Neural Parser Research Articles

Neural Shape Parsers for Constructive Solid Geometry.

Constructive solid geometry (CSG) is a geometric modeling technique that defines complex shapes by recursively applying boolean operations on primitives such as spheres and cylinders. We present CSGNet, a deep network architecture that takes as input a 2D or 3D shape and outputs a CSG program that models it. Parsing shapes into CSG programs is desirable as it yields a compact and interpretable generative model. However, the task is challenging since the space of primitives and their combinations can be prohibitively large. CSGNet uses a convolutional encoder and recurrent decoder based on deep networks to map shapes to modeling instructions in a feed-forward manner and is significantly faster than bottom-up approaches. We investigate two architectures for this task-a vanilla encoder (CNN) - decoder (RNN) and another architecture that augments the encoder with an explicit memory module based on the program execution stack. The stack augmentation improves the reconstruction quality of the generated shape and learning efficiency. Our approach is also more effective as a shape primitive detector compared to a state-of-the-art object detector. Finally, we demonstrate CSGNet can be trained on novel datasets without program annotations through policy gradient techniques.

Dependency Grammar Induction with a Neural Variational Transition-Based Parser

Dependency grammar induction is the task of learning dependency syntax without annotated training data. Traditional graph-based models with global inference achieve state-ofthe-art results on this task but they require O(n3) run time. Transition-based models enable faster inference with O(n) time complexity, but their performance still lags behind. In this work, we propose a neural transition-based parser for dependency grammar induction, whose inference procedure utilizes rich neural features with O(n) time complexity. We train the parser with an integration of variational inference, posterior regularization and variance reduction techniques. The resulting framework outperforms previous unsupervised transition-based dependency parsers and achieves performance comparable to graph-based models, both on the English Penn Treebank and on the Universal Dependency Treebank. In an empirical comparison, we show that our approach substantially increases parsing speed over graphbased models.

Parsing Chinese Sentences with Grammatical Relations

We report our work on building linguistic resources and data-driven parsers in the grammatical relation (GR) analysis for Mandarin Chinese. Chinese, as an analytic language, encodes grammatical information in a highly configurational rather than morphological way. Accordingly, it is possible and reasonable to represent almost all grammatical relations as bilexical dependencies. In this work, we propose to represent grammatical information using general directed dependency graphs. Both only-local and rich long-distance dependencies are explicitly represented. To create high-quality annotations, we take advantage of an existing TreeBank, namely, Chinese TreeBank (CTB), which is grounded on the Government and Binding theory. We define a set of linguistic rules to explore CTB’s implicit phrase structural information and build deep dependency graphs. The reliability of this linguistically motivated GR extraction procedure is highlighted by manual evaluation. Based on the converted corpus, data-driven, including graph- and transition-based, models are explored for Chinese GR parsing. For graph-based parsing, a new perspective, graph merging, is proposed for building flexible dependency graphs: constructing complex graphs via constructing simple subgraphs. Two key problems are discussed in this perspective: (1) how to decompose a complex graph into simple subgraphs, and (2) how to combine subgraphs into a coherent complex graph. For transition-based parsing, we introduce a neural parser based on a list-based transition system. We also discuss several other key problems, including dynamic oracle and beam search for neural transition-based parsing. Evaluation gauges how successful GR parsing for Chinese can be by applying data-driven models. The empirical analysis suggests several directions for future study.

Proactive neural processing of native and non-native speech

The impact of attention and language-experience in neural speech processing is typically assessed using sentences, words, and syllables that, when presented in isolation, may not engage the cortical speech network as well as more realistic continuous speech. Here, we explore the neuromodulatory effects of attention and language experience using continuous speech. We recorded electroencephalographic responses from native speakers of English and late Chinese-English bilinguals while listening to a story recorded in English. The story was mixed with a tone sequence, and listeners were instructed to focus either on the speech (attended speech condition) or the tone sequence (ignored speech condition). We used the multivariate temporal response function and the accuracy of a machine-learning based brain-to-speech decoder to quantify differences in cortical entrainment and speech-sound category processing. Our analyses revealed a more robust, context-independent neural encoding of speech-sound categories when they were attended and native. Interestingly, while cortical entrainment to speech was also enhanced by attention, the enhancement was stronger among non-native speakers. Our results suggest that while listeners can manage attention to improve the neural parsing of continuous speech via cortical entrainment, the benefits of attention for speech-sound category processing can be attenuated when the sounds are not native.The impact of attention and language-experience in neural speech processing is typically assessed using sentences, words, and syllables that, when presented in isolation, may not engage the cortical speech network as well as more realistic continuous speech. Here, we explore the neuromodulatory effects of attention and language experience using continuous speech. We recorded electroencephalographic responses from native speakers of English and late Chinese-English bilinguals while listening to a story recorded in English. The story was mixed with a tone sequence, and listeners were instructed to focus either on the speech (attended speech condition) or the tone sequence (ignored speech condition). We used the multivariate temporal response function and the accuracy of a machine-learning based brain-to-speech decoder to quantify differences in cortical entrainment and speech-sound category processing. Our analyses revealed a more robust, context-independent neural encoding of speech-sound categories when t...

Learning an Executable Neural Semantic Parser

This article describes a neural semantic parser that maps natural language utterances onto logical forms that can be executed against a task-specific environment, such as a knowledge base or a database, to produce a response. The parser generates tree-structured logical forms with a transition-based approach, combining a generic tree-generation algorithm with domain-general grammar defined by the logical language. The generation process is modeled by structured recurrent neural networks, which provide a rich encoding of the sentential context and generation history for making predictions. To tackle mismatches between natural language and logical form tokens, various attention mechanisms are explored. Finally, we consider different training settings for the neural semantic parser, including fully supervised training where annotated logical forms are given, weakly supervised training where denotations are provided, and distant supervision where only unlabeled sentences and a knowledge base are available. Experiments across a wide range of data sets demonstrate the effectiveness of our parser.

Analysis of Experiments on Statistical and Neural Parsing for a Morphologically Rich and Free Word Order Language Urdu

This article presents an analysis of experiments with statistical and neural parsing techniques for Urdu, a widely spoken South Asian language. We demonstrate state of the art constituency parsing results for an Urdu treebank. Urdu is a morphologically rich and is characterized by free word order. Language representation (e.g. input type, lemmatization, word clusters), part of speech tag set, phrase labels and the size of a training corpus are crucial for parsing such languages. In this article, probabilistic context-free grammars, data-oriented parsing, and recursive neural network based models have been experimented with several linguistic features which show improvements in the parsing results. Features include syntactic sub-categorization of POS tags, empirically learned horizontal and vertical markovizations and lexical head words. These features enable dependency information for case markers and add phrasal and lexical context to the parse trees. The data-oriented parsing and recursive neural network model give an f-score of 87.1 by considering gold POS tags in the test set, on textual input, they show a performance with f-scores of 83.4 and 84.2, respectively. To overcome the issue of data sparsity due to the morphological richness, lemmatization and unsupervised word clustering have been performed. A treebank should cover most probable word orders of the language so that models can learn various orders accurately. To analyze the order coverage of the treebank and learning capability of different parsers, a test set has been prepared conditioning different word orders. This test set is evaluated with the best performing parsing models and with gold POS tags, f-scores are above 90 and on textual input, the average f-score is 87.6.

Exploring Neural Methods for Parsing Discourse Representation Structures

Neural methods have had several recent successes in semantic parsing, though they have yet to face the challenge of producing meaning representations based on formal semantics. We present a sequence-to-sequence neural semantic parser that is able to produce Discourse Representation Structures (DRSs) for English sentences with high accuracy, outperforming traditional DRS parsers. To facilitate the learning of the output, we represent DRSs as a sequence of flat clauses and introduce a method to verify that produced DRSs are well-formed and interpretable. We compare models using characters and words as input and see (somewhat surprisingly) that the former performs better than the latter. We show that eliminating variable names from the output using De Bruijn indices increases parser performance. Adding silver training data boosts performance even further.

Surface Statistics of an Unknown Language Indicate How to Parse It

We introduce a novel framework for delexicalized dependency parsing in a new language. We show that useful features of the target language can be extracted automatically from an unparsed corpus, which consists only of gold part-of-speech (POS) sequences. Providing these features to our neural parser enables it to parse sequences like those in the corpus. Strikingly, our system has no supervision in the target language. Rather, it is a multilingual system that is trained end-to-end on a variety of other languages, so it learns a feature extractor that works well. We show experimentally across multiple languages: (1) Features computed from the unparsed corpus improve parsing accuracy. (2) Including thousands of synthetic languages in the training yields further improvement. (3) Despite being computed from unparsed corpora, our learned task-specific features beat previous work’s interpretable typological features that require parsed corpora or expert categorization of the language. Our best method improved attachment scores on held-out test languages by an average of 5.6 percentage points over past work that does not inspect the unparsed data (McDonald et al., 2011), and by 20.7 points over past “grammar induction” work that does not use training languages (Naseem et al., 2010).

From Characters to Time Intervals: New Paradigms for Evaluation and Neural Parsing of Time Normalizations.

This paper presents the first model for time normalization trained on the SCATE corpus. In the SCATE schema, time expressions are annotated as a semantic composition of time entities. This novel schema favors machine learning approaches, as it can be viewed as a semantic parsing task. In this work, we propose a character level multi-output neural network that outperforms previous state-of-the-art built on the TimeML schema. To compare predictions of systems that follow both SCATE and TimeML, we present a new scoring metric for time intervals. We also apply this new metric to carry out a comparative analysis of the annotations of both schemes in the same corpus.

Neural Character-level Dependency Parsing for Chinese

This paper presents a truly full character-level neural dependency parser together with a newly released character-level dependency treebank for Chinese, which has suffered a lot from the dilemma of defining word or not to model character interactions. Integrating full character-level dependencies with character embedding and human annotated character-level part-of-speech and dependency labels for the first time, we show an extra performance enhancement from the evaluation on Chinese Penn Treebank and SJTU (Shanghai Jiao Tong University) Chinese Character Dependency Treebank and the potential of better understanding deeper structure of Chinese sentences.

Attention-based Belief or Disbelief Feature Extraction for Dependency Parsing

Existing neural dependency parsers usually encode each word in a sentence with bi-directional LSTMs, and estimate the score of an arc from the LSTM representations of the head and the modifier, possibly missing relevant context information for the arc being considered. In this study, we propose a neural feature extraction method that learns to extract arc-specific features. We apply a neural network-based attention method to collect evidences for and against each possible head-modifier pair, with which our model computes certainty scores of belief and disbelief, and determines the final arc score by subtracting the score of disbelief from the one of belief. By explicitly introducing two kinds of evidences, the arc candidates can compete against each other based on more relevant information, especially for the cases where they share the same head or modifier. It makes possible to better discriminate two or more competing arcs by presenting their rivals (disbelief evidence). Experiments on various datasets show that our arc-specific feature extraction mechanism significantly improves the performance of bi-directional LSTM-based models by explicitly modeling long-distance dependencies. For both English and Chinese, the proposed model achieve a higher accuracy on dependency parsing task than most existing neural attention-based models.

A Neural Transition-Based Approach for Semantic Dependency Graph Parsing

Semantic dependency graph has been recently proposed as an extension of tree-structured syntactic or semantic representation for natural language sentences. It particularly features the structural property of multi-head, which allows nodes to have multiple heads, resulting in a directed acyclic graph(DAG) parsing problem. Yet most statistical parsers focused exclusively on shallow bi-lexical tree structures, DAG parsing remains under-explored. In this paper, we propose a neural transition-based parser, using a variant of list-based arc-eager transition algorithm for dependency graph parsing. Particularly, two non-trivial improvements are proposed for representing the key components of the transition system, to better capture the semantics of segments and internal sub-graph structures. We test our parser on the SemEval-2016 Task 9 dataset (Chinese) and the SemEval-2015 Task 18 dataset (English). On both benchmark datasets, we obtain superior or comparable results to the best performing systems. Our parser can be further improved with a simple ensemble mechanism, resulting in the state-of-the-art performance.

Towards Effective Parsing with Neural Networks: Inherent Generalisations and Bounded Resource Effects

This article explores how the effectiveness of learning to parse with neural networks can be improved by including two architectural features relevant to language: generalisations across syntactic constituents and bounded resource effects. A number of neural network parsers have recently been proposed, each with a different approach to the representational problem of outputting parse trees. In addition, some of the parsers have explicitly attempted to capture an important regularity within language, which is to generalise information across syntactic constituents. A further property of language is that natural bounds exist for the number of constituents which a parser need retain for later processing. Both the generalisations and the resource bounds may be captured in architectural features which enhance the effectiveness and efficiency of learning to parse with neural networks. We describe a number of different types of neural network parser, and compare them with respect to these two features. These features are both explicitly present in the Simple Synchrony Network parser, and we explore and illustrate their impact on the process of learning to parse in some experiments with a recursive grammar.

Extracting error productions from a neural network-based LR parser

It is well-known that traditional rule-based parsers have poor error recovery capability. This has limited their practicality in natural language processing, where robustness and flexibility are of primary concern. In view of this, we propose the neural network LR parser (NNLR) in which the shift-reduce parsing decision of the LR parser is simulated by a feedforward neural network. Being trained with a small set of grammatical sentences only, the NNLR is capable of parsing a significantly large number of erroneous sentences. To explore the knowledge encoded in the neural network that sustains its robust processing capacity, we analyze the NNLR in two ways. First, we show that in the NNLR, erroneous sentences are recovered as if the parser had filled in some of the empty slots in the original LR parsing table. An augmented parsing table is thus constructed. Second, a set of new grammar rules, commonly called error productions, can be extracted from the trained network. When being included in the original grammar, these rules allow certain erroneous sentences to be generated and parsed in addition to grammatical ones. In both analyses, the symbolic knowledge discovered is readily comprehensible, and it can potentially be re-used by the original LR parser to enhance its robustness.

Inductive inference from noisy examples using the hybrid finite state filter

Recurrent neural networks processing symbolic strings can be regarded as adaptive neural parsers. Given a set of positive and negative examples, picked up from a given language, adaptive neural parsers can effectively be trained to infer the language grammar. In this paper we use adaptive neural parsers to face the problem of inferring grammars from examples that are corrupted by a kind of noise that simply changes their membership.We propose a training algorithm, referred to as hybrid finite state filter (HFF), which is based on a parsimony principle that penalizes the development of complex rules.We report very promising experimental results showing that the proposed inductive inference scheme is indeed capable of capturing rules, while removing noise.