检索结果-内蒙古大学图书馆

International Joint Conference on Neural Networks (IJCNN)

作者： Liu, Hu Shi, Yuliang Zhang, Jianlin Wang, Xinjun Li, Hui Kong, Fanyu Shandong Univ Sch Software Jinan Peoples R China Dareway Software Co Ltd Jinan Peoples R China

ISBN: (纸本)9781665488679

text-to-sql aims to parse natural language problems into sql queries, which can provide a simple interface to access large databases enabling sql novices a quicker entry into databases. As the text-to-sql field is intensively studied, more and more models use GNNs to encode heterogeneous graph information in this task, and how to better obtain path information between nodes in database schema heterogeneous graphs and question-database schema heterogeneous graphs will greatly affect the effectiveness of the model parsing. Our work intends to explore the problem of solving the encoding of heterogeneous graph meta-paths in the text-to-sql task. Previous approaches usually use multi-layer GNNs to aggregate topological structure information between nodes. However, they ignored the structural information embedded at the edges and also failed to obtain nodes that are not directly connected but can provide contextual information through meta-paths. To solve the above problem, we propose Multi-Hop Relational Graph Attention Network based text-to-sql Parsing Model (MHRGATsql) for learning topological information between nodes while obtaining semantic information embedded in the edge topology. We use multi-hop attention to modify the relational graph attention network to diffuse the attention scores throughout the network, thus increasing the "receptive field" of each layer of RGAT. Experimental results on the large-scale cross-domain text-to-sql dataset Spider show that our model obtains an absolute improvement of 1.7% compared to the baseline and alleviates the over-smoothing problem in the deep network model.

关键词： Natural language processing text-to-sql Semantic parsing sql Graph neural network

来源：评论

学校读者我要写书评

暂无评论

sql-to-Schema Enhances Schema Linking in text-to-sql 35th

SQL-to-Schema Enhances Schema Linking in Text-to-SQL

引用

35th International Conference on Database and Expert Systems Applications (DEXA)

作者： Yang, Sun Su, Qiong Li, Zhishuai Li, Ziyue Mao, Hangyu Liu, Chenxi Zhao, Rui Peking Univ Beijing Peoples R China Guizhou Univ Guiyang Guizhou Peoples R China SenseTime Res Shanghai Peoples R China Nanyang Technol Univ Singapore Singapore

ISBN: (纸本)9783031683084;9783031683091

Sophisticated text-to-sql methods often face errors, such as schema-linking errors, join errors, nested errors, and group-by errors. To mitigate these, it's crucial to filter out unnecessary tables and columns, focusing the language model on relevant ones. Previous methods have attempted to sort tables and columns based on relevance or directly identify necessary elements, but these approaches suffer from long training times, high costs with GPT-4 tokens, or poor schema linking performance. We propose a two-step schema linking method: first, generate an initial sql query using the full database schema;then, extract the relevant tables and columns to form a concise schema. This method, tested with Code Llama and GPT-4, shows optimal performance compared to mainstream methods on the Spider dataset, reducing errors and improving efficiency in sql generation.

关键词： text-to-sql Schema Linking Large Language Model

来源：评论

学校读者我要写书评

暂无评论

Context-Dependent text-to-sql Generation with Intermediate Representation 20th

Context-Dependent Text-to-SQL Generation with Intermediate R...

引用

20th Pacific Rim International Conference on Artificial Intelligence

作者： Gao, Xuesong Zhao, Junfeng Inner Mongolia Univ Coll Comp Sci Hohhot Peoples R China

ISBN: (纸本)9789819970186;9789819970193

In recent years, the text-to-sql task has become a research hotspot in semantic analysis. Among them, context-dependent text-to-sql task has received more and more attention as it meets the needs of actual scenarios. The core of the problem is how to use historical interaction information and database schema to understand the context. Most existing research ignores the structure of sql queries and introduces low-level information such as variable names and parameters, and the mismatch problem between intents expressed in utterance and the implementation details in sql still exists. In this paper, SemQL is applied to serve as an intermediate representation between utterance and sql, meanwhile, the Coarse-to-Fine neural architecture is adopted to decompose decoding process of SemQL into two stages. We validated the performance of our model on SParC and Cosql datasets, which outperforms the existing ones and achieves excellent results on both datasets.

关键词： Semantic Parsing text-to-sql Intermediate Representation

来源：评论

学校读者我要写书评

暂无评论

Improving the Accuracy of text-to-sql Tools Based on Large Language Models for Real-World Relational Databases 35th

Improving the Accuracy of Text-to-SQL Tools Based on Large L...

引用

35th International Conference on Database and Expert Systems Applications (DEXA)

作者： Coelho, Gustavo M. C. Nascimento, Eduardo R. S. Izquierdo, Yenier T. Garcia, Grettel M. Feijo, Lucas Lemos, Melissa Garcia, Robinson L. S. de Oliveira, Aiko R. Pinheiro, Joao P. Casanova, Marco A. Pontificia Univ Catolica Rio de Janeiro Inst Tecgraf BR-22451900 Rio de Janeiro RJ Brazil Petrobras SA BR-20031912 Rio De Janeiro RJ Brazil Pontificia Univ Catolica Rio de Janeiro Dept Informat BR-22451900 Rio de Janeiro RJ Brazil

ISBN: (纸本)9783031683084;9783031683091

Real-world relational databases (RW-RDB) have large, complex schemas often expressed in terms alien to end-users. This scenario is challenging to LLM-based text-to-sql tools, that is, tools that translate Natural Language (NL) sentences into sql queries using a Large Language Model (LLM). Indeed, their accuracy on RW-RDBs is considerably less than that reported for well-known synthetic benchmarks. This paper then introduces a technique to improve the accuracy of LLM-based text-to-sql tools on RW-RDBs using Retrieval-Augmented Generation. The technique consists of two steps. Using the RW-RDB schema, the first step generates a synthetic dataset E of pairs (QN, QS), where QN is an NL sentence and QS is the corresponding sql translation. The core contribution of the paper is an algorithm that implements this first step. Given an input NL sentence QI, the second step retrieves pairs (QN, QS) from E based on the similarity of QI and QN, and prompts such pairs to the LLM to improve accuracy. To argue in favor of the proposed technique, the paper includes experiments with an RW-RDB, which is in production at an Energy company, and a well-known text-to-sql prompt strategy. It repeats the experiments with Mondial, an openly available database with a large schema. These experiments constitute a second contribution of the paper.

关键词： text-to-sql Retrieval-Augmented Generation RAG GPT Large Language Models Relational Databases

来源：评论

学校读者我要写书评

暂无评论

PS-sql: Phrase-based Schema-Linking with Pre-trained Language Models for text-to-sql Parsing 6

PS-SQL: Phrase-based Schema-Linking with Pre-trained Languag...

引用

6th International Conference on Natural Language Processing (ICNLP)

作者： Lan, Zhibo Li, Shuangyin South China Normal Univ Sch Comp Sci Guangzhou Peoples R China

ISBN: (纸本)9798350349122;9798350349115

In the text-to-sql task, a significant challenge is enabling parsers to generalize effectively across diverse domains. Key to solving is schema-linking, which involves mapping words to the pertinent columns or tables in the databases. Existing methods base on pre-trained language models (PLMs), which rely on token masking, have limitations in capturing the variety of schemas. Unlike single token, phrases offer richer semantics, and superior discrimination in determining whether one word corresponds to tables or columns. In this paper, we present an innovative approach named Phrase-based Schema-Linking for text-to-sql (PS-sql). By incorporating extracted phrases from the question, we enhance PLMs' ability to learn the mapping between tokens and schemas, leading to more robust schema-linking. We also introduce a mechanism to refine extracted phrases, reducing noise. During practical evaluations on several real-world datasets, PS-sql consistently delivers enhanced schema-linking precision, resulting in higher-quality sql query generation.

关键词： Semantic Parsing text-to-sql Pre-trained Language Model Schema-Linking

来源：评论

学校读者我要写书评

暂无评论

CMAT: Column-Mask-Augmented Training for text-to-sql Parsers 28th

CMAT: Column-Mask-Augmented Training for Text-to-SQL Parsers

引用

28th International Conference on Neural Information Processing

作者： Chang, Chen Peking Univ Beijing Peoples R China Peking Univ Ctr Data Sci Beijing Peoples R China

ISBN: (纸本)9783030923099;9783030923105

When it comes to text-to-sql tasks, the model needs to learn context-based representations of schema along with natural language utterances. We present a simple and effective method for text-to-sql tasks, Column-Mask-Augmented Training (CMAT), to make up for the insufficiency of training data. To exploit the synthesized data, we propose the clause prediction (CP) object for multi-task learning, which forces the model to capture contextual features of the schema items. Besides, we add the fuzzy match and subword match to the schema linking strategy in RAT-sql. As a result, our method significantly increases the recall and F1 value of schema linking and achieves a competitive result with RAT-sql and GraPPa on Spider.

关键词： text-to-sql Multi-task learning Data augmentation

来源：评论

学校读者我要写书评

暂无评论

Towards Robustness of Large Language Models on text-to-sql Task: An Adversarial and Cross-Domain Investigation 1

引用

32nd International Conference on Artificial Neural Networks (ICANN)

作者： Zhang, Weixu Wang, Yu Fan, Ming Xi An Jiao Tong Univ Xian 710049 Peoples R China

ISBN: (数字)9783031441929

ISBN: (纸本)9783031441912;9783031441929

Recent advances in large language models (LLMs) like Chat-GPT have led to impressive results on various natural language processing (NLP) challenges including text-to-sql task, which aims to automatically generate sql queries from natural language questions. However, these languagemodels are still subject to vulnerabilities such as adversarial attacks, domain shift and lack of robustness, which can greatly affect their performance and reliability. In this paper, we conduct a comprehensive evaluation of large language models, such as ChatGPT, on their robustness in text-to-sql tasks. We assess the impact of adversarial and domain generalization perturbations on LLMs using seven datasets, five of which are popular robustness evaluation benchmarks for text-to-sql tasks and two are synthetic adversarial datasets generated by ChatGPT. Our experiments show that while LLMs exhibit promise as zero-shot text-to-sql parsers, their performances degrade under adversarial and domain generalization perturbations, with varying degrees of robustness depending on the type and level of perturbations applied. We also explore the impact of usage-related factors such as prompt design on the performance and robustness of LLMs. Our study provides insights into the limitations and potential directions for future research to enhance the performance and robustness of LLMs on text-to-sql and other NLP tasks.

关键词： Large language model ChatGPT text-to-sql Robustness Adversarial attacks

来源：评论

学校读者我要写书评

暂无评论

An Experimental Research of text-to-sql for Heterogeneous Data in Large Language Models 20th

An Experimental Research of Text-to-SQL for Heterogeneous Da...

引用

20th International Conference on Intelligent Computing (ICIC)

作者： Yang, Weiwei Wang, Xiaoliang Chen, Bosheng Liu, Yong Wang, Bing Wang, Hui Wang, Xiaoke Zhua, Haitao Wang, Zhehao China Elect Digital Innovat Co Ltd Beijing Peoples R China

ISBN: (纸本)9789819756629;9789819756636

The large language model (LLMs) technology has become a new paradigm for text-to-sql task. However, when faced with heterogeneous data, there is still a lack of comprehensive and effective solutions for text-to-sql task, especially in terms of high-quality corpus construction methods and low-resource training strategies. To address this challenge, this paper systematically studies text-to-sql technology in the field of heterogeneous data and innovatively proposes the HD-sqlmethod. This method achieves the construction of text-to-sql corpora for heterogeneous data that are secure, efficient, and low-cost. Based on this corpus, we further explore supervised fine-tuning strategies for low-resource LLMs on heterogeneous data. Compared to baseline models, significant performance improvements are achieved on the Heterogeneous Data Test set, further validating the practicality and effectiveness of the HD-sql method. Overall, this paper proposes a feasible solution for text-to-sql task with heterogeneous data. We hope that our work can advance research in this field and provide strong technical support for practical applications.

关键词： text-to-sql Large Language Model Supervised Fine-Tuning Heterogeneous Data

来源：评论

学校读者我要写书评

暂无评论

GS-sql: Modeling Spatial Semantics in Spatial text-to-sql

GS-SQL: Modeling Spatial Semantics in Spatial Text-to-SQL

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Zhang, Xu Xiao, Feiyang Yan, Liang Zhang, Zhiqiang Chongqing Univ Posts & Telecommun Dept Comp Sci & Technol Chongqing Peoples R China Minist Nat Resources Key Lab Monitoring Evaluat & Early Warning Terr S Chongqing Peoples R China Minist Culture & Tourism Key Lab Tourism Multisource Data Percept & Decis Chongqing Peoples R China Chongqing Planning & Nat Resources Informat Ctr Chongqing Peoples R China

ISBN: (纸本)9798350359329;9798350359312

Conventional text-to-sql research tackles the problem of solving user questions in natural language by generating the corresponding sql queries. Most of the recent works are dedicated to improving model's robustness and generalizability in cross- domain settings. However, model's capability in solving geography-related questions remains unexploited. In this paper we propose GS-sql, a new framework that jointly model the schema item alignment and geospatial semantics in the question. The proposed framework consists of an improved abstract syntax tree for representing spatial queries, a novel spatial entity tagging module for locating entities in the question, and a spatial semantics extraction module for determining the spatial relationship between the entities. Then we propose GeoSpatialSpider, a dataset that introduces geospatial queries, requiring model to yield spatial functions and nested sql inside functions. Finally we evaluate the proposed method on our dataset Experimental results show the effectiveness of our abstract syntax tree and GS-sql in parsing geospatial semantics while preserving traditional textto-sql capabilities.

关键词： text-to-sql Semantic Parsing Code Generation

来源：评论

学校读者我要写书评

暂无评论

TSCL-sql: Two-Stage Curriculum Learning Framework for text-to-sql

引用

Journal of Donghua University(English Edition) 2023年第4期40卷 421-427页

作者：尹枫程路易王秋月王志军杜明徐波 School of Computer Science and Technology Donghua UniversityShanghai 201620China

text-to-sql is the task of translating a natural language query into a structured query language. Existing text-to-sql approaches focus on improving the model’s architecture while ignoring the relationship between queries and table schemas and the differences in difficulty between examples in the dataset. To tackle these challenges, a two-stage curriculum learning framework for text-to-sql(TSCL-sql) is proposed in this paper. To exploit the relationship between the queries and the table schemas, a schema identification pre-training task is proposed to make the model choose the correct table schema from a set of candidates for a specific query. To leverage the differences in difficulty between examples, curriculum learning is applied to the text-to-sql task, accompanied by an automatic curriculum learning solution, including a difficulty scorer and a training scheduler. Experiments show that the framework proposed in this paper is effective.

关键词： text-to-sql curriculum learning semantic parsing

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：