DOCSLIB.ORG

LIMIT-BERT : Linguistics Informed Multi-Task BERT

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
LIMIT-BERT : Linguistics Informed Multi-Task BERT LIMIT-BERT : Linguistics Informed Multi-Task BERT Junru Zhou1;2;3 , Zhuosheng Zhang 1;2;3, Hai Zhao1;2;3∗, Shuailiang Zhang 1;2;3 1Department of Computer Science and Engineering, Shanghai Jiao Tong University 2Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering, Shanghai Jiao Tong University, Shanghai, China 3MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University fzhoujunru,[email protected], [email protected] Abstract so on (Zhou and Zhao, 2019; Zhou et al., 2020; Ouchi et al., 2018; He et al., 2018b; Li et al., 2019), In this paper, we present Linguistics Informed when taking the latter as downstream tasks for Multi-Task BERT (LIMIT-BERT) for learning the former. In the meantime, introducing linguis- language representations across multiple lin- tic clues such as syntax and semantics into the guistics tasks by Multi-Task Learning. LIMIT- BERT includes five key linguistics tasks: Part- pre-trained language models may furthermore en- Of-Speech (POS) tags, constituent and de- hance other downstream tasks such as various Nat- pendency syntactic parsing, span and depen- ural Language Understanding (NLU) tasks (Zhang dency semantic role labeling (SRL). Differ- et al., 2020a,b). However, nearly all existing lan- ent from recent Multi-Task Deep Neural Net- guage models are usually trained on large amounts works (MT-DNN), our LIMIT-BERT is fully of unlabeled text data (Peters et al., 2018; Devlin linguistics motivated and thus is capable of et al., 2019), without explicitly exploiting linguis- adopting an improved masked training objec- tive according to syntactic and semantic con- tic knowledge. Such observations motivate us to stituents. Besides, LIMIT-BERT takes a semi- jointly consider both types of tasks, pre-training supervised learning strategy to offer the same language models, and solving linguistics inspired large amount of linguistics task data as that NLP problems. We argue such a treatment may for the language model training. As a re- benefit from two-fold. (1) Joint learning is a better sult, LIMIT-BERT not only improves linguis- way to let the former help the latter in a bidirec- tics tasks performance, but also benefits from tional mode, rather than in a unidirectional mode, a regularization effect and linguistics infor- taking the latter as downstream tasks of the former. mation that leads to more general representa- tions to help adapt to new tasks and domains. (2) Naturally empowered by linguistic clues from LIMIT-BERT outperforms the strong baseline joint learning, pre-trained language models will be Whole Word Masking BERT on both depen- more powerful for enhancing downstream tasks. dency and constituent syntactic/semantic pars- Thus we propose Linguistics Informed Multi-Task ing, GLUE benchmark, and SNLI task. Our BERT (LIMIT-BERT), making an attempt to in- practice on the proposed LIMIT-BERT also en- corporate linguistic knowledge into pre-training ables us to release a well pre-trained model for language representation models. The proposed multi-purpose of natural language processing tasks once for all. LIMIT-BERT is implemented in terms of Multi- Task Learning (MTL) (Caruana, 1993) which has 1 Introduction shown useful, by alleviating overfitting to a spe- cific task, thus making the learned representations Recently, pre-trained language models have shown universal across tasks. greatly effective across a range of linguistics in- Since universal language representations are spired natural language processing (NLP) tasks learning by leveraging large amounts of unlabeled such as syntactic parsing, semantic parsing and data which has quite different data volume com- pared with linguistics tasks dataset such as Penn ∗ Corresponding author. This paper was partially sup- 1 ported by National Key Research and Development Program Treebank (PTB) (Marcus et al., 1993). of China (No. 2017YFB0304100), Key Projects of Na- To alleviate such data unbalance on multi-task tional Natural Science Foundation of China (U1836222 and 61733011), Huawei-SJTU long term AI project, Cutting-edge 1PTB is an English treebank with syntactic tree annotation Machine reading comprehension and language model. which only contains 50k sentences. 4450 Findings of the Association for Computational Linguistics: EMNLP 2020, pages 4450–4461 November 16 - 20, 2020. c 2020 Association for Computational Linguistics learning, we apply semi-supervised learning ap- is better at disclosing phrasal continuity, while proach that uses a pre-trained linguistics model2 to the dependency structure is better at indicating annotate large amounts of unlabeled text data and to dependency relation among words. combine with gold linguistics task dataset as our fi- Semantic role labeling (SRL) is dedicated to rec- nal training data. For such pre-processing, it is easy ognizing the predicate-argument structure of a sen- to train our LIMIT-BERT on large amounts of data tence, such as who did what to whom, where and with many tasks concurrently by simply summing when, etc. For argument annotation, there are two up all the concerned losses together. Moreover, formulizations. One is based on text spans, namely since every sentence has labeled with predicted span-based SRL. The other is dependency-based syntax and semantics, we can furthermore improve SRL, which annotates the syntactic head of argu- the masked training objective by fully exploiting ment rather than the entire argument span. SRL is the known syntactic or semantic constituents dur- an important method to obtain semantic informa- ing the language model training process. Unlike tion beneficial to a wide range of NLP tasks (Zhang the previous work MT-DNN (Liu et al., 2019b) et al., 2019; Mihaylov and Frank, 2019). which only fine-tunes BERT on GLUE tasks, our BERT is typically trained on quite large un- LIMIT-BERT is trained on large amounts of data labeled text datasets, BooksCorpus and English in a semi-supervised way and firmly supported by Wikipedia, which have 13GB plain text, while the explicit linguistic clues. datasets for specific linguistics tasks are less than We verify the effectiveness and applicability 100MB. Thus we employ semi-supervised learn- of LIMIT-BERT on Propbank semantic parsing ing to alleviate such data unbalance on multi-task 3 in both span style (CoNLL-2005) (Carreras and learning by using a pre-trained linguistics model Marquez` , 2005) and dependency style, (CoNLL- to label BooksCorpus and English Wikipedia data. 2009) (Hajicˇ et al., 2009) and Penn Treebank (PTB) The pre-trained model jointly learns POS tags and (Marcus et al., 1993) for both constituent and de- the four types of structures on semantics and syn- pendency syntactic parsing. Our empirical results tax, in which the latter is from the XLNet version show that semantics and syntax can indeed ben- of (Zhou et al., 2020), giving state-of-the-art or efit the language representation model via multi- comparable performance for the concerned four task learning and outperforms the strong baseline parsing tasks. During training, we set 10% proba- Whole Word Masking BERT (BERTWWM). bility to use gold syntactic parsing and SRL data: Penn Treebank (PTB) (Marcus et al., 1993), span 2 Tasks and Datasets style SRL (CoNLL-2005) (Carreras and Marquez` , LIMIT-BERT includes five types of downstream 2005) and dependency style SRL (CoNLL-2009) tasks: Part-Of-Speech, constituent and dependency (Hajicˇ et al., 2009). syntactic parsing, span and dependency semantic 2.1 Linguistics-Guided Mask Strategy role labeling (SRL). Both span (constituent) and dependency are two BERT applies two training objectives: Masked broadly-adopted annotation styles for either seman- Language Model (LM) and Next Sentence Predic- tics or syntax, which have been well studied and tion (NSP) based on WordPiece embeddings (Wu discussed from both linguistic and computational et al., 2016) with a 30,000 token vocabulary. For perspectives (Chomsky, 1981; Li et al., 2019). Masked LM training objective, BERT uses training Constituency parsing aims to build a data generator to choose 15% of the token posi- constituency-based parse tree from a sentence tions at random for mask replacement and predict that represents its syntactic structure according to the masked tokens4. Since using different mask- a phrase structure grammar. While dependency ing strategy can improve model performance such parsing identifies syntactic relations (such as as the Whole Word Masking5 which masks all of an adjective modifying a noun) between word the tokens corresponding to a word at once, we pairs in a sentence. The constituent structure further improve the masking strategy by exploit- 2The model may jointly predict syntax and semantics for 4Actually, BERT applies three replacement strategies: (1) both span and dependency annotation styles, which is from the [MASK] token 80% of the time (2) random token 10% of (Zhou et al., 2020) and joint learning with POS tag. the time (3) the unchanged i-th token 10% of the time. This 3It is also called semantic role labeling (SRL) for the se- work uses the same replacement strategies. mantic parsing task over the Propbank. 5https://github.com/huggingface/transformers 4451 A0 A1 tactic and semantic scorers and decoders. We take A0 A1 multi-task learning (MTL) (Caruana, 1993) sharing Federal Paper Board sells paper and wood products . the parameters of token representation and Trans- A2 former encoder, while language modeling layers A1 A0 and the top task-specific layers have independent Span and Dependency SRL parameters. The training procedure is simple that federal paper board [MASK] paper and wood [MASK] . we just sum up the language model loss with task- (a) Semantic Phrase Masking. specific losses together. S (1,9) 3.2 Token Representation NP VP Following BERT token representation (Devlin (1,3) (4,8) 9 et al., 2019), the first token is always the [CLS] VBZ NP NNP NNP NNP (5,8) token. If input X is packed by a sentence pair Board sells Federal Paper 4 3 1 2 NN CC NN NNS X1; X2, we separate the two sentences with a spe- paper and wood products 5 6 7 8 cial token [SEP] (”packed by” means connect two Constituent Syntactic Tree sentences as BERT training).
Load more

Recommended publications
  • Abstraction and Generalisation in Semantic Role Labels: Propbank
    Abstraction and Generalisation in Semantic Role Labels: PropBank, VerbNet or both? Paola Merlo Lonneke Van Der Plas Linguistics Department Linguistics Department University of Geneva University of Geneva 5 Rue de Candolle, 1204 Geneva 5 Rue de Candolle, 1204 Geneva Switzerland Switzerland [email protected] [email protected] Abstract The role of theories of semantic role lists is to obtain a set of semantic roles that can apply to Semantic role labels are the representa- any argument of any verb, to provide an unam- tion of the grammatically relevant aspects biguous identifier of the grammatical roles of the of a sentence meaning. Capturing the participants in the event described by the sentence nature and the number of semantic roles (Dowty, 1991). Starting from the first proposals in a sentence is therefore fundamental to (Gruber, 1965; Fillmore, 1968; Jackendoff, 1972), correctly describing the interface between several approaches have been put forth, ranging grammar and meaning. In this paper, we from a combination of very few roles to lists of compare two annotation schemes, Prop- very fine-grained specificity. (See Levin and Rap- Bank and VerbNet, in a task-independent, paport Hovav (2005) for an exhaustive review). general way, analysing how well they fare In NLP, several proposals have been put forth in in capturing the linguistic generalisations recent years and adopted in the annotation of large that are known to hold for semantic role samples of text (Baker et al., 1998; Palmer et al., labels, and consequently how well they 2005; Kipper, 2005; Loper et al., 2007). The an- grammaticalise aspects of meaning.
    [Show full text]
  • A Novel Large-Scale Verbal Semantic Resource and Its Application to Semantic Role Labeling
    VerbAtlas: a Novel Large-Scale Verbal Semantic Resource and Its Application to Semantic Role Labeling Andrea Di Fabio}~, Simone Conia}, Roberto Navigli} }Department of Computer Science ~Department of Literature and Modern Cultures Sapienza University of Rome, Italy {difabio,conia,navigli}@di.uniroma1.it Abstract The automatic identification and labeling of ar- gument structures is a task pioneered by Gildea We present VerbAtlas, a new, hand-crafted and Jurafsky(2002) called Semantic Role Label- lexical-semantic resource whose goal is to ing (SRL). SRL has become very popular thanks bring together all verbal synsets from Word- to its integration into other related NLP tasks Net into semantically-coherent frames. The such as machine translation (Liu and Gildea, frames define a common, prototypical argu- ment structure while at the same time pro- 2010), visual semantic role labeling (Silberer and viding new concept-specific information. In Pinkal, 2018) and information extraction (Bas- contrast to PropBank, which defines enumer- tianelli et al., 2013). ative semantic roles, VerbAtlas comes with In order to be performed, SRL requires the fol- an explicit, cross-frame set of semantic roles lowing core elements: 1) a verb inventory, and 2) linked to selectional preferences expressed in terms of WordNet synsets, and is the first a semantic role inventory. However, the current resource enriched with semantic information verb inventories used for this task, such as Prop- about implicit, shadow, and default arguments. Bank (Palmer et al., 2005) and FrameNet (Baker We demonstrate the effectiveness of VerbAtlas et al., 1998), are language-specific and lack high- in the task of dependency-based Semantic quality interoperability with existing knowledge Role Labeling and show how its integration bases.
    [Show full text]
  • Generating Lexical Representations of Frames Using Lexical Substitution
    Generating Lexical Representations of Frames using Lexical Substitution Saba Anwar Artem Shelmanov Universitat¨ Hamburg Skolkovo Institute of Science and Technology Germany Russia [email protected] [email protected] Alexander Panchenko Chris Biemann Skolkovo Institute of Science and Technology Universitat¨ Hamburg Russia Germany [email protected] [email protected] Abstract Seed sentence: I hope PattiHelper can helpAssistance youBenefited party soonTime . Semantic frames are formal linguistic struc- Substitutes for Assistance: assist, aid tures describing situations/actions/events, e.g. Substitutes for Helper: she, I, he, you, we, someone, Commercial transfer of goods. Each frame they, it, lori, hannah, paul, sarah, melanie, pam, riley Substitutes for Benefited party: me, him, folk, her, provides a set of roles corresponding to the sit- everyone, people uation participants, e.g. Buyer and Goods, and Substitutes for Time: tomorrow, now, shortly, sooner, lexical units (LUs) – words and phrases that tonight, today, later can evoke this particular frame in texts, e.g. Sell. The scarcity of annotated resources hin- Table 1: An example of the induced lexical represen- ders wider adoption of frame semantics across tation (roles and LUs) of the Assistance FrameNet languages and domains. We investigate a sim- frame using lexical substitutes from a single seed sen- ple yet effective method, lexical substitution tence. with word representation models, to automat- ically expand a small set of frame-annotated annotated resources. Some publicly available re- sentences with new words for their respective sources are FrameNet (Baker et al., 1998) and roles and LUs. We evaluate the expansion PropBank (Palmer et al., 2005), yet for many lan- quality using FrameNet.
    [Show full text]
  • Distributional Semantics, Wordnet, Semantic Role
    Semantics Kevin Duh Intro to NLP, Fall 2019 Outline • Challenges • Distributional Semantics • Word Sense • Semantic Role Labeling The Challenge of Designing Semantic Representations • Q: What is semantics? • A: The study of meaning • Q: What is meaning? • A: … We know it when we see it • These sentences/phrases all have the same meaning: • XYZ corporation bought the stock. • The stock was bought by XYZ corporation. • The purchase of the stock by XYZ corporation... • The stock purchase by XYZ corporation... But how to formally define it? Meaning Sentence Representation Example Representations Sentence: “I have a car” Logic Formula Graph Representation Key-Value Records Example Representations Sentence: “I have a car” As translation in “Ich habe ein Auto” another language There’s no single agreed-upon representation that works in all cases • Different emphases: • Words or Sentences • Syntax-Semantics interface, Logical Inference, etc. • Different aims: • “Deep (and narrow)” vs “Shallow (and broad)” • e.g. Show me all flights from BWI to NRT. • Do we link to actual flight records? • Or general concept of flying machines? Outline • Challenges • Distributional Semantics • Word Sense • Semantic Role Labeling Distributional Semantics 10 Learning Distributional Semantics from large text dataset • Your pet dog is so cute • Your pet cat is so cute • The dog ate my homework • The cat ate my homework neighbor(dog) overlaps-with neighbor(cats) so meaning(dog) is-similar-to meaning(cats) 11 Word2Vec implements Distribution Semantics Your pet dog is so cute 1. Your pet - is so | dog 2. dog | Your pet - is so From: Mikolov, Tomas; et al. (2013). "Efficient Estimation of Word 12 Representations in Vector Space” Latent Semantic Analysis (LSA) also implements Distributional Semantics Pet-peeve: fundamentally, neural approaches aren’t so different from classical LSA.
    [Show full text]
  • Foundations of Natural Language Processing Lecture 16 Semantic
    Language is Flexible Foundations of Natural Language Processing Often we want to know who did what to whom (when, where, how and why) • Lecture 16 But the same event and its participants can have different syntactic realizations. Semantic Role Labelling and Argument Structure • Sandy broke the glass. vs. The glass was broken by Sandy. Alex Lascarides She gave the boy a book. vs. She gave a book to the boy. (Slides based on those of Schneider, Koehn, Lascarides) Instead of focusing on syntax, consider the semantic roles (also called 13 March 2020 • thematic roles) defined by each event. Alex Lascarides FNLP Lecture 16 13 March 2020 Alex Lascarides FNLP Lecture 16 1 Argument Structure and Alternations Stanford Dependencies Mary opened the door • The door opened John slices bread with a knife • This bread slices easily The knife slices cleanly Mary loaded the truck with hay • Mary loaded hay onto the the truck The truck was loaded with hay (by Mary) The hay was loaded onto the truck (by Mary) John gave a present to Mary • John gave Mary a present cf Mary ate the sandwich with Kim! Alex Lascarides FNLP Lecture 16 2 Alex Lascarides FNLP Lecture 16 3 Syntax-Semantics Relationship Outline syntax = semantics • 6 The semantic roles played by different participants in the sentence are not • trivially inferable from syntactical relations . though there are patterns! • The idea of semantic roles can be combined with other aspects of meaning • (beyond this course). Alex Lascarides FNLP Lecture 16 4 Alex Lascarides FNLP Lecture 16 5 Commonly used thematic
    [Show full text]
  • Merging Propbank, Nombank, Timebank, Penn Discourse Treebank and Coreference James Pustejovsky, Adam Meyers, Martha Palmer, Massimo Poesio
    Merging PropBank, NomBank, TimeBank, Penn Discourse Treebank and Coreference James Pustejovsky, Adam Meyers, Martha Palmer, Massimo Poesio Abstract annotates the temporal features of propositions Many recent annotation efforts for English and the temporal relations between propositions. have focused on pieces of the larger problem The Penn Discourse Treebank (Miltsakaki et al of semantic annotation, rather than initially 2004a/b) treats discourse connectives as producing a single unified representation. predicates and the sentences being joined as This paper discusses the issues involved in arguments. Researchers at Essex were merging four of these efforts into a unified responsible for the coreference markup scheme linguistic structure: PropBank, NomBank, the developed in MATE (Poesio et al, 1999; Poesio, Discourse Treebank and Coreference 2004a) and have annotated corpora using this Annotation undertaken at the University of scheme including a subset of the Penn Treebank Essex. We discuss resolving overlapping and (Poesio and Vieira, 1998), and the GNOME conflicting annotation as well as how the corpus (Poesio, 2004a). This paper discusses the various annotation schemes can reinforce issues involved in creating a Unified Linguistic each other to produce a representation that is Annotation (ULA) by merging annotation of greater than the sum of its parts. examples using the schemata from these efforts. Crucially, all individual annotations can be kept 1. Introduction separate in order to make it easy to produce alternative annotations of a specific type of The creation of the Penn Treebank (Marcus et al, semantic information without need to modify the 1993) and the word sense-annotated SEMCOR annotation at the other levels. Embarking on (Fellbaum, 1997) have shown how even limited separate annotation efforts has the advantage of amounts of annotated data can result in major allowing researchers to focus on the difficult improvements in complex natural language issues in each area of semantic annotation and understanding systems.
    [Show full text]
  • Radio Frequency Identification Based Smart
    ISSN (Online) 2394-6849 International Journal of Engineering Research in Electronics and Communication Engineering (IJERECE) Vol 4, Issue 11, November 2017 Unstructured Text to DBPEDIA RDF Triples – Entity Extraction [1] Monika S G, [2] Chiranjeevi S, [3] Harshitha A, [4] Harshitha M, [5] V K Tivari, [6] Raghevendra Rao [1 – 5] Department of Electronics and Communication Engineering, Sri Sairam College of Engineering, Anekal, Bengaluru. [5] Assistant Professor, Department of Electronic and Communication Engineering, Sri Sairam College of Engineering, Anekal, Bengaluru. [6].Assistant Professor, Department of Computer Science and Engineering, Sri Sairam College of Engineering, Anekal, Bengaluru Abstract:- In the means of current technologies Use of data, information has grown significantly over the last few years. The information processing facing an issue like where the data is originating from multiple sources in an uncontrolled environment. The reason for the uncontrolled environment is the data gathered beyond the organization and generated by many people working outside the organization. The intent of this paper is delving into this unformatted information and build the framework in such a way that the information becomes more managed and used in the organization. Case and point for resume submitted for particular positions should become searchable. In this framework, we try and solve the problem and provide suggestions on how to solve other similar problem. In this paper, we describe an end-to-end system that automatically extracts RDF triples describing entity relations and properties from unstructured text. This system is based on a pipeline of text processing modules that includes an asemantic parser and a co-reference solver.
    [Show full text]
  • Frame-Based Continuous Lexical Semantics Through Exponential Family Tensor Factorization and Semantic Proto-Roles
    Frame-Based Continuous Lexical Semantics through Exponential Family Tensor Factorization and Semantic Proto-Roles Francis Ferraro and Adam Poliak and Ryan Cotterell and Benjamin Van Durme Center for Language and Speech Processing Johns Hopkins University fferraro,azpoliak,ryan.cotterell,[email protected] Abstract AGENT GOAL We study how different frame annota- ATTEMPT tions complement one another when learn- She said Bill would try the same tactic again. ing continuous lexical semantics. We Figure 1: A simple frame analysis. learn the representations from a tensorized skip-gram model that consistently en- codes syntactic-semantic content better, with multiple 10% gains over baselines. with judgements about multiple underlying prop- erties about what is likely true of the entity fill- 1 Introduction ing the role. For example, SPR talks about how likely it is for Bill to be a willing participant in the Consider “Bill” in Fig.1: what is his involve- ATTEMPT. The answer to this and other simple ment with the words “would try,” and what does judgments characterize Bill and his involvement. this involvement mean? Word embeddings repre- Since SPR both captures the likelihood of certain sent such meaning as points in a real-valued vec- properties and characterizes roles as groupings of tor space (Deerwester et al., 1990; Mikolov et al., properties, we can view SPR as representing a type 2013). These representations are often learned by of continuous frame semantics. exploiting the frequency that the word cooccurs with contexts, often within a user-defined window We are interested in capturing these SPR-based (Harris, 1954; Turney and Pantel, 2010).
    [Show full text]
  • Verbnet and Propbank
    The English VerbNet PropBank VerbNet and PropBank Word and VerbNets for Semantic Processing DAAD Summer School in Advanced Language Engineering, Kathmandu University, Nepal Day 3 Annette Hautli 1 / 34 The English VerbNet PropBank Recap Levin (1993) Detailed analysis of syntactic alternations in English Verbs with a common meaning partake in the same syntactic alternations Verb classes that share syntactic properties ! How can we use this information in nlp? 2 / 34 The English VerbNet Lexical information PropBank The VerbNet API Today 1 The English VerbNet Lexical information The VerbNet API 2 PropBank 3 / 34 The English VerbNet Lexical information PropBank The VerbNet API Overview Hierarchically arranged verb classes Inspired by Levin (1993) and further extended for nlp purposes Levin: 240 classes (47 top level, 193 second level), around 3000 lemmas VerbNet: 471 classes, around 4000 lemmas Explicit links to other resources, e.g. WordNet, FrameNet 4 / 34 The English VerbNet Lexical information PropBank The VerbNet API Outline 1 The English VerbNet Lexical information The VerbNet API 2 PropBank 5 / 34 The English VerbNet Lexical information PropBank The VerbNet API Components Thematic roles 23 thematic roles are used in VerbNet Each argument is assigned a unique thematic role When verbs have more than one sense, they usually have different thematic roles 6 / 34 The English VerbNet Lexical information PropBank The VerbNet API Components Role Example Actor Susan was chitchatting. Agent Cynthia nibbled on the carrot. Asset Carmen purchased a dress for $50. Attribute The price of oil soared. Beneficiary Sandy sang a song to me. Cause My heart is pounding from fear.
    [Show full text]
  • Arxiv:2005.08206V1 [Cs.CL] 17 May 2020 and the Projection Method Itself
    Building a Hebrew Semantic Role Labeling Lexical Resource from Parallel Movie Subtitles Ben Eyal, Michael Elhadad Dept. of Computer Science, Ben-Gurion University of the Negev Beer Sheva, Israel [email protected], [email protected] Abstract We present a semantic role labeling resource for Hebrew built semi-automatically through annotation projection from English. This corpus is derived from the multilingual OpenSubtitles dataset and includes short informal sentences, for which reliable linguistic annotations have been computed. We provide a fully annotated version of the data including morphological analysis, dependency syntax and semantic role labeling in both FrameNet and PropBank styles. Sentences are aligned between English and Hebrew, both sides include full annotations and the explicit mapping from the English arguments to the Hebrew ones. We train a neural SRL model on this Hebrew resource exploiting the pre-trained multilingual BERT transformer model, and provide the first available baseline model for Hebrew SRL as a reference point. The code we provide is generic and can be adapted to other languages to bootstrap SRL resources. Keywords: SRL, FrameNet, PropBank, Hebrew, Cross-lingual Linguistic Annotations 1. Introduction jection. We provide full source code and datasets for the described method https://github.com/bgunlp/ Semantic role labeling (SRL) is the task that consists of an- hebrew_srl. The resulting dataset is the first available notating sentences with labels that answer questions such as SRL resource for Hebrew. “Who did what to whom, when, and where?”, the answers to these questions are called “roles.” 1.1. FrameNet Two major SRL annotation schemes have been used in re- This work is part of a larger effort to develop a Hebrew cent years, FrameNet (Baker et al., 1998) and PropBank FrameNet (Hayoun and Elhadad, 2016).
    [Show full text]
  • Slides Are Modified from Prof
    Dependency Parse TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: AA AA AA AA AAAA A A A A A A A Dependency Tags aux – auxiliary auxpass – passive auxiliary cop -- copula conj – conjunct cc – coordination ref -- referent subj – subject nsubj – nominal subject nsubjpass – passive nominal subject csubj – clausal subject det – determiner prep – prepositional modifier Dependency Tags comp – complement mod -- modifier obj – object dobj – direct object iobj – indirect object pobj – object of preposition attr – attribute ccomp – clausal complement with internal subject xcomp – clausal complement with external subject acomp – adjectival complement compl -- complementizer Dependency Tags mod – modifier advcl – adverbial clause modifier tmod – temporal modifier rcmod – relative clause modifier amod – adjectival modifier infmod – infinitival modifier partmod – participial modifier appos – appositional modifier nn – noun compound modifier poss – possession modifier Exercise We learned dependency parsers Exercise We learned dependency parsers nsubj(learned-2, I-1) amod(parsers-4, dependency-3) dobj(learned-2, parsers-4) Exercise I am excited about my project. Exercise I am excited about my project. dependencies: nsubj(excited-3, I-1) cop(excited-3, am-2) prep(excited-3, about-4) poss(project-6, my-5) pobj(about-4, project-6) Exercise I am excited about my project. “collapsed” version of dependencies: nsubj(excited-3, I-1) cop(excited-3, am-2) poss(project-6, my-5) prep_about(excited-3,
    [Show full text]
  • A Semantic Parsing and Reasoning-Based Approach to Knowledge Base Question Answering
    The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-21) A Semantic Parsing and Reasoning-Based Approach to Knowledge Base Question Answering Ibrahim Abdelaziz∗, Srinivas Ravishankar∗, Pavan Kapanipathi∗, Salim Roukos, Alexander Gray IBM Research IBM T.J. Watson Research Center Yorktown Heights, NY, USA fibrahim.abdelaziz1, srini, [email protected], fkapanipa, [email protected] Abstract associated PropBank information to a KB such as DBpedia. Finally, the logic forms are used by Logical Neural Network Knowledge Base Question Answering (KBQA) is a task where existing techniques have faced significant challenges, such as (LNN) (Riegel et al. 2020), a neuro-symbolic reasoner that the need for complex question understanding, reasoning, and enables DTQA to perform different types of reasoning. DTQA large training datasets. In this work, we demonstrate Deep comprises several modules which operate within a pipeline Thinking Question Answering (DTQA), a semantic parsing and are individually trained for their specific tasks. and reasoning-based KBQA system. DTQA (1) integrates mul- This demonstration shows an interpretable methodology tiple, reusable modules that are trained specifically for their for KBQA that achieves state-of-the-art results on two promi- individual tasks (e.g. semantic parsing, entity linking, and rela- nent datasets (QALD-9, LC-QuAD-1). DTQA is a system of tionship linking), eliminating the need for end-to-end KBQA systems that demonstrates: (1) the use of task-general seman- training data; (2) leverages semantic parsing and a reasoner for tic parsing via AMR; (2) an approach that aligns AMR to KB improved question understanding. DTQA is a system of sys- triples; and (3) the use of a neuro-symbolic reasoner for the tems that achieves state-of-the-art performance on two popular KBQA datasets.
    [Show full text]