微软研究院）；Xi’an Jiaotong University（西安交通大学）
Numerical Question Answering is the task of answering questions that require numerical capabilities. Previous works introduce general adversarial attacks to Numerical Question Answering, while not systematically exploring numerical capabilities specific to the topic. In this paper, we propose to conduct numerical capability diagnosis on a series of Numerical Question Answering systems and datasets. A series of numerical capabilities are highlighted, and corresponding dataset perturbations are designed. Empirical results indicate that existing systems are severely challenged by these perturbati…
University of California, Santa Barbara（加州大学圣巴巴拉分校）；J.P. Morgan（摩根大通）
With the recent advance in large pre-trained language models, researchers have achieved record performances in NLP tasks that mostly focus on language pattern matching. The community is experiencing the shift of the challenge from how to model language to the imitation of complex reasoning abilities like human beings. In this work, we investigate the application domain of finance that involves real-world, complex numerical reasoning. We propose a new large-scale dataset, ConvFinQA, aiming to study the chain of numerical reasoning in conversational question answering. Our dataset poses great ch…
New York University（纽约大学）；AWS AI（AWS人工智能）；AWS AI Labs（AWS人工智能实验室）
Recent work has found that multi-task training with a large number of diverse tasks can uniformly improve downstream performance on unseen target tasks. In contrast, literature on task transferability has established that the choice of intermediate tasks can heavily affect downstream task performance. In this work, we aim to disentangle the effect of scale and relatedness of tasks in multi-task representation learning. We find that, on average, increasing the scale of multi-task learning, in terms of the number of tasks, indeed results in better learned representations than smaller multi-task …
Recent studies have revealed a security threat to natural language processing (NLP) models, called the Backdoor Attack. Victim models can maintain competitive performance on clean samples while behaving abnormally on samples with a specific trigger word inserted. Previous backdoor attacking methods usually assume that attackers have a certain degree of data knowledge, either the dataset which users would use or proxy datasets for a similar task, for implementing the data poisoning procedure. However, in this paper, we find that it is possible to hack the model in a data-free way by modifying o…
In this paper, our main contributions are that embeddings from relatively smaller corpora can outperform ones from larger corpora and we make the new Swedish analogy test set publicly available. To achieve a good network performance in natural language processing (NLP) downstream tasks, several factors play important roles: dataset size, the right hyper-parameters, and well-trained embeddings. We show that, with the right set of hyper-parameters, good network performance can be reached even on smaller datasets. We evaluate the embeddings at both the intrinsic and extrinsic levels. The embeddin…
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…
King’s College London（伦敦国王学院）；Amazon（亚马逊）；University of Bahrain（巴林大学）；IIT Madras（印度理工学院马德拉斯分校）；Dasman Diabetes Institute（达斯曼糖尿病研究所）；King Abdulaziz University（阿卜杜勒阿齐兹国王大学）；KFUPM（法赫德国王石油与矿业大学）；University of Tübingen（图宾根大学）；INSEA（摩洛哥国家统计与应用经济学院）；University of Cape Town（开普敦大学）；Hugging Face, Inc（Hugging Face公司）；Saudi Electronic University（沙特电子大学）
Masader (Alyafeai et al., 2021) created a metadata structure to be used for cataloguing Arabic NLP datasets. However, developing an easy way to explore such a catalogue is a challenging task. In order to give the optimal experience for users and researchers exploring the catalogue, several design and user experience challenges must be resolved. Furthermore, user interactions with the website may provide an easy approach to improve the catalogue. In this paper, we introduce Masader Plus, a web interface for users to browse Masader. We demonstrate data exploration, filtration, and a simple API t…
What is the relation between a word and its description, or a word and its embedding? Both descriptions and embeddings are semantic representations of words. But, what information from the original word remains in these representations? Or more importantly, which information about a word do these two representations share? Definition Modeling and Reverse Dictionary are two opposite learning tasks that address these questions. The goal of the Definition Modeling task is to investigate the power of information laying inside a word embedding to express the meaning of the word in a humanly underst…
Monash University（莫纳什大学）；Ant Group（蚂蚁集团）；The Australian National University（澳大利亚国立大学）；Data61 CSIRO；Lehigh University（里海大学）
Natural language processing (NLP) tasks, ranging from text classification to text generation, have been revolutionised by the pre-trained language models, such as BERT. This allows corporations to easily build powerful APIs by encapsulating fine-tuned BERT models for downstream tasks. However, when a fine-tuned BERT model is deployed as a service, it may suffer from different attacks launched by malicious users. In this work, we first present how an adversary can steal a BERT-based API service (the victim/target model) on multiple benchmark datasets with limited prior knowledge and queries. We…
Sogang University（西江大学）；Seoul National University（首尔大学）；Language Technologies Institute, Carnegie Mellon University（卡内基梅隆大学语言技术研究所）
This paper describes our approach to the task of identifying offensive languages in a multilingual setting. We investigate two data augmentation strategies: using additional semi-supervised labels with different thresholds and cross-lingual transfer with data selection. Leveraging the semi-supervised dataset resulted in performance improvements compared to the baseline trained solely with the manually-annotated dataset. We propose a new metric, Translation Embedding Distance, to measure the transferability of instances for cross-lingual data selection. We also introduce various preprocessing s…
We present the Language Interpretability Tool (LIT), an open-source platform for visualization and understanding of NLP models. We focus on core questions about model behavior: Why did my model make this prediction? When does it perform poorly? What happens under a controlled change in the input? LIT integrates local explanations, aggregate analysis, and counterfactual generation into a streamlined, browser-based interface to enable rapid exploration and error analysis. We include case studies for a diverse set of workflows, including exploring counterfactuals for sentiment analysis, measuring…
University of Toronto（多伦多大学）；Vector Institute（向量研究所）；University of Western Ontario（西安大略大学）；St. Michael’s Hospital（圣迈克尔医院）
Clinical machine learning is increasingly multimodal, collected in both structured tabular formats and unstructured forms such as freetext. We propose a novel task of exploring fairness on a multimodal clinical dataset, adopting equalized odds for the downstream medical prediction tasks. To this end, we investigate a modality-agnostic fairness algorithm - equalized odds post processing - and compare it to a text-specific fairness algorithm: debiased clinical word embeddings. Despite the fact that debiased word embeddings do not explicitly address equalized odds of protected groups, we show tha…
Designing shared neural architecture plays an important role in multi-task learning. The challenge is that finding an optimal sharing scheme heavily relies on the expert knowledge and is not scalable to a large number of diverse tasks. Inspired by the promising work of neural architecture search (NAS), we apply reinforcement learning to automatically find possible shared architecture for multi-task learning. Specifically, we use a controller to select from a set of shareable modules and assemble a task-specific architecture, and repeat the same procedure for other tasks. The controller is trai…
Allen Institute for Artificial Intelligence（艾伦人工智能研究所）；Harvard John A. Paulson School of Engineering and Applied Sciences（哈佛大学约翰·A·保尔森工程与应用科学学院）；MIT Computer Science and Artificial Intelligence Laboratory（麻省理工学院计算机科学与人工智能实验室）；University of Washington（华盛顿大学）
Contextual word representations derived from large-scale neural language models are successful across a diverse set of NLP tasks, suggesting that they encode useful and transferable features of language. To shed light on the linguistic knowledge they capture, we study the representations produced by several recent pretrained contextualizers (variants of ELMo, the OpenAI transformer language model, and BERT) with a suite of seventeen diverse probing tasks. We find that linear models trained on top of frozen contextual representations are competitive with state-of-the-art task-specific models in…
Université de Lorraine（洛林大学）；CNRS（法国国家科学研究中心）；LORIA（洛林计算机科学与应用实验室）；HIIT（赫尔辛基信息技术研究所）；Aalto University（阿尔托大学）
Analogical proportions are statements expressed in the form "A is to B as C is to D" and are used for several reasoning and classification tasks in artificial intelligence and natural language processing (NLP). In this paper, we focus on morphological tasks and we propose a deep learning approach to detect morphological analogies. We present an empirical study to see how our framework transfers across languages, and that highlights interesting similarities and differences between these languages. In view of these results, we also discuss the possibility of building a multilingual morphological…
IT University of Copenhagen（哥本哈根信息技术大学）；Technical University of Denmark（丹麦技术大学）
We investigate the problem of determining the predictive confidence (or, conversely, uncertainty) of a neural classifier through the lens of low-resource languages. By training models on sub-sampled datasets in three different languages, we assess the quality of estimates from a wide array of approaches and their dependence on the amount of available data. We find that while approaches based on pre-trained models and ensembles achieve the best results overall, the quality of uncertainty estimates can surprisingly suffer with more data. We also perform a qualitative analysis of uncertainties on…
News archives are an invaluable primary source for placing current events in historical context. But current search engine tools do a poor job at uncovering broad themes and narratives across documents. We present Rookie: a practical software system which uses natural language processing (NLP) to help readers, reporters and editors uncover broad stories in news archives. Unlike prior work, Rookie's design emerged from 18 months of iterative development in consultation with editors and computational journalists. This process lead to a dramatically different approach from previous academic syste…
Pre-trained language models have achieved state-of-the-art results in various Natural Language Processing (NLP) tasks. GPT-3 has shown that scaling up pre-trained language models can further exploit their enormous potential. A unified framework named ERNIE 3.0 was recently proposed for pre-training large-scale knowledge enhanced models and trained a model with 10 billion parameters. ERNIE 3.0 outperformed the state-of-the-art models on various NLP tasks. In order to explore the performance of scaling up ERNIE 3.0, we train a hundred-billion-parameter model called ERNIE 3.0 Titan with up to 260…
Luleå University of Technology（吕勒奥理工大学）；Masakhane；CIS
Dialogue generation is an important NLP task fraught with many challenges. The challenges become more daunting for low-resource African languages. To enable the creation of dialogue agents for African languages, we contribute the first high-quality dialogue datasets for 6 African languages: Swahili, Wolof, Hausa, Nigerian Pidgin English, Kinyarwanda & Yorùbá. These datasets consist of 1,500 turns each, which we translate from a portion of the English multi-domain MultiWOZ dataset. Subsequently, we investigate & analyze the effectiveness of modelling through transfer learning by utilziing state…
Why can pre-trained language models (PLMs) learn universal representations and effectively adapt to broad NLP tasks differing a lot superficially? In this work, we empirically find evidence indicating that the adaptations of PLMs to various few-shot tasks can be reparameterized as optimizing only a few free parameters in a unified low-dimensional intrinsic task subspace, which may help us understand why PLMs could easily adapt to various NLP tasks with small-scale data. To find such a subspace and examine its universality, we propose an analysis pipeline called intrinsic prompt tuning (IPT). S…
Bogazici University（博阿齐奇大学）；IBM Research（IBM研究院）；University of Zurich（苏黎世大学）
Text-based representations of chemicals and proteins can be thought of as unstructured languages codified by humans to describe domain-specific knowledge. Advances in natural language processing (NLP) methodologies in the processing of spoken languages accelerated the application of NLP to elucidate hidden knowledge in textual representations of these biochemical entities and then use it to construct models to predict molecular properties or to design novel molecules. This review outlines the impact made by these advances on drug discovery and aims to further the dialogue between medicinal che…
Public debate forums provide a common platform for exchanging opinions on a topic of interest. While recent studies in natural language processing (NLP) have provided empirical evidence that the language of the debaters and their patterns of interaction play a key role in changing the mind of a reader, research in psychology has shown that prior beliefs can affect our interpretation of an argument and could therefore constitute a competing alternative explanation for resistance to changing one's stance. To study the actual effect of language use vs. prior beliefs on persuasion, we provide a ne…
University of West Bohemia（西波希米亚大学）；NTIS—New Technologies for the Information Society（NTIS—新信息技术学会）
The word embedding methods have been proven to be very useful in many tasks of NLP (Natural Language Processing). Much has been investigated about word embeddings of English words and phrases, but only little attention has been dedicated to other languages. Our goal in this paper is to explore the behavior of state-of-the-art word embedding methods on Czech, the language that is characterized by very rich morphology. We introduce new corpus for word analogy task that inspects syntactic, morphosyntactic and semantic properties of Czech words and phrases. We experiment with Word2Vec and GloVe al…
Cross-model Transferability among Large Language Models on the Platonic Representations of Concepts
Sichuan University, China（四川大学）；Engineering Research Center of Machine Learning and Industry Intelligence, Ministry of Education, China（教育部机器学习与工业智能工程研究中心）
In this work we explore recent advances in Recurrent Neural Networks for large scale Language Modeling, a task central to language understanding. We extend current models to deal with two key challenges present in this task: corpora and vocabulary sizes, and complex, long term structure of language. We perform an exhaustive study on techniques such as character Convolutional Neural Networks or Long-Short Term Memory, on the One Billion Word Benchmark. Our best single model significantly improves state-of-the-art perplexity from 51.3 down to 30.0 (whilst reducing the number of parameters by a f…
Large pre-trained neural language models have supported the effectiveness of many NLP tasks, yet are still prone to generating toxic language hindering the safety of their use. Using empathetic data, we improve over recent work on controllable text generation that aims to reduce the toxicity of generated text. We find we are able to dramatically reduce the size of fine-tuning data to 7.5-30k samples while at the same time making significant improvements over state-of-the-art toxicity mitigation of up to 3.4% absolute reduction (26% relative) from the original work on 2.3m samples, by strategic…
