Search

Named entity recognition (NER) is one fundamental task in natural language processing, which is typically addressed by neural condition random field (CRF) models, regarding the task as a sequence labeling problem. Sentence-level information has been shown positive for the task. Equipped with sophisticated neural structures such as long-short term memory network (LSTM), implicit sentence-level global information can be exploited fully, and has also been demonstrated effective in previous studies. In this work, we propose a new method for better learning of these sentence-level features in an explicit manner. Concretely, we suggest an auxiliary task, namely sentence-level named type prediction (i.e., determining whether a sentence includes a certain kind of named type), to supervise the feature representation learning globally. We conduct experiments on six benchmark datasets of various languages to evaluate our method. The results show that our final model is highly effective, resulting in significant improvements and leading to highly competitive results on all datasets.

Character-level Chinese dependency parsing has been a concern of several studies that naturally handle word segmentation, POS (Part of Speech) tagging and dependency parsing jointly in an end-to-end way. Previous work mostly concentrates on a transition-based framework for this task because of its easy adaption, which is extremely important when feature representation relies heavily on the decoding strategy, particularly under the traditional statistical setting. Recently, on the one hand, sophisticated deep neural networks and deep contextualized word representations have greatly weakened the dependence between feature representation and decoding. On the other hand, (first-order) graph-based models, especially the biaffine parsers, are straightforward for dependency parsing, and meanwhile they can yield competitive parsing performance. In this paper, we make a comprehensive investigation of the deep graph-based character-level dependency parsing for Chinese. We start from an extension of a standard graph-based biaffine parser, and then exploit Chinese BERT as well as our improved encoders based on transformers to enhance the character-level dependency parsing model. We conduct a series of experiments on the Chinese benchmark datasets, showing the performances of various graph-based character-level models and analyzing the advantages of the character-level dependency parsing under the deep neural setting.

Fine-grained opinion extraction has received increasing interests in the natural language processing community. It usually involves several subtasks. Recently, joint methods and neural models have been investigated by several studies, achieving promising performance by using graph-based models such as conditional random field. In this work, we propose a novel end-to-end neural model alternatively for joint opinion extraction, by using a transition-based framework. First, we exploit multi-layer bi-directional long short term memory (LSTM) networks to encode the input sentences, and then decode incrementally based on partial output results dominated by a transition system. We use global normalization and beam search for training and decoding. Experiments on a standard benchmark show that the proposed end-to-end model can achieve competitive results compared with the state-of-the-art neural models of opinion extraction.

With the rapid development of web and mobile applications, as well as their wide adoption in different domains, more and more personal data is provided, consciously or unconsciously, to different application providers. Privacy policy is an important medium for users to understand what personal information has been collected and used. As data privacy protection is becoming a critical social issue, there are laws and regulations being enacted in different countries and regions, and the most representative one is the EU General Data Protection Regulation (GDPR). It is thus important to detect compliance issues among regulations, e.g., GDPR, with privacy policies, and provide intuitive results for data subjects (i.e., users), data collection party (i.e., service providers) and the regulatory authorities. In this work, we target to solve the problem of compliance analysis between GDPR (Article 13) and privacy policies. We format the task into a combination of a sentence classification step and a rule-based analysis step. We manually curate a corpus of 36,610 labeled sentences from 304 privacy policies, and benchmark our corpus with several standard sentence classifiers. We also conduct a rule-based analysis to detect compliance issues and a user study to evaluate the usability of our approach. The web-based tool AutoCompliance is publicly accessible 1.

Research on overlapped and discontinuous named entity recognition (NER) has received increasing attention. The majority of previous work focuses on either overlapped or discontinuous entities. In this paper, we propose a novel span-based model that can recognize both overlapped and discontinuous entities jointly. The model includes two major steps. First, entity fragments are recognized by traversing over all possible text spans, thus, overlapped entities can be recognized. Second, we perform relation classification to judge whether a given pair of entity fragments to be overlapping or succession. In this way, we can recognize not only discontinuous entities, and meanwhile doubly check the overlapped entities. As a whole, our model can be regarded as a relation extraction paradigm essentially. Experimental results on multiple benchmark datasets (i.e., CLEF, GENIA and ACE05) show that our model is highly competitive for overlapped and discontinuous NER., Comment: Accepted in the main conference of ACL 2021

Crowdsourcing is regarded as one prospective solution for effective supervised learning, aiming to build large-scale annotated training data by crowd workers. Previous studies focus on reducing the influences from the noises of the crowdsourced annotations for supervised models. We take a different point in this work, regarding all crowdsourced annotations as gold-standard with respect to the individual annotators. In this way, we find that crowdsourcing could be highly similar to domain adaptation, and then the recent advances of cross-domain methods can be almost directly applied to crowdsourcing. Here we take named entity recognition (NER) as a study case, suggesting an annotator-aware representation learning model that inspired by the domain adaptation methods which attempt to capture effective domain-aware features. We investigate both unsupervised and supervised crowdsourcing learning, assuming that no or only small-scale expert annotations are available. Experimental results on a benchmark crowdsourced NER dataset show that our method is highly effective, leading to a new state-of-the-art performance. In addition, under the supervised setting, we can achieve impressive performance gains with only a very small scale of expert annotations., ACL-IJCNLP 2021 main conf, long paper; corrected the wrong reference for "argument retrieval" in first paragraph of Introduction

Prior studies show that cross-lingual semantic role labeling (SRL) can be achieved by model transfer under the help of universal features. In this paper, we fill the gap of cross-lingual SRL by proposing an end-to-end SRL model that incorporates a variety of universal features and transfer methods. We study both the bilingual transfer and multi-source transfer, under gold or machine-generated syntactic inputs, pre-trained high-order abstract features, and contextualized multilingual word representations. Experimental results on the Universal Proposition Bank corpus indicate that performances of the cross-lingual SRL can vary by leveraging different cross-lingual features. In addition, whether the features are gold-standard also has an impact on performances. Precisely, we find that gold syntax features are much more crucial for cross-lingual SRL, compared with the automatically-generated ones. Moreover, universal dependency structure features are able to give the best help, and both pre-trained high-order features and contextualized word representations can further bring significant improvements., Accepted at TASLP

Syntactic and semantic parsing has been investigated for decades, which is one primary topic in the natural language processing community. This article aims for a brief survey on this topic. The parsing community includes many tasks, which are difficult to be covered fully. Here we focus on two of the most popular formalizations of parsing: constituent parsing and dependency parsing. Constituent parsing is majorly targeted to syntactic analysis, and dependency parsing can handle both syntactic and semantic analysis. This article briefly reviews the representative models of constituent parsing and dependency parsing, and also dependency graph parsing with rich semantics. Besides, we also review the closely-related topics such as cross-domain, cross-lingual and joint parsing models, parser application as well as corpus development of parsing in the article., SCIENCE CHINA Technological Sciences

In this paper, we present Chinese lexical fusion recognition, a new task which could be regarded as one kind of coreference recognition. First, we introduce the task in detail, showing the relationship with coreference recognition and differences from the existing tasks. Second, we propose an end-to-end model for the task, handling mentions as well as coreference relationship jointly. The model exploits the state-of-the-art contextualized BERT representations as an encoder, and is further enhanced with the sememe knowledge from HowNet by graph attention networks. We manually annotate a benchmark dataset for the task and then conduct experiments on it. Results demonstrate that our final model is effective and competitive for the task. Detailed analysis is offered for comprehensively understanding the new task and our proposed model.

Many efforts of research are devoted to semantic role labeling (SRL) which is crucial for natural language understanding. Supervised approaches have achieved impressing performances when large-scale corpora are available for resource-rich languages such as English. While for the low-resource languages with no annotated SRL dataset, it is still challenging to obtain competitive performances. Cross-lingual SRL is one promising way to address the problem, which has achieved great advances with the help of model transferring and annotation projection. In this paper, we propose a novel alternative based on corpus translation, constructing high-quality training datasets for the target languages from the source gold-standard SRL annotations. Experimental results on Universal Proposition Bank show that the translation-based method is highly effective, and the automatic pseudo datasets can improve the target-language SRL performances significantly., Comment: Accepted at ACL 2020

Joint models have shown stronger capabilities for Chinese word segmentation and POS tagging, and have received great interests in the community of Chinese natural language processing. In this paper, we follow this line of work, presenting a simple yet effective sequence-to-sequence neural model for the joint task, based on a well-defined transition system, by using long short term memory neural network structures. We conduct experiments on five different datasets. The results demonstrate that our proposed model is highly competitive. By using well-trained character-level embeddings, the proposed neural joint model is able to obtain the best-reported performances in the literature.

While part-of-speech (POS) tagging and dependency parsing are observed to be closely related, existing work on joint modeling with manually crafted feature templates suffers from the feature sparsity and incompleteness problems. In this paper, we propose an approach to joint POS tagging and dependency parsing using transition-based neural networks. Three neural network based classifiers are designed to resolve shift/reduce, tagging, and labeling conflicts. Experiments show that our approach significantly outperforms previous methods for joint POS tagging and dependency parsing across a variety of natural languages.

Named entity recognition (NER) is an important task for biomedical text mining. Most prior work focused on recognizing regular entities that consist of continuous word sequences and are not overlapped with each other. In this paper, we propose a neural network model called Bi-LSTM-CRF that consists of bidirectional (Bi) long short-term memories (LSTMs) and conditional random fields (CRFs) to identify regular entities and the components of irregular entities. Then the components are combined to build final irregular entities according to manually designed rules. Furthermore, we propose a novel model called NerOne that consists of the Bi-LSTM-CRF network and another Bi-LSTM network. The Bi-LSTM-CRF network performs the same task as the aforementioned model, and the Bi-LSTM network determines whether two components should be combined. Therefore, NerOne automatically combines the components instead of using manually designed rules. We evaluate our models on two datasets for recognizing regular and irregular biomedical entities. Experimental results show that, with less feature engineering, the performances of our models are comparable with those of state-of-the-art systems. We show that the method of automatically combining the components is as effective as the method of manually designing rules. Our work can facilitate the research on biomedical text mining.

The rapid spread of preliminary scientific evidence is raising concerns on its role in producing misinformation during the COVID-19 pandemic. This research investigated how the communication of uncertainty about preliminary evidence affects the spread of its inferred misinformation in a Weibo case study. In total, 3439 Weibo posts and 10,380 reposts regarding the misinformation of pets transmitting COVID-19 were analyzed. The results showed that attitude ambiguity toward the preliminary evidence and the stage when the evidence was first released with uncertainty were associated with higher numbers of likes and retweets of misinformation posts. Our study highlights the internal sources of misinformation and revisits the contextual perspective in misinformation studies. Published version This research was supported by the Open Funding Project of the State Key Laboratory of Communication Content Cognition (grant number: 20G01) and the MICRON-NISTH Advancing Curiosity on Responsible AI Grant (Reg. No.: 200604393R).

Syntax has been demonstrated highly effective in neural machine translation (NMT). Previous NMT models integrate syntax by representing 1-best tree outputs from a well-trained parsing system, e.g., the representative Tree-RNN and Tree-Linearization methods, which may suffer from error propagation. In this work, we propose a novel method to integrate source-side syntax implicitly for NMT. The basic idea is to use the intermediate hidden representations of a well-trained end-to-end dependency parser, which are referred to as syntax-aware word representations (SAWRs). Then, we simply concatenate such SAWRs with ordinary word embeddings to enhance basic NMT models. The method can be straightforwardly integrated into the widely-used sequence-to-sequence (Seq2Seq) NMT models. We start with a representative RNN-based Seq2Seq baseline system, and test the effectiveness of our proposed method on two benchmark datasets of the Chinese-English and English-Vietnamese translation tasks, respectively. Experimental results show that the proposed approach is able to bring significant BLEU score improvements on the two datasets compared with the baseline, 1.74 points for Chinese-English translation and 0.80 point for English-Vietnamese translation, respectively. In addition, the approach also outperforms the explicit Tree-RNN and Tree-Linearization methods., NAACL 2019

Semantic role labeling (SRL), also known as shallow semantic parsing, is an important yet challenging task in NLP. Motivated by the close correlation between syntactic and semantic structures, traditional discrete-feature-based SRL approaches make heavy use of syntactic features. In contrast, deep-neural-network-based approaches usually encode the input sentence as a word sequence without considering the syntactic structures. In this work, we investigate several previous approaches for encoding syntactic trees, and make a thorough study on whether extra syntax-aware representations are beneficial for neural SRL models. Experiments on the benchmark CoNLL-2005 dataset show that syntax-aware SRL approaches can effectively improve performance over a strong baseline with external word representations from ELMo. With the extra syntax-aware representations, our approaches achieve new state-of-the-art 85.6 F1 (single model) and 86.6 F1 (ensemble) on the test data, outperforming the corresponding strong baselines with ELMo by 0.8 and 1.0, respectively. Detailed error analysis are conducted to gain more insights on the investigated approaches., Comment: AAAI 2019

Treebank translation is a promising method for cross-lingual transfer of syntactic dependency knowledge. The basic idea is to map dependency arcs from a source treebank to its target translation according to word alignments. This method, however, can suffer from imperfect alignment between source and target words. To address this problem, we investigate syntactic transfer by code mixing, translating only confident words in a source treebank. Cross-lingual word embeddings are leveraged for transferring syntactic knowledge to the target from the resulting code-mixed treebank. Experiments on University Dependency Treebanks show that code-mixed treebanks are more effective than translated treebanks, giving highly competitive performances among cross-lingual parsing methods., Comment: 10 pages

Opinion role labeling (ORL) is an important task for fine-grained opinion mining, which identifies important opinion arguments such as holder and target for a given opinion trigger. The task is highly correlative with semantic role labeling (SRL), which identifies important semantic arguments such as agent and patient for a given predicate. As predicate agents and patients usually correspond to opinion holders and targets respectively, SRL could be valuable for ORL. In this work, we propose a simple and novel method to enhance ORL by utilizing SRL, presenting semantic-aware word representations which are learned from SRL. The representations are then fed into a baseline neural ORL model as basic inputs. We verify the proposed method on a benchmark MPQA corpus. Experimental results show that the proposed method is highly effective. In addition, we compare the method with two representative methods of SRL integration as well, finding that our method can outperform the two methods significantly, achieving 1.47% higher F-scores than the better one.

Dependency parsing has been an important task in the natural language processing (NLP) community. Supervised methods have achieved great success these years. However, these models can suffer significant performance loss when test domain differs from the training domain. In this paper, we adopt the Bi-Affine parser as our baseline. To explore domain-specific information and domain-independent information for cross-domain dependency parsing, we apply an ensemble-style self-training and adversarial learning, respectively. We finally combine the two strategies to enhance our baseline model and our final system was ranked the first of at NLPCC2019 shared task on cross-domain dependency parsing.

Dependency syntax has been demonstrated highly useful for a number of natural language processing (NLP) tasks. Typical approaches of utilizing dependency syntax include Tree-RNN and Tree-Linearization, both of which exploit explicit 1-best tree outputs from a well-trained parser as inputs. However, these approaches may suffer from error propagation due to the inevitable errors contained in the 1-best tree outputs. In this work, we propose a novel approach to integrate dependency syntax without using the discrete tree outputs. The key idea is to use the intermediate hidden representations of a well-trained encoder-decoder dependency parser, which are referred to as Dep endency-based S yntax- A ware W ord R epresentations (Dep-SAWRs). Then, we simply concatenate such Dep-SAWRs with the conventional context-insensitive word embeddings to compose input word representations, without requiring to modify the model architecture of the downstream tasks. We evaluate the proposed method on four kinds of typical NLP tasks, including sentence classification, sentence matching, sequence labeling and machine translation. Experimental results show that the proposed approach is highly promising. On the one hand, it can utilize dependency syntax effectively, bringing consistently better performance on the four tasks compared with baselines without using syntax. On the other hand, the proposed method can outperform the Tree-RNN and Tree-Linearization approaches in most settings, and meanwhile are highly efficient in syntax integration. In addition, the proposed method would be easily extendable to encoding other structural attributes of language.

Sentiment analysis (also known as opinion mining) is an active research area in natural language processing. The task aims at identifying, extracting, and organizing sentiments from user-generated texts in social networks, blogs, or product reviews. Over the past two decades, many studies in the literature exploit machine learning approaches to solve sentiment analysis tasks from different perspectives. Since the performance of a machine learner heavily depends on the choices of data representation, many studies devote to building powerful feature extractor with domain expertise and careful engineering. Recently, deep learning approaches emerge as powerful computational models that discover intricate semantic representations of texts automatically from data without feature engineering. These approaches have improved the state of the art in many sentiment analysis tasks, including sentiment classification, opinion extraction, fine-grained sentiment analysis, etc. In this paper, we give an overview of the successful deep learning approaches sentiment analysis tasks at different levels.

Neural network models have recently received heated research attention in the natural language processing community. Compared with traditional models with discrete features, neural models have two main advantages. First, they take low-dimensional, real-valued embedding vectors as inputs, which can be trained over large raw data, thereby addressing the issue of feature sparsity in discrete models. Second, deep neural networks can be used to automatically combine input features, and including non-local features that capture semantic patterns that cannot be expressed using discrete indicator features. As a result, neural network models have achieved competitive accuracies compared with the best discrete models for a range of NLP tasks.

Background Extracting biomedical entities and their relations from text has important applications on biomedical research. Previous work primarily utilized feature-based pipeline models to process this task. Many efforts need to be made on feature engineering when feature-based models are employed. Moreover, pipeline models may suffer error propagation and are not able to utilize the interactions between subtasks. Therefore, we propose a neural joint model to extract biomedical entities as well as their relations simultaneously, and it can alleviate the problems above. Results Our model was evaluated on two tasks, i.e., the task of extracting adverse drug events between drug and disease entities, and the task of extracting resident relations between bacteria and location entities. Compared with the state-of-the-art systems in these tasks, our model improved the F1 scores of the first task by 5.1% in entity recognition and 8.0% in relation extraction, and that of the second task by 9.2% in relation extraction. Conclusions The proposed model achieves competitive performances with less work on feature engineering. We demonstrate that the model based on neural networks is effective for biomedical entity and relation extraction. In addition, parameter sharing is an alternative method for neural models to jointly process this task. Our work can facilitate the research on biomedical text mining.

Recent work on opinion mining typically focuses on subtasks such as aspect mining or polarity classification, ignoring the detailed explanatory evidences that account for one certain user opinion. In this paper, we study the extraction of explanatory expressions, by modeling the problem based on conditional random field (CRF). We compare the effectiveness of both discrete and neural features, and further integrate them. We evaluate the models on two datasets from two different domains which have been annotated with ground-truth explanatory expression. Results show that the neural CRF model performs better than the discrete CRF. After a combination of the discrete and neural features, our final CRF mode achieves the top-performing results.

Extracting Lives_In relations between bacteria and their locations involves two steps, namely bacteria/location entity recognition and Lives_In relation classification. Previous work solved this task by pipeline models, which may suffer error propagation and cannot utilize the interactions between these steps. We follow the line of work using joint models, which perform two subtasks simultaneously to obtain better performances. A state-of-the-art neural joint model for relation extraction in the Automatic Content Extraction (ACE) task is adapted to our task. Furthermore, we propose two strategies to improve this model. First, a novel relation is suggested in the second step to detect the errors in the first step, thus this relation can correct some errors in the first step. Second, we replace the original greedy-search decoding with beam-search, and train the model with early-update techniques. Experimental results on a standard dataset for this task show that our adapted model achieves better precisions than other systems. After adding the novel relation, we gain a nearly 2% improvement of F1 for Lives_In relation extraction. When beam-search is used, the F1 is further improved by 6%. These demonstrate that our proposed strategies are effective for this task. However, additional experiments show that the performance improvement in another dataset of bacteria and location extraction is not significant. Therefore, whether our methods are effective for other relation extraction tasks needs to be further investigated.

Neural networks have shown promising results for relation extraction. State-of-the-art models cast the task as an end-to-end problem, solved incrementally using a local classifier. Yet previous work using statistical models have demonstrated that global optimization can achieve better performances compared to local classification. We build a globally optimized neural model for end-to-end relation extraction, proposing novel LSTM features in order to better learn context representations. In addition, we present a novel method to integrate syntactic information to facilitate global learning, yet requiring little background on syntactic grammars thus being easy to extend. Experimental results show that our proposed model is highly effective, achieving the best performances on two standard benchmarks.

The Chinese language is a character-based language, with no explicit separators between words like English. Traditionally, word segmentation is conducted to convert Chinese sentences into word sequences, thus the same framework of English sentiment analysis can be exploited for Chinese. These work uses a specified word segmentor as a prerequisite step, yet ignores the fact that different segmentation styles exist in Chinese word segmentation, such as CTB, PKU, MSR and etc. In this paper, we study the influences of these heterogeneous segmentations for Chinese sentiment analysis, and then integrate these segmentations, based on both discrete and neural models. Experimental results show that different segmentations do affect the final performances, and the integrated models can achieve better performances.

Discrete and Neural models are two mainstream methods for Chinese POS tagging nowadays. Both have achieved state-of-the-art performances. In this paper, we compare the two kinds of models empirically, and further investigate the combination methods of them. In particular, as the pre-trained word embeddings are exploited under the neural setting, one can regard neural models as semi-supervised setting. To make a fairer comparison of the discrete and the neural models, we incorporate word clusters for both models as well as their combination, since it has been generally accepted that word clusters can encode similar information as pre-trained word embeddings.

In this paper, we presents a stance detection system for NLPCC-ICCPOL 2016 share task 4. Our Stance Detection System can determinate whether the author of Weibo text is in favor of the given target, against the given target, or neither. We exploit LSTMs model and the average F score of our system is 56.56%. In contrast to the traditional target/aspect sentiment, the given target may not be preserved in Weibo text. We model the task as a classification problem, exploiting LSTMs as the basic part of classifier.

We investigate a combination of a traditional linear sparse feature model and a multi-layer neural network model for deterministic transition-based dependency parsing, by integrating the sparse features into the neural model. Correlations are drawn between the hybrid model and previous work on integrating word embedding features into a discrete linear model. By analyzing the results of various parsers on web-domain parsing, we show that the integrated model is a better way to combine traditional and embedding features compared with previous methods.

We propose an event-driven model for headline generation. Given an input document, the system identifies a key event chain by extracting a set of structural events that describe them. Then a novel multi-sentence compression algorithm is used to fuse the extracted events, generating a headline for the document. Our model can be viewed as a novel combination of extractive and abstractive headline generation, combining the advantages of both methods using event structures. Standard evaluation shows that our model achieves the best performance compared with previous state-of-the-art systems.

We propose a transition-based model for joint word segmentation, POS tagging and text normalization. Different from previous methods, the model can be trained on standard text corpora, overcoming the lack of annotated microblog corpora. To evaluate our model, we develop an annotated corpus based on microblogs. Experimental results show that our joint model can help improve the performance of word segmentation on microblogs, giving an error reduction in segmentation accuracy of 12.02%, compared to the traditional approach.

Open domain targeted sentiment is the joint information extraction task that finds target mentions together with the sentiment towards each mention from a text corpus. The task is typically modeled as a sequence labeling problem, and solved using state-of-the-art labelers such as CRF. We empirically study the effect of word embeddings and automatic feature combinations on the task by extending a CRF baseline using neural networks, which have demonstrated large potentials for sentiment analysis. Results show that the neural model can give better results by significantly increasing the recall. In addition, we propose a novel integration of neural and discrete features, which combines their relative advantages, leading to significantly higher results compared to both baselines.

Chinese grammar engineering has been a much debated task. Whilst semantic information has been reconed crucial for Chinese syntactic analysis and downstream applications, existing Chinese treebanks lack a consistent and strict sentential semantic formalism. In this paper, we introduce a semantics oriented grammar for Chinese, designed to provide basic supports for tasks such as automatic semantic parsing and sentence generation. It has a directed acyclic graph structure with a simple yet expressive label set, and leverages elementary predication to support logical form conversion. To our knowledge, it is the first Chinese grammar representation capable of direct transformation into logical forms.

Recent work on Chinese analysis has led to large-scale annotations of the internal structures of words, enabling characterlevel analysis of Chinese syntactic structures. In this paper, we investigate the problem of character-level Chinese dependency parsing, building dependency trees over characters. Character-level information can benefit downstream applications by offering flexible granularities for word segmentation while improving wordlevel dependency parsing accuracies. We present novel adaptations of two major shift-reduce dependency parsing algorithms to character-level parsing. Experimental results on the Chinese Treebank demonstrate improved performances over word-based parsing methods.

We address the problem of Chinese semantic dependency parsing. Dependency parsing is traditionally oriented to syntax analysis, which we denote by syntactic dependency parsing to distinguish it from semantic dependency parsing. In this paper, firstly we compare Chinese semantic dependency parsing and syntactic dependency parsing systematically, showing that syntactic dependency parsing can potentially improve the performance of semantic dependency parsing. Thus then we suggest an approach based on quasi-synchronous grammar to incorporate the auto-parsed syntactic dependency tree into semantic dependency parsing. We conduct experiments on the Chinese semantic dependency parsing corpus of SemEval-2012. Finally we achieve 65.25% LAS on test corpus, gaining increases of 2.45% compared to the top result of 62.80% in SemEval-2012.