DOCSLIB.ORG

Learning from Executions for Semantic Parsing

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Learning from Executions for Semantic Parsing Learning from Executions for Semantic Parsing Bailin Wang, Mirella Lapata and Ivan Titov Institute for Language, Cognition and Computation School of Informatics, University of Edinburgh [email protected], {mlap, ititov}@inf.ed.ac.uk Abstract list all 3 star rated thai restaurants Semantic parsing aims at translating natu- Program Candidates Gold Exe ral language (NL) utterances onto machine- select restaurant where star_rating = thai 77 interpretable programs, which can be executed select restaurant where cuisine > 3 77 against a real-world environment. The ex- select where 73 pensive annotation of utterance-program pairs restaurant star_rating = 3 has long been acknowledged as a major bot- select restaurant where star_rating = 3 tleneck for the deployment of contemporary and cuisine = thai 33 neural models to real-life applications. In this work, we focus on the task of semi-supervised Figure 1: Candidate programs for an utterance can be learning where a limited amount of annotated classified by executability (Exe); note that the gold pro- data is available together with many unlabeled gram is always in the set of executable programs. We NL utterances. Based on the observation that propose to ultilize the weak yet freely available signal programs which correspond to NL utterances of executablility for learning. must be always executable, we propose to en- courage a parser to generate executable pro- grams for unlabeled utterances. Due to the ronment against which they are executed (e.g., a large search space of executable programs, knowledge base, a relational database). An alter- conventional methods that use approximations native to annotation is to collect answers (or deno- based on beam-search such as self-training and tations) of programs, rather than programs them- top-k marginal likelihood training, do not per- selves (Liang et al., 2013; Berant et al., 2013). In form as well. Instead, we view the problem this work, we focus on the more extreme setting of learning from executions from the perspec- tive of posterior regularization and propose a where there are no annotations available for a large set of new training objectives. Experimen- number of utterances. This setting resembles a tal results on OVERNIGHT and GEOQUERY common real-life scenario where massive numbers show that our new objectives outperform con- of user utterances can be collected when deploying ventional methods, bridging the gap between a semantic parser (Iyer et al., 2017). Effectively semi-supervised and supervised learning. utilizing the unlabeled data makes it possible for 1 Introduction a semantic parser to improve over time without human involvement. arXiv:2104.05819v1 [cs.CL] 12 Apr 2021 Semantic parsing is the task of mapping nat- Our key observation is that not all candidate pro- ural language (NL) utterances to meaning repre- grams for an utterance will be semantically valid. sentations (aka programs) that can be executed This implies that only some candidate programs against a real-world environment such as a knowl- can be executed and obtain non-empty execution edge base or a relational database. While neural results.1 As illustrated in Figure1, executability is sequence-to-sequence models (Dong and Lapata, a weak signal that can differentiate between seman- 2016; Jia and Liang, 2016a) have achieved much tically valid and invalid programs. On unlabeled success in this task in recent years, they usually re- utterances, we can encourage a parser to only focus quire a large amount of labeled data (i.e., utterance- on executable programs ignoring non-executable program pairs) for training. However, annotat- ones. Moreover, the executability of a program ing utterances with programs is expensive as it 1In the rest of this paper, we extend the meaning of ‘exe- requires expert knowledge of meaning representa- cutability’, and use it to refer to the case where a program is tions (e.g., lambda calculus, SQLs) and the envi- executable and obtains non-empty results. can be obtained from an executor for free without data as unlabeled. Empirical results show that our requiring human effort. Executability has previ- method can substantially boost the performance of ously been used to guide the decoding of a semantic a parser, trained only on labeled data, by utilizing parser (Wang et al., 2018). We take a step further a large amount of unlabeled data. to directly use this weak signal for learning from Our contributions are summarized as follows: unlabeled utterances. • We show how to exploit unlabeled utterances To learn from executability, we resort to by taking advantage of their executability. marginal likelihood training, i.e., maximizing the • To better learn from executability, we propose marginal likelihood of all executable programs for a set of new objectives based on posterior reg- an unlabeled NL utterance. However, the space ularization. of all possible programs is exponentially large, as well as the space of executable ones. Hence, simply • Our method can help a base parser achieve marginalizing over all executable programs is in- substantially better performance by utilizing tractable. Typical approximations use beam search unlabeled data. to retrieve a handful of (‘seen’) programs, which Our code, datasets, and splits are publicly avail- are used to approximate the entire space. Using able at https://github.com/berlino/ such approximations can lead to optimization get- tensor2struct-public. ting trapped in undesirable local minima. For ex- ample, we observe that encouraging a model to 2 Related Work exploit seen executable programs hinders explo- ration and reinforces the preference for shorter pro- Semi-Supervised Semantic Parsing In the con- grams, as discussed in Section 6.3. This happens text of semantic parsing, semi-supervised models because shorter programs are both more likely to using limited amounts of parallel data and large be among ‘seen’ programs (probably due to using amounts of unlabeled data treat either utterances locally-normalized autoregressive modeling) and or programs as discrete latent variables and in- more likely to be executable. To alleviate these duce them in the framework of generative mod- issues, we derive three new alternative objectives, els (Kociskýˇ et al., 2016; Yin et al., 2018). A chal- relying on a new interpretation of marginal likeli- lenge with these methods is that (combinatorially) hood training from the perspective of posterior reg- complex discrete variables make optimization very ularization. Our proposed objectives encode two hard, even with the help of variational inference. In kinds of inductive biases: 1) discouraging seen this work, we seek to directly constrain the discrimi- non-executable programs, which plays a similar native parser with signals obtained from executions. role to encouraging seen executable ones but does Our method can potentially be integrated into these not share its drawback of hindering exploration; generative models to regularize discrete variables. 2) encouraging sparsity among executable pro- (Underspecified) Sequence-Level Rewards grams, which encourages a parser to only focus on There have been attempts in recent years to a subset of executable programs by softly injecting integrate sequence-level rewards into sequence- a sparsity constraint. This is desirable, as there are to-sequence training as a way of accommodating only one or few correct programs for each utter- task-specific objectives. For example, BLEU can ance (see Figure1), and an accurate parser should be optimized for coherent text generation (Bosse- assign probability mass only to this subset. We col- lut et al., 2018) and machine translation (Wu lectively call these objectives X-PR, as a shorthand et al., 2018) via reinforcement learning or beam- for Execution-guided Posterior Regularization. search (Wiseman and Rush, 2016). In this work, We conduct experiments on two representative we resort to marginal likelihood training to exploit semantic parsing tasks: text-to-LF (logical form) binary executability rewards for semantic parsing parsing over a knowledge base and text-to-SQL (i.e., whether a program is executable or not), (Zelle and Mooney, 1996) parsing over a relational which has been shown to be more effective than database. Concretely, we evaluate our methods on REINFORCE (Guu et al., 2017). the OVERNIGHT (Wang et al., 2015a) and GEO- More importantly, our binary reward is under- QUERY datasets. We simulate the semi-supervised specified, i.e., there exist many spurious programs learning setting by treating 70% of the training that enjoy the same reward as the gold program. This issue of learning from underspecified re- grams. Specifically, they are defined as follows: wards underlies many weakly-supervised tasks, e.g., learning from denotations (Liang et al., 2013; Lsup(x; y) = − log p(yjx; θ) (2) X Berant et al., 2013), weakly supervised question Lunsup(x) = − log R(y)p(yjx; θ) (3) answering (Min et al., 2019). Previous work tried y to model latent alignments (Wang et al., 2019) be- where R(y) is a binary reward function that returns tween NL and programs to alleviate this issue. In 1 if y is executable and 0 otherwise. In practice, this this work, we take an orthogonal direction and pro- function is implemented by running a task-specific pose several training objectives that alleviate the executor, e.g., a SQL executor. impact of spurious programs. Another alternative to unsupervised loss is RE- INFORCE (Sutton et al., 1999), i.e., maximize the Execution for Semantic Parsing Execution has expected R(y) with respect to p(yjx; θ). However, been utilized in semantic parsing (Wang et al., as presented in Guu et al.(2017), this objective 2018) and the related area of program synthe- usually underperforms MML, which is consistent sis (Chen et al., 2019). These approaches ex- with our initial experiments.2 ploit the execution of partial programs to guide the search for plausible complete programs. Al- 3.2 Self-Training and Top-K MML though partial execution is feasible for SQL-style MML in Equation (3) requires marginalizing programs, it cannot be trivially extended to general over all executable programs which is intractable.
Load more

Recommended publications
  • Semantic Parsing for Single-Relation Question Answering
    Semantic Parsing for Single-Relation Question Answering Wen-tau Yih Xiaodong He Christopher Meek Microsoft Research Redmond, WA 98052, USA scottyih,xiaohe,meek @microsoft.com { } Abstract the same question. That is to say that the problem of mapping from a question to a particular relation We develop a semantic parsing framework and entity in the KB is non-trivial. based on semantic similarity for open do- In this paper, we propose a semantic parsing main question answering (QA). We focus framework tailored to single-relation questions. on single-relation questions and decom- At the core of our approach is a novel semantic pose each question into an entity men- similarity model using convolutional neural net- tion and a relation pattern. Using convo- works. Leveraging the question paraphrase data lutional neural network models, we mea- mined from the WikiAnswers corpus by Fader et sure the similarity of entity mentions with al. (2013), we train two semantic similarity mod- entities in the knowledge base (KB) and els: one links a mention from the question to an the similarity of relation patterns and re- entity in the KB and the other maps a relation pat- lations in the KB. We score relational tern to a relation. The answer to the question can triples in the KB using these measures thus be derived by finding the relation–entity triple and select the top scoring relational triple r(e1, e2) in the KB and returning the entity not to answer the question. When evaluated mentioned in the question. By using a general se- on an open-domain QA task, our method mantic similarity model to match patterns and re- achieves higher precision across different lations, as well as mentions and entities, our sys- recall points compared to the previous ap- tem outperforms the existing rule learning system, proach, and can improve F1 by 7 points.
    [Show full text]
  • Semantic Parsing with Semi-Supervised Sequential Autoencoders
    Semantic Parsing with Semi-Supervised Sequential Autoencoders Toma´sˇ Kociskˇ y´†‡ Gabor´ Melis† Edward Grefenstette† Chris Dyer† Wang Ling† Phil Blunsom†‡ Karl Moritz Hermann† †Google DeepMind ‡University of Oxford tkocisky,melisgl,etg,cdyer,lingwang,pblunsom,kmh @google.com { } Abstract This is common in situations where we encode nat- ural language into a logical form governed by some We present a novel semi-supervised approach grammar or database. for sequence transduction and apply it to se- mantic parsing. The unsupervised component While such an autoencoder could in principle is based on a generative model in which latent be constructed by stacking two sequence transduc- sentences generate the unpaired logical forms. ers, modelling the latent variable as a series of dis- We apply this method to a number of semantic crete symbols drawn from multinomial distributions parsing tasks focusing on domains with lim- ited access to labelled training data and ex- creates serious computational challenges, as it re- tend those datasets with synthetically gener- quires marginalising over the space of latent se- ated logical forms. quences Σx∗. To avoid this intractable marginalisa- tion, we introduce a novel differentiable alternative for draws from a softmax which can be used with 1 Introduction the reparametrisation trick of Kingma and Welling (2014). Rather than drawing a discrete symbol in Neural approaches, in particular attention-based Σx from a softmax, we draw a distribution over sym- sequence-to-sequence models, have shown great bols from a logistic-normal distribution at each time promise and obtained state-of-the-art performance step. These serve as continuous relaxations of dis- for sequence transduction tasks including machine crete samples, providing a differentiable estimator translation (Bahdanau et al., 2015), syntactic con- of the expected reconstruction log likelihood.
    [Show full text]
  • SQL Generation from Natural Language: a Sequence-To- Sequence Model Powered by the Transformers Architecture and Association Rules
    Journal of Computer Science Original Research Paper SQL Generation from Natural Language: A Sequence-to- Sequence Model Powered by the Transformers Architecture and Association Rules 1,2Youssef Mellah, 1Abdelkader Rhouati, 1El Hassane Ettifouri, 2Toumi Bouchentouf and 2Mohammed Ghaouth Belkasmi 1NovyLab Research, Novelis, Paris, France 2Mohammed First University Oujda, National School of Applied Sciences, LaRSA/SmartICT Lab, Oujda, Morocco Article history Abstract: Using Natural Language (NL) to interacting with relational Received: 12-03-2021 databases allows users from any background to easily query and analyze Revised: 21-04-2021 large amounts of data. This requires a system that understands user Accepted: 28-04-2021 questions and automatically converts them into structured query language such as SQL. The best performing Text-to-SQL systems use supervised Corresponding Author: Youssef Mellah learning (usually formulated as a classification problem) by approaching NovyLab Research, Novelis, this task as a sketch-based slot-filling problem, or by first converting Paris, France questions into an Intermediate Logical Form (ILF) then convert it to the And corresponding SQL query. However, non-supervised modeling that LARSA Laboratory, ENSAO, directly converts questions to SQL queries has proven more difficult. In Mohammed First University, this sense, we propose an approach to directly translate NL questions into Oujda, Morocco SQL statements. In this study, we present a Sequence-to-Sequence Email: [email protected] (Seq2Seq) parsing model for the NL to SQL task, powered by the Transformers Architecture exploring the two Language Models (LM): Text-To-Text Transfer Transformer (T5) and the Multilingual pre-trained Text-To-Text Transformer (mT5). Besides, we adopt the transformation- based learning algorithm to update the aggregation predictions based on association rules.
    [Show full text]
  • Semantic Parsing with Neural Hybrid Trees
    Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence (AAAI-17) Semantic Parsing with Neural Hybrid Trees Raymond Hendy Susanto, Wei Lu Singapore University of Technology and Design 8 Somapah Road, Singapore 487372 {raymond susanto, luwei}@sutd.edu.sg Abstract QUERY : answer(STATE) We propose a neural graphical model for parsing natural lan- STATE : exclude(STATE, STATE) guage sentences into their logical representations. The graph- ical model is based on hybrid tree structures that jointly rep- : state( ) : next to( ) resent both sentences and semantics. Learning and decoding STATE all STATE STATE are done using efficient dynamic programming algorithms. The model is trained under a discriminative setting, which STATE : stateid(STATENAME) allows us to incorporate a rich set of features. Hybrid tree structures have shown to achieve state-of-the-art results on STATENAME :(tx ) standard semantic parsing datasets. In this work, we propose a novel model that incorporates a rich, nonlinear featurization Which states do not border Texas ? by a feedforward neural network. The error signals are com- puted with respect to the conditional random fields (CRFs) objective using an inside-outside algorithm, which are then Figure 1: An example tree-structured semantic representa- backpropagated to the neural network. We demonstrate that tion and its corresponding natural language sentence. by combining the strengths of the exact global inference in the hybrid tree models and the power of neural networks to extract high level features, our model is able to achieve new explicit labelings between words and meaning representa- state-of-the-art results on standard benchmark datasets across tion tokens are not available in the training data.
    [Show full text]
  • Learning Logical Representations from Natural Languages with Weak Supervision and Back-Translation
    Learning Logical Representations from Natural Languages with Weak Supervision and Back-Translation Kaylin Hagopiany∗ Qing Wangy Georgia Institute of Technology IBM Research Atlanta, GA, USA Yorktown Heights, NY, USA [email protected] [email protected] Tengfei Ma Yupeng Gao Lingfei Wu IBM Research AI IBM Research AI IBM Research AI Yorktown Heights, NY, USA Yorktown Heights, NY, USA Yorktown Heights, NY, USA [email protected] [email protected] [email protected] Abstract Semantic parsing with strong supervision requires a large parallel corpus for train- ing, whose production may be impractical due to the excessive human labour needed for annotation. Recently, weakly supervised approaches based on reinforce- ment learning have been studied, wherein only denotations, but not the logical forms, are available. However, reinforcement learning approaches based on, e.g., policy gradient training, suffer reward sparsity and progress slowly. In this work, we explore enriching the policy gradient framework with a back-translation mecha- nism that boosts the reward signal. In this vein, we jointly optimize the expected reward of obtaining the correct denotation and the likelihood of reconstructing the natural language utterance. Initial experimental results suggest the proposed method slightly outperforms a state-of-the-art semantic parser trained with policy gradient. Introduction Semantic parsing is the task of converting natural language utterances to logical representations that a machine can understand and execute as problem solving steps. There has recently been a surge of interest in developing machine learning methods for semantic parsing, owing to the creation of large- scale text corpora. There has also been a proliferation of logical representations used in this setting, including logical forms [28], domain specific languages [26], python code [24], SQL queries [31], and bash commands [20].
    [Show full text]
  • A Sequence to Sequence Architecture for Task-Oriented Semantic Parsing
    Don’t Parse, Generate! A Sequence to Sequence Architecture for Task-Oriented Semantic Parsing Subendhu Rongali∗ Luca Soldaini University of Massachusetts Amherst Amazon Alexa Search Amherst, MA, USA Manhattan Beach, CA, USA [email protected] [email protected] Emilio Monti Wael Hamza Amazon Alexa Amazon Alexa AI Cambridge, UK New York, NY, USA [email protected] [email protected] ABSTRACT ArtistName Virtual assistants such as Amazon Alexa, Apple Siri, and Google Play the song don't stop believin by Journey Assistant often rely on a semantic parsing component to understand PlaySongIntent SongName which action(s) to execute for an utterance spoken by its users. Traditionally, rule-based or statistical slot-filling systems have been Intent: PlaySongIntent used to parse “simple” queries; that is, queries that contain a single Slots: SongName(don't stop believin), action and can be decomposed into a set of non-overlapping entities. ArtistName(Journey) More recently, shift-reduce parsers have been proposed to process more complex utterances. These methods, while powerful, impose specific limitations on the type of queries that can be parsed; namely, Figure 1: Semantic parsing of a “simple” query. Simple they require a query to be representable as a parse tree. queries define single action (intent) and can be decomposed In this work, we propose a unified architecture based on Se- into a set of non-overlapping entities (slots). quence to Sequence models and Pointer Generator Network to handle both simple and complex queries. Unlike other works, our 1 INTRODUCTION approach does not impose any restriction on the semantic parse Adoption of intelligent voice assistants such as Amazon Alexa, schema.
    [Show full text]
  • Neural Semantic Parsing Graham Neubig
    CS11-747 Neural Networks for NLP Neural Semantic Parsing Graham Neubig Site https://phontron.com/class/nn4nlp2018/ Tree Structures of Syntax • Dependency: focus on relations between words ROOT I saw a girl with a telescope • Phrase structure: focus on the structure of the sentence S VP PP NP NP PRP VBD DT NN IN DT NN I saw a girl with a telescope Representations of Semantics • Syntax only gives us the sentence structure • We would like to know what the sentence really means • Specifically, in an grounded and operationalizable way, so a machine can • Answer questions • Follow commands • etc. Meaning Representations • Special-purpose representations: designed for a specific task • General-purpose representations: designed to be useful for just about anything • Shallow representations: designed to only capture part of the meaning (for expediency) Parsing to Special-purpose Meaning Representations Example Special-purpose Representations • A database query language for sentence understanding • A robot command and control language • Source code in a language such as Python (?) Example Query Tasks • Geoquery: Parsing to Prolog queries over small database (Zelle and Mooney 1996) • Free917: Parsing to Freebase query language (Cai and Yates 2013) • Many others: WebQuestions, WikiTables, etc. Example Command and Control Tasks • Robocup: Robot command and control (Wong and Mooney 2006) • If this then that: Commands to smartphone interfaces (Quirk et al. 2015) Example Code Generation Tasks • Hearthstone cards (Ling et al. 2015) • Django commands (Oda
    [Show full text]
  • Practical Semantic Parsing for Spoken Language Understanding
    Practical Semantic Parsing for Spoken Language Understanding Marco Damonte∗ Rahul Goel Tagyoung Chung University of Edinburgh Amazon Alexa AI [email protected] fgoerahul,[email protected] Abstract the intent’s slots. This architecture is popular due to its simplicity and robustness. On the other hand, Executable semantic parsing is the task of con- verting natural language utterances into logi- Q&A, which need systems to produce more com- cal forms that can be directly used as queries plex structures such as trees and graphs, requires a to get a response. We build a transfer learn- more comprehensive understanding of human lan- ing framework for executable semantic pars- guage. ing. We show that the framework is effec- One possible system that can handle such a task tive for Question Answering (Q&A) as well as is an executable semantic parser (Liang, 2013; for Spoken Language Understanding (SLU). Kate et al., 2005). Given a user utterance, an exe- We further investigate the case where a parser on a new domain can be learned by exploit- cutable semantic parser can generate tree or graph ing data on other domains, either via multi- structures that represent logical forms that can be task learning between the target domain and used to query a knowledge base or database. In an auxiliary domain or via pre-training on the this work, we propose executable semantic pars- auxiliary domain and fine-tuning on the target ing as a common framework for both uses cases domain. With either flavor of transfer learn- by framing SLU as executable semantic parsing ing, we are able to improve performance on that unifies the two use cases.
    [Show full text]
  • Arxiv:2009.14116V1 [Cs.CL] 29 Sep 2020
    A Survey on Semantic Parsing from the Perspective of Compositionality Pawan Kumar Srikanta Bedathur Indian Institute of Technology Delhi, Indian Institute of Technology Delhi, New Delhi, Hauzkhas-110016 New Delhi, Hauzkhas-110016 [email protected] [email protected] Abstract compositionality: evident by the choice of the for- mal language and the construction mechanism and Different from previous surveys in seman- lexical variation: grounding the words/phrases to tic parsing (Kamath and Das, 2018) and appropriate KB entity/relation. knowledge base question answering(KBQA) (Chakraborty et al., 2019; Zhu et al., 2019; Language Compositionality: According to Hoffner¨ et al., 2017) we try to takes a dif- Frege’s principle of compositionality: “the ferent perspective on the study of semantic meaning of a whole is a function of the meaning parsing. Specifically, we will focus on (a) meaning composition from syntactical struc- of the parts. The study of formal semantics ture (Partee, 1975), and (b) the ability of for natural language to appropriately represent semantic parsers to handle lexical variation sentence meaning has a long history (Szab, 2017). given the context of a knowledge base (KB). Most semantic formalism follow the tradition of In the following section after an introduction Montague’s grammar (Partee, 1975) i.e. there is of the field of semantic parsing and its uses a one-to-one correspondence between syntax and in KBQA, we will describe meaning represen- semantics e.g. CCG (Zettlemoyer and Collins, tation using grammar formalism CCG (Steed- 2005). We will not be delving into the concept of man, 1996).
    [Show full text]
  • Learning Contextual Representations for Semantic Parsing with Generation-Augmented Pre-Training
    Learning Contextual Representations for Semantic Parsing with Generation-Augmented Pre-Training Peng Shi1*, Patrick Ng2, Zhiguo Wang2, Henghui Zhu2, Alexander Hanbo Li2, Jun Wang2, Cicero Nogueira dos Santos2, Bing Xiang2 1 University of Waterloo, 2 AWS AI Labs [email protected], fpatricng,zhiguow,henghui,hanboli,juwanga,cicnog,[email protected] Abstract Pain Point 1: Fail to match and detect the column mentions. Utterance: Which professionals live in a city containing the Most recently, there has been significant interest in learn- substring ’West’? List his or her role, street, city and state. ing contextual representations for various NLP tasks, by Prediction: SELECT role code, street, state leveraging large scale text corpora to train large neural lan- FROM Professionals WHERE city LIKE ’%West%’ guage models with self-supervised learning objectives, such Error: Missing column city in SELECT clause. as Masked Language Model (MLM). However, based on a pi- lot study, we observe three issues of existing general-purpose Pain Point 2: Fail to infer columns based on cell values. language models when they are applied to text-to-SQL se- Utterance: Give the average life expectancy for countries in mantic parsers: fail to detect column mentions in the utter- Africa which are republics? ances, fail to infer column mentions from cell values, and fail Prediction: SELECT Avg(LifeExpectancy) FROM to compose complex SQL queries. To mitigate these issues, country WHERE Continent = ’Africa’ we present a model pre-training framework, Generation- Error: Missing GovernmentForm = ’Republic’. Augmented Pre-training (GAP), that jointly learns repre- Pain Point 3 sentations of natural language utterances and table schemas : Fail to compose complex target SQL.
    [Show full text]
  • On the Use of Parsing for Named Entity Recognition
    applied sciences Review On the Use of Parsing for Named Entity Recognition Miguel A. Alonso * , Carlos Gómez-Rodríguez and Jesús Vilares Grupo LyS, Departamento de Ciencias da Computación e Tecnoloxías da Información, Universidade da Coruña and CITIC, 15071 A Coruña, Spain; [email protected] (C.G.-R.); [email protected] (J.V.) * Correspondence: [email protected] Abstract: Parsing is a core natural language processing technique that can be used to obtain the structure underlying sentences in human languages. Named entity recognition (NER) is the task of identifying the entities that appear in a text. NER is a challenging natural language processing task that is essential to extract knowledge from texts in multiple domains, ranging from financial to medical. It is intuitive that the structure of a text can be helpful to determine whether or not a certain portion of it is an entity and if so, to establish its concrete limits. However, parsing has been a relatively little-used technique in NER systems, since most of them have chosen to consider shallow approaches to deal with text. In this work, we study the characteristics of NER, a task that is far from being solved despite its long history; we analyze the latest advances in parsing that make its use advisable in NER settings; we review the different approaches to NER that make use of syntactic information; and we propose a new way of using parsing in NER based on casting parsing itself as a sequence labeling task. Keywords: natural language processing; named entity recognition; parsing; sequence labeling Citation: Alonso, M.A; Gómez- 1.
    [Show full text]
  • Neural Semantic Parsing for Syntax-Aware Code Generation
    Neural Semantic Parsing for Syntax-Aware Code Generation Masterarbeit zur Erlangung des Master-Grades Master of Science im Studiengang Tech- nische Informatik an der Fakultät für Informations-, Medien- und Elektrotechnik der Technischen Hochschule Köln vorgelegt von: Artur Baranowski Matrikel-Nr.: 11130536 Adresse: Heidestraße 144 51147 Köln Erstprüfer: Prof. Dr. Hubert Randerath Zweitprüfer: Prof. Dr. Nico Hochgeschwender Köln, 23. September 2020 Abstract I Abstract The task of mapping natural language expressions to logical forms is referred to as semantic parsing. The syntax of logical forms that are based on programming or query languages, such as Python or SQL, is defined by a formal grammar. In this thesis, we present an efficient neural semantic parser that exploits the underlying grammar of logical forms to enforce well-formed expressions. We use an encoder-decoder model for sequence prediction. Syntactically valid programs are guaranteed by means of a bottom- up shift-reduce parser, that keeps track of the set of viable tokens at each decoding step. We show that the proposed model outperforms the standard encoder-decoder model across datasets and is competitive with comparable grammar-guided semantic parsing approaches. Keywords: semantic parsing, machine learning, natural language processing, code generation, formal grammars, parsing theory Statutory Declaration II Statutory Declaration I herewith declare that I have composed the present thesis myself and without use of any other than the cited sources and aids. Sentences or parts of sentences quoted literally are marked as such; other references with regard to the statement and scope are indicated by full details of the publications concerned. The thesis in the same or similar form has not been submitted to any examination body and has not been published.
    [Show full text]