Frame-Based Continuous Lexical Semantics Through Exponential Family Tensor Factorization and Semantic Proto-Roles

Total Page:16

File Type:pdf, Size:1020Kb

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). When properties and expectations within word embed- built from large-scale sources, like Wikipedia or dings. We present a method that learns frame- web crawls, embeddings capture general charac- enriched embeddings from millions of documents teristics of words and allow for robust downstream that have been semantically parsed with multiple applications (Kim, 2014; Das et al., 2015). different frame analyzers (Ferraro et al., 2014). Frame semantics generalize word meanings to Our method leverages Cotterell et al.(2017)’s that of analyzing structured and interconnected la- formulation of Mikolov et al.(2013)’s popular beled “concepts” and abstractions (Minsky, 1974; skip-gram model as exponential family principal Fillmore, 1976, 1982). These concepts, or roles, component analysis (EPCA) and tensor factor- arXiv:1706.09562v1 [cs.CL] 29 Jun 2017 implicitly encode expected properties of that word. ization. This paper’s primary contributions are: In a frame semantic analysis of Fig.1, the segment (i) enriching learned word embeddings with mul- “would try” triggers the ATTEMPT frame, filling tiple, automatically obtained frames from large, the expected roles AGENT and GOAL with “Bill” disparate corpora; and (ii) demonstrating these and “the same tactic,” respectively. While frame enriched embeddings better capture SPR-based semantics provide a structured form for analyzing properties. In so doing, we also generalize Cot- words with crisp, categorically-labeled concepts, terell et al.’s method to arbitrary tensor dimen- the encoded properties and expectations are im- sions. This allows us to include an arbitrary plicit. What does it mean to fill a frame’s role? amount of semantic information when learning Semantic proto-role (SPR) theory, motivated by embeddings. Our variable-size tensor factoriza- Dowty(1991)’s thematic proto-role theory, offers tion code is available at https://github.com/ an answer to this. SPR replaces categorical roles fmof/tensor-factorization. 2 Frame Semantics and Proto-Roles words (CBOW)—repopularized these methods. We focus on SG, which predicts the context i Frame semantics currently used in NLP have a rich around a word j, with learned representations ci history in linguistic literature. Fillmore(1976)’s and wj, respectively, as p(context i j word j) / frames are based on a word’s context and prototyp- | | exp (ci wj) = exp (1 (ci wj)) ; where is the ical concepts that an individual word evokes; they Hadamard (pointwise) product. Traditionally, the intend to represent the meaning of lexical items by context words i are those words within a small mapping words to real world concepts and shared window of j and are trained with negative sam- experiences. Frame-based semantics have inspired pling (Goldberg and Levy, 2014). many semantic annotation schemata and datasets, such as FrameNet (Baker et al., 1998), PropBank 3.1 Skip-Gram as Matrix Factorization (Palmer et al., 2005), and Verbnet (Schuler, 2005), Levy and Goldberg(2014b), and subsequently as well as composite resources (Hovy et al., 2006; Keerthi et al.(2015), showed how vectors learned 1 Palmer, 2009; Banarescu et al., 2012). under SG with the negative sampling are, under Thematic Roles and Proto Roles These re- certain conditions, the factorization of (shifted) sources map words to their meanings through positive pointwise mutual information. Cotterell discrete/categorically labeled frames and roles; et al.(2017) showed that SG is a form of ex- sometimes, as in FrameNet, the roles can be very ponential family PCA that factorizes the matrix descriptive (e.g., the DEGREE role for the AF- of word/context cooccurrence counts (rather than FIRM OR DENY frame), while in other cases, as shifted positive PMI values). With this interpre- in PropBank, the roles can be quite general (e.g., tation, they generalize SG from matrix to tensor ARG0). Regardless of the actual schema, the roles factorization, and provide a theoretical basis for are based on thematic roles, which map a predi- modeling higher-order SG (or additional context, cate’s arguments to a semantic representation that such as morphological features of words) within a makes various semantic distinctions among the ar- word embeddings framework. 2 guments (Dowty, 1989). Dowty(1991) claims Specifically, Cotterell et al. recast higher-order that thematic role distinctions are not atomic, i.e., SG as maximizing the log-likelihood they can be deconstructed and analyzed at a lower X level. Instead of many discrete thematic roles, Xijk log p(context i j word j; feature k) (1) Dowty(1991) argues for proto-thematic roles, e.g. ijk PROTO-AGENT rather than AGENT, where dis- X exp (1|(ci wj ak)) = Xijk log P ; (2) tinctions in proto-roles are based on clusterings of 0 exp (1|(ci0 wj ak)) ijk i logical entailments. That is, PROTO-AGENTs of- ten have certain properties in common, e.g., ma- where Xijk is a cooccurrence count 3-tensor of nipulating other objects or willingly participating words j, surrounding contexts i, and features k. in an action; PROTO-PATIENTs are often changed or affected by some action. By decomposing the 3.2 Skip-Gram as n-Tensor Factorization meaning of roles into properties or expectations When factorizing an n-dimensional tensor to in- that can be reasoned about, proto-roles can be seen clude an arbitrary number of L annotations, we as including a form of vector representation within replace feature k in Equation (1) and ak in Equa- structured frame semantics. tion (2) with each annotation type l and vector αl included. Xi;j;k becomes Xi;j;l1;:::lL , representing 3 Continuous Lexical Semantics the number of times word j appeared in context i with features l through l . We maximize Word embeddings represent word meanings as el- 1 L ements of a (real-valued) vector space (Deerwester X Xi;j;l1;:::;lL log βi;j;l1;:::;lL word2vec et al., 1990). Mikolov et al.(2013)’s i;j;l1;:::;lL methods—skip-gram (SG) and continuous bag of | βi;j;l1;:::;lL / exp (1 (ci wj αl1 · · · αlL )) : 1See Petruck and de Melo(2014) for detailed descriptions on frame semantics’ contributions to applied NLP tasks. 4 Experiments 2Thematic role theory is rich, and beyond this paper’s scope (Whitehead, 1920; Davidson, 1967; Cresswell, 1973; Our end goal is to use multiple kinds of au- Kamp, 1979; Carlson, 1984). tomatically obtained, “in-the-wild” frame se- mantic parses in order to improve the seman- windowed frame 232 35.9 (triggers) tic content—specifically SPR-type information— # target words within learned lexical embeddings. We utilize ma- 404 45.7 (triggers) jority portions of the Concretely Annotated New # surrounding 232 531 (role fillers) York Times and Wikipedia corpora from Ferraro words 404 2,305 (role fillers) et al.(2014). These have been annotated with Table 1: Vocabulary sizes, in thousands, extracted from Fer- three frame semantic parses: FrameNet from Das raro et al.(2014)’s data with both the standard sliding context et al.(2010), and both FrameNet and PropBank window approach (x3) and the frame-based approach (x4). Upper numbers (Roman) are for newswire; lower numbers from Wolfe et al.(2016). In total, we use nearly (italics) are Wikipedia. For both corpora, 800 total FrameNet five million frame-annotated documents. frame types and 5100 PropBank frame types are extracted. Extracting Counts The baseline extraction we consider is a standard sliding window: for each to enable any arbitrary dimensional tensor fac- word wj seen ≥ T times, extract all words wi two torization, as described in x3.2. We learn 100- to the left and right of wj. These counts, forming a dimensional embeddings for words that appear at matrix, are then used within standard word2vec. least 100 times from 15 negative samples.4 The We also follow Cotterell et al.(2017) and augment implementation is available at https://github. the above with the signed number of tokens sepa- com/fmof/tensor-factorization. rating wi and wj, e.g., recording that wi appeared Metric We evaluate our learned (trigger) embed- two to the left of wj; these counts form a 3-tensor. dings w via QVEC (Tsvetkov et al., 2015). QVEC To turn semantic parses into tensor counts, we uses canonical correlation analysis to measure the first identify relevant information from the parses.
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]
  • Enriching an Explanatory Dictionary with Framenet and Propbank Corpus Examples
    Proceedings of eLex 2019 Enriching an Explanatory Dictionary with FrameNet and PropBank Corpus Examples Pēteris Paikens 1, Normunds Grūzītis 2, Laura Rituma 2, Gunta Nešpore 2, Viktors Lipskis 2, Lauma Pretkalniņa2, Andrejs Spektors 2 1 Latvian Information Agency LETA, Marijas street 2, Riga, Latvia 2 Institute of Mathematics and Computer Science, University of Latvia, Raina blvd. 29, Riga, Latvia E-mail: [email protected], [email protected] Abstract This paper describes ongoing work to extend an online dictionary of Latvian – Tezaurs.lv – with representative semantically annotated corpus examples according to the FrameNet and PropBank methodologies and word sense inventories. Tezaurs.lv is one of the largest open lexical resources for Latvian, combining information from more than 300 legacy dictionaries and other sources. The corpus examples are extracted from Latvian FrameNet and PropBank corpora, which are manually annotated parallel subsets of a balanced text corpus of contemporary Latvian. The proposed approach augments traditional lexicographic information with modern cross-lingually interpretable information and enables analysis of word senses from the perspective of frame semantics, which is substantially different from (complementary to) the traditional approach applied in Latvian lexicography. In cases where FrameNet and PropBank corpus evidence aligns well with the word sense split in legacy dictionaries, the frame-semantically annotated corpus examples supplement the word sense information with clarifying usage examples and commonly used semantic valence patterns. However, the annotated corpus examples often provide evidence of a different sense split, which is often more coarse-grained and, thus, may help dictionary users to cluster and comprehend a fine-grained sense split suggested by the legacy sources.
    [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]
  • Developing a Large Scale Framenet for Italian: the Iframenet Experience
    Developing a large scale FrameNet for Italian: the IFrameNet experience Roberto Basili° Silvia Brambilla§ Danilo Croce° Fabio Tamburini§ ° § Dept. of Enterprise Engineering Dept. of Classic Philology and Italian Studies University of Rome Tor Vergata University of Bologna {basili,croce}@info.uniroma2.it [email protected], [email protected] ian Portuguese, German, Spanish, Japanese, Swe- Abstract dish and Korean. All these projects are based on the idea that English. This paper presents work in pro- most of the Frames are the same among languages gress for the development of IFrameNet, a and that, thanks to this, it is possible to adopt large-scale, computationally oriented, lexi- Berkeley’s Frames and FEs and their relations, cal resource based on Fillmore’s frame se- with few changes, once all the language-specific mantics for Italian. For the development of information has been cut away (Tonelli et al. 2009, IFrameNet linguistic analysis, corpus- Tonelli 2010). processing and machine learning techniques With regard to Italian, over the past ten years are combined in order to support the semi- several research projects have been carried out at automatic development and annotation of different universities and Research Centres. In par- the resource. ticular, the ILC-CNR in Pisa (e.g. Lenci et al. 2008; Johnson and Lenci 2011), FBK in Trento (e.g. Italiano. Questo articolo presenta un work Tonelli et al. 2009, Tonelli 2010) and the Universi- in progress per lo sviluppo di IFrameNet, ty of Rome, Tor Vergata (e.g. Pennacchiotti et al. una risorsa lessicale ad ampia copertura, 2008, Basili et al.
    [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]
  • Integrating Wordnet and Framenet Using a Knowledge-Based Word Sense Disambiguation Algorithm
    Integrating WordNet and FrameNet using a knowledge-based Word Sense Disambiguation algorithm Egoitz Laparra and German Rigau IXA group. University of the Basque Country, Donostia {egoitz.laparra,german.rigau}@ehu.com Abstract coherent groupings of words belonging to the same frame. This paper presents a novel automatic approach to In that way we expect to extend the coverage of FrameNet partially integrate FrameNet and WordNet. In that (by including from WordNet closely related concepts), to way we expect to extend FrameNet coverage, to en- enrich WordNet with frame semantic information (by port- rich WordNet with frame semantic information and ing frame information to WordNet) and possibly to extend possibly to extend FrameNet to languages other than FrameNet to languages other than English (by exploiting English. The method uses a knowledge-based Word local wordnets aligned to the English WordNet). Sense Disambiguation algorithm for linking FrameNet WordNet1 [12] (hereinafter WN) is by far the most lexical units to WordNet synsets. Specifically, we ex- ploit a graph-based Word Sense Disambiguation algo- widely-used knowledge base. In fact, WN is being rithm that uses a large-scale knowledge-base derived used world-wide for anchoring different types of seman- from WordNet. We have developed and tested four tic knowledge including wordnets for languages other than additional versions of this algorithm showing a sub- English [4], domain knowledge [17] or ontologies like stantial improvement over previous results. SUMO [22] or the EuroWordNet Top Concept Ontology [3]. It contains manually coded information about En- glish nouns, verbs, adjectives and adverbs and is organized around the notion of a synset.
    [Show full text]
  • The Hebrew Framenet Project
    The Hebrew FrameNet Project Avi Hayoun, Michael Elhadad Dept. of Computer Science Ben-Gurion University Beer Sheva, Israel {hayounav,elhadad}@cs.bgu.ac.il Abstract We present the Hebrew FrameNet project, describe the development and annotation processes and enumerate the challenges we faced along the way. We have developed semi-automatic tools to help speed the annotation and data collection process. The resource currently covers 167 frames, 3,000 lexical units and about 500 fully annotated sentences. We have started training and testing automatic SRL tools on the seed data. Keywords: FrameNet, Hebrew, frame semantics, semantic resources 1. Introduction frames and their structures in natural language. Recent years have seen growing interest in the task 1.2. FrameNet in Other Languages of Semantic Role Labeling (SRL) of natural language The original FrameNet project has been adapted and text (sometimes called “shallow semantic parsing”). ported to multiple languages. The most active interna- The task is usually described as the act of identify- tional FrameNet teams include the Swedish FrameNet ing the semantic roles, which are the set of semantic (SweFN) covering close to 1,200 frames with 34K LUs properties and relationships defined over constituents (Ahlberg et al., 2014); the Japanese FrameNet (JFN) of a sentence, given a semantic context. with 565 frames, 8,500 LUs and 60K annotated ex- The creation of resources that document the realiza- ample sentences (Ohara, 2013); and FrameNet Brazil tion of semantic roles in natural language texts, such (FN-Br) covering 179 frames, 196 LUs and 12K anno- as FrameNet (Fillmore and Baker, 2010; Ruppenhofer tated sentences (Torrent and Ellsworth, 2013).
    [Show full text]
  • INF5830 Introduction to Semantic Role Labeling
    INF5830 Introduction to Semantic Role Labeling Andrey Kutuzov University of Oslo Language Technology Group With thanks to Lilja Øvrelid, Martha Palmer and Dan Jurafsky INF5830 Introduction to Semantic Role Labeling 1(36) Semantic Role Labeling INF5830 Introduction to Semantic Role Labeling 2(36) Introduction Contents Introduction Semantic roles in general PropBank: Proto-roles FrameNet: Frame Semantics Summary INF5830 Introduction to Semantic Role Labeling 2(36) Introduction Semantics I Study of meaning, expressed in language; I Morphemes, words, phrases, sentences; I Lexical semantics; I Sentence semantics; I (Pragmatics: how the context affects meaning). INF5830 Introduction to Semantic Role Labeling 3(36) Introduction Semantics I Linguistic knowledge: meaning I Meaningful or not: I Word { flick vs blick I Sentence { John swims vs John metaphorically every I Several meanings (WSD): I Words { fish I Sentence { John saw the man with the binoculars I Same meaning (semantic similarity): I Word { sofa vs couch I Sentence { John gave Hannah a gift vs John gave a gift to Hannah I Truth conditions: I All kings are male I Molybdenum conducts electricity I Entailment: I Alfred murdered the librarian I The librarian is dead I Participant roles: John is the `giver', Hannah is the `receiver' INF5830 Introduction to Semantic Role Labeling 4(36) Introduction Representing events I We want to understand the event described by these sentences: 1. IBM bought Spark 2. IBM acquired Spark 3. Spark was acquired by IBM 4. The owners of Spark sold it to IBM I Dependency parsing is insufficient. UDPipe will give us simple relations between verbs and arguments: 1. (buy, nsubj, IBM), (buy, obj, Spark) 2.
    [Show full text]
  • Using Frame Semantics in Natural Language Processing
    Using Frame Semantics in Natural Language Processing Apoorv Agarwal Daniel Bauer Owen Rambow Dept. of Computer Science Dept. of Computer Science CCLS Columbia University Columbia University Columbia University New York, NY New York, NY New York, NY [email protected] [email protected] [email protected] Abstract the notion of a social event (Agarwal et al., 2010), We summarize our experience using a particular kind of event which involves (at least) FrameNet in two rather different projects two people such that at least one of them is aware in natural language processing (NLP). of the other person. If only one person is aware We conclude that NLP can benefit from of the event, we call it Observation (OBS): for FrameNet in different ways, but we sketch example, someone is talking about someone else some problems that need to be overcome. in their absence. If both people are aware of the event, we call it Interaction (INR): for example, 1 Introduction one person is telling the other a story. Our claim We present two projects at Columbia in which we is that links in social networks are in fact made use FrameNet. In these projects, we do not de- up of social events: OBS social events give rise velop basic NLP tools for FrameNet, and we do to one-way links, and INR social events to two- not develop FramNets for new languages: we sim- way links. For more information, see (Agarwal ply use FrameNet or a FrameNet parser in an NLP and Rambow, 2010; Agarwal et al., 2013a; Agar- application.
    [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]
  • 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.
    [Show full text]