Dynamic Hybrid Relation Exploration Network for Cross-Domain Context-Dependent Semantic Parsing

Hui, Binyuan; Geng, Ruiying; Ren, Qiyu; Li, Binhua; Li, Yongbin; Sun, Jian; Huang, Fei; Si, Luo; Zhu, Pengfei; Zhu, Xiaodan

doi:10.1609/aaai.v35i14.17550

Cited by 29 publications

(14 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Numerous studies have harnessed previously generated SQL queries to address extended dependencies and enhance parsing accuracy [8], [15], [16]. Additionally, research by Cai and Wan [14] and Hui et al [11] has leveraged graph neural networks to jointly encode multi-turn questions and schema information. Building upon the accomplishments of pre-trained models like T5, BERT, ALM, GanLM, and BART [22]- [25], ScoRE [9] and Star [26] design pre-training framework which leverage contextual information to enrich natural language (NL) utterance and table schema representations for text-to-SQL conversations.…”

Section: Related Workmentioning

confidence: 99%

“…This entails the model's ability to effectively establish the entity mapping relationship between the user query and the database schema, while also comprehending the underlying intent of the current inquiry in the context provided. Several prior studies [9], [11]- [13] have employed neural network encoders that con- catenate the current question, question context, and schema. Concurrently, a number of approaches have directly incorporated historically generated SQL queries [8], [14]- [16] to aid the model in SQL parsing for the present question.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

SE-HCL: Schema Enhanced Hybrid Curriculum Learning for Multi-Turn Text-to-SQL

Zhang,

Zhou,

Huang

2024

IEEE Access

View full text Add to dashboard Cite

Existing multi-turn Text-to-SQL approaches, mainly use data in a randomized order when training the model, ignoring the rich structural information contained in the dialog and schema. In this paper, we propose to use curriculum learning (CL) to better leverage the curriculum structure of schema, query, and dialog for multi-turn question-query pairs. We design a model-agnostic framework named Schema Enhanced Hybrid Curriculum Learning (SE-HCL) for multi-turn Text-to-SQL to help the models gain a full contextual semantic understanding. Concretely, We measure the difficulty of the data from both a structural and model perspective. In terms of data structure, we mainly consider the turns of the question and the complexity of the schema and SQL query. Accordingly, we designed a data course module to dynamically adjust the difficulty of the data based on the convergence of the model and the schema enhancement method we designed. In terms of the model, we propose a scoring module that will judge the difficulty of a problem based on whether the model could solve the question effectively. Finally, we will consider both aspects and design a hybrid curriculum to determine the flow of model training. Our experiments show that our proposed method improves SQL-generated performance over previous state-of-the-art models on SparC and CoSQL, especially for hard and long-turn questions.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

SE-HCL: Schema Enhanced Hybrid Curriculum Learning for Multi-Turn Text-to-SQL

Zhang,

Zhou,

Huang

2024

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Moreover, we train all the models on WikiBio and WikiPerson from scratch, and the training cost is rather expensive: 2.5 days using 4 NVIDIA V100 32G GPUs. Lastly, this paper does not compare the pre-trained language models (PLMs) (Devlin et al, 2019;Raffel et al, 2020;Hui et al, 2021Hui et al, , 2022, though our approach may also benefit from some pre-trained table encoders, such as TAPAS (Müller et al, 2021). The main reasons why we do not consider PLMs are that PLMs will bring an unfair comparison and bring more variables and may make our work lose focus.…”

Section: Limitationsmentioning

confidence: 99%

Plan-then-Seam: Towards Efficient Table-to-Text Generation

Li¹,

Geng²,

Fang³

et al. 2023

Preprint

View full text Add to dashboard Cite

Table -to-text generation aims at automatically generating text to help people conveniently obtain salient information in tables. Recent works explicitly decompose the generation process into content planning and surface generation stages, employing two autoregressive networks for them respectively. However, they are computationally expensive due to the nonparallelizable nature of autoregressive decoding and the redundant parameters of two networks. In this paper, we propose the first totally non-autoregressive table-to-text model (Plan-then-Seam, PTS) that produces its outputs in parallel with one single network. PTS firstly writes and calibrates one plan of the content to be generated with a novel rethinking pointer predictor, and then takes the plan as the context for seaming to decode the description. These two steps share parameters and perform iteratively to capture token interdependency while keeping parallel decoding. Experiments on two public benchmarks show that PTS achieves 3.0 ∼ 5.6 times speedup for inference time, reducing 50% parameters, while maintaining as least comparable performance against strong two-stage table-to-text competitors 1 .

show abstract

“…Two columns of VXI additional bus pins are used to design the storage pool of image information resources in college sports multimedia teaching. At the input and output devices, multithread control bus design and serial output bus control structure design are adopted [13][14] . Based on ZigBee networking design method, the output end of college sports multimedia teaching platform under moving image analysis mode is established.…”

Section: Multimedia Teaching Resource Database Permission Configurati...mentioning

confidence: 99%

“…Wherein, N is the statistical sample number of image information resources in college sports multimedia teaching, and X and Y are fuzzy constraint factors that users' visits exceed the load limit of the system, respectively. The optimal solution of dynamic distribution of image information resources in college sports multimedia teaching is obtained, and the distribution length of image information resources in college sports multimedia teaching is N, which will be converted into xi La = strings, thus establishing a scheduling model of image information resources in college sports multimedia teaching [14] .…”

Section: Multimedia Usermentioning

confidence: 99%

Realization of college sports multimedia teaching platform based on cloud computing motion image analysis

2022

Preprint

View full text Add to dashboard Cite

In order to improve the ability of college sports multimedia teaching, dynamic image analysis and data management, a design scheme of college sports multimedia teaching platform based on cloud computing mode was proposed. The method of dynamic integrated allocation and fuzzy scheduling of multimedia teaching resources is adopted. The characteristic values and dynamic analysis of moving images of college sports multimedia teaching are extracted. By using hierarchical structure and image integration scheduling method, this paper analyzes the characteristic values of image analysis in college sports multimedia teaching and dynamic analysis of moving images. By using the methods of adaptive cooperative control and resource matching scheduling, the control objective allocation function of college sports multimedia teaching and dynamic analysis of moving images is constructed. The correlation of image information resources in college sports multimedia teaching is mined by using the ultra micro pixel feature matching method, and a big data mining and information fusion model for college sports multimedia education and dynamic analysis of motion images is constructed. Use cloud computing technology to provide users with various services, and use machine learning technology to analyze teaching data. Through priority scheduling and dynamic allocation methods, the dynamic allocation of teaching resources and dynamic analysis of moving images in the process of college sports multimedia teaching are realized. In ASP NET technology, cloud computing and Silverlight technology have realized the software development and design of college sports multimedia teaching platform under the motion image analysis mode. The simulation results show that using this method to design the college sports multimedia teaching platform under the motion image analysis mode can improve the dynamic distribution and mining ability of image information resources in college sports multimedia education. The platform has a high level of data fusion and a strong ability to analyze motion images.

show abstract

Dynamic Hybrid Relation Exploration Network for Cross-Domain Context-Dependent Semantic Parsing

Cited by 29 publications

References 27 publications

SE-HCL: Schema Enhanced Hybrid Curriculum Learning for Multi-Turn Text-to-SQL

SE-HCL: Schema Enhanced Hybrid Curriculum Learning for Multi-Turn Text-to-SQL

Plan-then-Seam: Towards Efficient Table-to-Text Generation

Realization of college sports multimedia teaching platform based on cloud computing motion image analysis

Contact Info

Product

Resources

About