Filters








28,212 Hits in 3.2 sec

Improving Compositional Generalization in Semantic Parsing [article]

Inbar Oren, Jonathan Herzig, Nitish Gupta, Matt Gardner, Jonathan Berant
2020 arXiv   pre-print
In this work, we investigate compositional generalization in semantic parsing, a natural test-bed for compositional generalization, as output programs are constructed from sub-components.  ...  We analyze a wide variety of models and propose multiple extensions to the attention module of the semantic parser, aiming to improve compositional generalization.  ...  In this paper, we thoroughly analyze the impact of different modeling choices on compositional generalization in 5 semantic parsing datasets-four that are text-to-SQL datasets, and DROP, a dataset for  ... 
arXiv:2010.05647v1 fatcat:mj7nlhfbizgbdbmwk4yrbezd4m

Improving Compositional Generalization in Semantic Parsing

Inbar Oren, Jonathan Herzig, Nitish Gupta, Matt Gardner, Jonathan Berant
2020 Findings of the Association for Computational Linguistics: EMNLP 2020   unpublished
In this work, we investigate compositional generalization in semantic parsing, a natural test-bed for compositional generalization, as output programs are constructed from sub-components.  ...  We analyze a wide variety of models and propose multiple extensions to the attention module of the semantic parser, aiming to improve compositional generalization.  ...  In this paper, we thoroughly analyze the impact of different modeling choices on compositional generalization in 5 semantic parsing datasets-four that are text-to-SQL datasets, and DROP, a dataset for  ... 
doi:10.18653/v1/2020.findings-emnlp.225 fatcat:ncfdgbci5vhszbj2tn2bdn2njm

SUBS: Subtree Substitution for Compositional Semantic Parsing [article]

Jingfeng Yang, Le Zhang, Diyi Yang
2022 arXiv   pre-print
Although sequence-to-sequence models often achieve good performance in semantic parsing for i.i.d. data, their performance is still inferior in compositional generalization.  ...  However, prior work only leveraged superficial grammar or rules for data augmentation, which resulted in limited improvement.  ...  This work is funded in part by a grant from Amazon and Salesforce.  ... 
arXiv:2205.01538v1 fatcat:2yikyid4yzg23ffjhkgpneen7u

Deep Image Harmonization

Yi-Hsuan Tsai, Xiaohui Shen, Zhe Lin, Kalyan Sunkavalli, Xin Lu, Ming-Hsuan Yang
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
Compositing is one of the most common operations in photo editing. To generate realistic composites, the appearances of foreground and background need to be adjusted to make them compatible.  ...  In this work, we propose an end-to-end deep convolutional neural network for image harmonization, which can capture both the context and semantic information of the composite images during harmonization  ...  This work is supported in part by the NSF CAREER Grant #1149783, NSF IIS Grant #1152576, and a gift from Adobe. Portions of this work were performed while Y.-H. Tsai was an intern at Adobe Research.  ... 
doi:10.1109/cvpr.2017.299 dblp:conf/cvpr/TsaiSLSL017 fatcat:gd2ww7pfdnferdowtqiez6jx3m

Fish Transporters and Miracle Homes: How Compositional Distributional Semantics can Help NP Parsing

Angeliki Lazaridou, Eva Maria Vecchi, Marco Baroni
2013 Conference on Empirical Methods in Natural Language Processing  
In this work, we argue that measures that have been shown to quantify the degree of semantic plausibility of phrases, as obtained from their compositionally-derived distributional semantic representations  ...  We show that our plausibility cues outperform a strong baseline and significantly improve performance when used in combination with state-of-the-art features.  ...  Besides paving the way to a more general integration of compositional distributional semantics in syntactic parsing, the proposed methodology provides a new way to evaluate composition functions.  ... 
dblp:conf/emnlp/LazaridouVB13 fatcat:aiwc77plnjhipmmj4kretsxllq

Deep Image Harmonization [article]

Yi-Hsuan Tsai, Xiaohui Shen, Zhe Lin, Kalyan Sunkavalli, Xin Lu, Ming-Hsuan Yang
2017 arXiv   pre-print
Compositing is one of the most common operations in photo editing. To generate realistic composites, the appearances of foreground and background need to be adjusted to make them compatible.  ...  In this work, we propose an end-to-end deep convolutional neural network for image harmonization, which can capture both the context and semantic information of the composite images during harmonization  ...  In order to further improve harmonization results, it is natural to consider semantics of the composite foreground region.  ... 
arXiv:1703.00069v1 fatcat:72z2nc6ttbb65k4t6y3l4ufihy

Non-Autoregressive Semantic Parsing for Compositional Task-Oriented Dialog [article]

Arun Babu, Akshat Shrivastava, Armen Aghajanyan, Ahmed Aly, Angela Fan, Marjan Ghazvininejad
2021 arXiv   pre-print
In this work, we propose a non-autoregressive approach to predict semantic parse trees with an efficient seq2seq model architecture.  ...  Our novel architecture achieves up to an 81% reduction in latency on TOP dataset and retains competitive performance to non-pretrained models on three different semantic parsing datasets.  ...  ., 2020) , an extension of the compositional form proposed in for task oriented semantic parsing.  ... 
arXiv:2104.04923v1 fatcat:cmpiibt3v5co5fjxf7vhssuvvm

Iterative Utterance Segmentation for Neural Semantic Parsing [article]

Yinuo Guo, Zeqi Lin, Jian-Guang Lou, Dongmei Zhang
2020 arXiv   pre-print
Experiments on Geo, ComplexWebQuestions, and Formulas show that our framework can consistently improve performances of neural semantic parsers in different domains.  ...  On data splits that require compositional generalization, our framework brings significant accuracy gains: Geo 63.1 to 81.2, Formulas 59.7 to 72.7, ComplexWebQuestions 27.1 to 56.3.  ...  Therefore, the lack of compositional generalization ability is still a challenging problem in neural semantic parsing (Finegan-Dollak et al. 2018; Keysers et al. 2020 ).  ... 
arXiv:2012.07019v1 fatcat:sciq2qg7unaqffgylcpu7drvmu

Compositional Generalization via Semantic Tagging [article]

Hao Zheng, Mirella Lapata
2021 arXiv   pre-print
Experimental results on three semantic parsing datasets show that the proposed approach consistently improves compositional generalization across model architectures, domains, and semantic formalisms.  ...  Although neural sequence-to-sequence models have been successfully applied to semantic parsing, they fail at compositional generalization, i.e., they are unable to systematically generalize to unseen compositions  ...  Conclusions We presented a two-stage decoding framework, aiming to improve compositional generalization in neural semantic parsing.  ... 
arXiv:2010.11818v2 fatcat:ymqwzq2qfbeuri2aycjhce2kvi

Semantic Parsing in Task-Oriented Dialog with Recursive Insertion-based Encoder [article]

Elman Mansimov, Yi Zhang
2022 arXiv   pre-print
We introduce a Recursive INsertion-based Encoder (RINE), a novel approach for semantic parsing in task-oriented dialog.  ...  At the generation time, the model constructs the semantic parse tree by recursively inserting the predicted non-terminal labels at the predicted positions until termination.  ...  Finally in Section 2.3 we describe the generation procedure of semantic parse tree given the input utterance.  ... 
arXiv:2109.04500v2 fatcat:tdre55dztzh3bgvbudqxlyerze

Semantic Parsing in Task-Oriented Dialog with Recursive Insertion-Based Encoder

Elman Mansimov, Yi Zhang
2022 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
We introduce a Recursive INsertion-based Encoder (RINE), a novel approach for semantic parsing in task-oriented dialog.  ...  At the generation time, the model constructs the semantic parse tree by recursively inserting the predicted non-terminal labels at the predicted positions until termination.  ...  Validity of Generated Trees In this section we compare the validity of the semantic parse trees generated by our and baseline approaches.  ... 
doi:10.1609/aaai.v36i10.21355 fatcat:sfh7kylegrgoflncyuey3zebzi

Evaluating the Impact of Model Scale for Compositional Generalization in Semantic Parsing [article]

Linlu Qiu, Peter Shaw, Panupong Pasupat, Tianze Shi, Jonathan Herzig, Emily Pitler, Fei Sha, Kristina Toutanova
2022 arXiv   pre-print
Meanwhile, recent work has shown considerable improvements on many NLP tasks from model scaling. Can scaling up model size also improve compositional generalization in semantic parsing?  ...  We observe that fine-tuning generally has flat or negative scaling curves on out-of-distribution compositional generalization in semantic parsing evaluations.  ...  compositional generalization in semantic parsing?  ... 
arXiv:2205.12253v1 fatcat:wpffg53mlfce5hp5dt3ta4ri4m

Empirically-motivated Generalizations of CCG Semantic Parsing Learning Algorithms

Jesse Glass, Alexander Yates
2014 Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics  
Learning algorithms for semantic parsing have improved drastically over the past decade, as steady improvements on benchmark datasets have shown.  ...  In this paper we investigate whether they can generalize to a novel biomedical dataset that differs in important respects from the traditional geography and air travel benchmark datasets.  ...  In an effort to investigate and improve the generalization capacity of existing learning algorithms for semantic parsing, we develop a novel, natural experimental setting, and we test whether current semantic  ... 
doi:10.3115/v1/e14-1037 dblp:conf/eacl/GlassY14 fatcat:aczjodkqa5axtagccnbbvhniq4

Compositional Generalization in Dependency Parsing [article]

Emily Goodwin, Siva Reddy, Timothy J. O'Donnell, Dzmitry Bahdanau
2022 arXiv   pre-print
To test compositional generalization in semantic parsing, Keysers et al. (2020) introduced Compositional Freebase Queries (CFQ).  ...  Dependency parsing, however, lacks a compositional generalization benchmark.  ...  To address the issues with the CFQ semantic parsing benchmark, we studied compositional generalization in syntactic parsing.  ... 
arXiv:2110.06843v2 fatcat:yhlen62dmva67jahhwbnpoyucu

GLR parsing with multiple grammars for natural language queries

Helen Meng, Po-Chui Luk, Kui Xu, Fuliang Weng
2002 ACM Transactions on Asian Language Information Processing  
Parser composition imparts a robust parsing capability in our framework, and hence obtains a higher understanding performance when compared to using a single GLR parser.  ...  A parser composition technique then combines the parsers' outputs to produce an overall parse that is the same as the output parse of single parser.  ...  By comparison, our semantic interpreter only generates the key-value pair "time=tonight" in our frame, and the PI is not available to us for incorporation in our generated semantic frames.  ... 
doi:10.1145/568954.568956 fatcat:4f3as7wqqreetpceow7fjnhdni
« Previous Showing results 1 — 15 out of 28,212 results