Pith. sign in

REVIEW 4 major objections 5 minor 48 references

SpatialNLI: A Spatial Domain Natural Language Interface to Databases Using Spatial Comprehension

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read SpatialNLI resolves ambiguous spatial phrases with an external comprehension model, injects the predicted type symbols into the question, and reports 90.4% denotation accuracy on Geoquery and 100% on Restaurant.

desk verdict The paper has a sensible idea—external spatial comprehension feeding type symbols into a seq2seq parser—but the unstated SC train/test split and a three-question margin make the SOTA claim not yet credible. read the letter →

arxiv 1908.10917 v1 pith:2IM7LYNV submitted 2019-08-28 cs.CL

Jingjing Li , Wenlu Wang , Wei-Shinn Ku , Yingtao Tian , Haixun Wang This is my paper · ORCID
classification cs.CL MSC 68T5068P15
keywords naturallanguageinterfacespatialsemanticssemanticparsingmachinecomprehensionsequence-to-sequencemodeldatabasesdataaugmentationquerytranslation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that a spatial natural language interface can handle context-dependent spatial meaning by splitting the job in two: an external spatial comprehension model decides what an ambiguous phrase means, such as whether "Mississippi" is a state or a river, and a sequence-to-sequence parser handles the remaining question structure. The disambiguating meaning is inserted into the question as special symbols, so the parser never has to guess it from sparse training data. On the Geoquery benchmark the full system reaches 90.4% denotation accuracy, above the 88.2% reported for a strong syntax-based parser, and on Restaurant it matches the best prior result at 100%. If this is right, the approach offers a general recipe for adding domain-specific encyclopedic knowledge to a neural semantic parser without retraining the parser on that knowledge.

What carries the argument

The load-bearing mechanism is a two-stage pipeline: SpatialMapper, a phrase-detection step that finds keywords and values by exact string match, edit distance, and pretrained word-vector cosine similarity against database contents; and the spatial comprehension model, which uses stacked LSTMs over the question and the candidate type with a shared attentive LSTM layer to predict true or false for each question-phrase-type triple, with the ambiguous phrase enclosed in special symbols. SpatialInjection writes the chosen type and a placeholder for the value into the question as special tokens, so the sequence-to-sequence parser can copy a single token instead of generating long entity names, and query recovery substitutes the original text back. The paper also augments the training data by shuffling prepositional phrases and recombining existing questions.

What would settle it

Run the full SpatialNLI pipeline on Geoquery test questions in which every ambiguous value is replaced by a synonym that is absent from the database and far from the original in the embedding space, such as a nickname for a river. If denotation accuracy drops sharply, the central claim is bounded by the phrase-detection step rather than by the comprehension model; if accuracy holds, the pipeline generalizes beyond exact or near-exact matches.

Watch

Extended reading notes

Core claim

The central claim is that spatial semantics should be computed outside the semantic parser and then physically inserted into the input text. The paper constructs an external spatial comprehension model that, for each ambiguous phrase, predicts the phrase's type by attending over the words of the question; the predicted type is inserted as a symbol such as "riverid" between special markers, the value itself is replaced by a placeholder symbol, the modified question is fed to a sequence-to-sequence model with a copying mechanism, and the resulting logical form is recovered by substituting the original phrases back. The reported outcome is 90.4% denotation accuracy on Geoquery, exceeding the 88.2% of a strong syntax-based parser, and 100% on Restaurant. Ablation experiments attribute most of the gain to symbol injection and type feeding, with spatial comprehension adding 4% on Geoquery and 3.9% on Restaurant.

Load-bearing premise

The pipeline assumes the phrase-detection step finds every keyword and value that matters and that the comprehension model predicts the right type for each ambiguous phrase; a miss or a wrong type propagates unchanged into the final query because the parser copies the injected symbol.

Editorial extensions

If this is right

  • A neural semantic parser can be improved on spatial benchmarks by externalizing entity-type disambiguation instead of learning it from sparse question-query pairs.
  • Replacing multi-token entity names with placeholder symbols shortens the generation path and should reduce entity-copy errors, which the Restaurant ablation supports.
  • Prepositional-phrase shuffling and question recombination add training structures without new human annotation.
  • Because the comprehension model reaches 98.1% question-level accuracy on the Geoquery test set, almost all ambiguous point-of-interest cases are resolved before the parser sees the question.
  • Jointly training on Geoquery and Restaurant reaches 90.7% on Geoquery, suggesting the injected symbols serve as a shared representation across the two domains.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A wrong type judgment by the comprehension model is never corrected downstream: the sequence-to-sequence model copies the injected symbol, so a single misclassification can turn into an executable query against the wrong table, and residual error on ambiguous questions is likely concentrated there.
  • The phrase-detection step is a hidden precondition: a phrase missed by string match, edit distance, or word-vector similarity never reaches the comprehension model, so the end-to-end upper bound depends on the detector as much as on the neural components.
  • The same externalize-and-inject pattern could transfer to other domains with encyclopedic ambiguity, such as biomedical entity types or temporal expressions, whenever the type vocabulary is fixed and discoverable from the database schema.
  • The paper does not state whether the comprehension model's train/test split matches the parser's 600/280 Geoquery split, so the component accuracies and the end-to-end accuracy may not have been measured under identical conditions.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. This paper proposes SpatialNLI, a natural-language interface to spatial databases. The system first detects spatial keywords and values with a mapper based on exact string match, edit distance, and GloVe cosine similarity; an external spatial comprehension model (a bidirectional attentive LSTM) then classifies the type of ambiguous point-of-interest names; the predicted type and detected terms are injected as special symbols into the question; a seq2seq parser with copying translates the modified question into a logical form; and a recovery module replaces symbols with the original phrases. The paper evaluates on Geoquery and Restaurant, reporting 90.4% and 100% denotation accuracy, respectively, and ablates the comprehension model, type feeding, information injection, copy mechanism, and data augmentation.

Significance. If the reported numbers are trustworthy, the paper makes a useful contribution by separating contextual spatial-semantic disambiguation from structural parsing and by showing that simple symbol injection can improve an existing seq2seq semantic parser. The evaluation is systematic: all ablation components are named, case studies illustrate the injection/recovery loop, and the code is released. However, the central claim of outperforming the state of the art rests on a 1.1-point margin on 280 Geoquery test questions, with no error bars or significance testing, and the training/test separation of the spatial comprehension model is not reported. These issues are load-bearing for the headline comparison, while the Restaurant result ties, rather than beats, the existing 100% baseline.

major comments (4)
  1. [Section 7.3 / Table 1] The train/test split used for the spatial comprehension (SC) model is not reported. Because the SC training labels in Figure 6 are read off the ground-truth logical form, the claimed 4.0-point gain from the '- Spatial Comprehension' ablation and the 90.4% headline accuracy are fully interpretable only if the SC model never sees a test question's logical form during training and if the SC split coincides with the parser's standard 600/280 Geoquery split. Please state the exact SC split, the label-generation rule, and confirm that no test-question-derived record is used in SC training.
  2. [Table 2 / Section 7.4] The headline comparison lacks error bars and significance testing. On 280 Geoquery test questions, 90.4% versus JL16's 89.3% is about three questions, and the margin over TRANX's 88.2% is similarly narrow. The '- Spatial Comprehension' ablation at 86.4% is the output of a random-guess procedure whose variance is not reported. Please report means and standard deviations over multiple random seeds and a significance test, for example bootstrap or McNemar, for the comparisons that support the claim of outperforming the state of the art.
  3. [Section 7.4 / Figure 10] The paper does not analyze the failure path that its own case study demonstrates: when the SC model or the random-guess baseline selects the wrong type (stateid instead of riverid), the injected symbol is copied by the seq2seq model and the recovered query is wrong. Because the end-to-end gain is attributed to the comprehension model, the paper should report how many test questions have SC type errors and how often those errors propagate to denotation mismatches, rather than only reporting the aggregate Acc_qu.
  4. [Section 4.1 / Algorithm 2] SpatialMapper relies on exact string match, an edit-distance threshold tau_ed, and a GloVe cosine threshold tau_sem, but the paper reports no sensitivity analysis and no count of test questions for which a relevant keyword or value is missed. A phrase missed by the mapper is never disambiguated or injected, so the contribution of the comprehension model is bounded by the mapper's recall. Please report detection recall on the test sets and, if possible, the effect of varying the thresholds.
minor comments (5)
  1. [Section 3, paragraph following Figure 4] The sentence claiming that 'stateid will be inferred instead of riverid' appears to invert the correct and incorrect types: with the ground-truth type stateid(Mississippi), the failure case is riverid being inferred instead of stateid.
  2. [Section 7.3, Table 1 paragraph] The statement that 'we prove that Accrcd and Accqu are optimized simultaneously' is not supported by any proof in the paper; either provide a short argument or rephrase it as an empirical observation.
  3. [Abstract and Section 1] The abstract and introduction say SpatialNLI outperforms state-of-the-art methods, but Table 2(b) shows a tie with FKZ18 at 100% on Restaurant; please qualify the claim, for example 'matches or exceeds' or restrict the outperformance claim to Geoquery.
  4. [Section 7.1, Restaurant dataset] The paper does not report the training/test split used for Restaurant; since Table 2 reports a test accuracy, please specify the split or clarify how the 251 pairs are divided.
  5. [Section 5.1 / Table 2] The seq2seq model is described as following prior work [9], but [9] is not included in Table 2; please clarify whether [9] is an applicable baseline and, if so, add it to the comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the end-to-end accuracy claim is tested on external benchmark datasets, and the spatial comprehension model is a separately trained classifier rather than a restatement of the parser's output.

full rationale

The paper's central claim is the denotation accuracy reported in Table 2 on Geoquery and Restaurant, both external benchmark datasets, so the headline result is not defined in terms of the system's own inputs. The spatial comprehension model is trained on question/POI-type records (Figure 6) and evaluated separately in Table 1; at inference, its predicted type is injected into the question, and the '- Spatial Comprehension' ablation replaces this prediction with a random guess. The paper quotes no construction by which the test-time type is taken from the ground-truth logical form. The seq2seq parser with symbol injection follows the same authors' prior work [9], but that citation supplies an architectural technique; the contribution is evaluated by ablations against external denotation accuracy rather than by citing [9] as proof, and the copying mechanism is also credited to the external JL16 [37]. Query recovery (Algorithm 4) is deterministic replacement of generated symbols with the originally detected phrases, which is bookkeeping rather than a predicted quantity. The unstated alignment of the spatial-comprehension train/test split with the parser's 600/280 Geoquery split is a potential experimental-reporting concern, but no quoted passage shows the spatial comprehension model being fitted to test logical forms, and no fitted parameter is renamed as a prediction. Under the rule that circularity must be exhibited by a specific reduction, no circular step is established.

Assumptions & free parameters 6 free parameters · 6 assumptions · 1 invented entities

The central method relies on benchmark assumptions and a set of hand-chosen hyperparameters. The most consequential choices are the unreported detection thresholds in the SpatialMapper, the positive-sample replication factor for training the comprehension model, and the seq2seq architecture sizes. The special symbols are invented method-internal tokens, not independently testable entities.

free parameters (6)
  • semantic distance threshold tau_sem = not reported in paper (example: semantic_distance('place','spot') < 0.368)
    Algorithm 2 uses this threshold to accept database matches; without a stated value the detection component is not fully specified.
  • edit distance threshold tau_ed = not reported
    Algorithm 2 uses this threshold for fuzzy keyword matching; no value is given in the paper.
  • positive sample replication factor = 2
    Section 7.3 replicates positive samples 2x to balance spatial comprehension training; chosen by hand.
  • spatial comprehension hidden units = 200
    Section 7.3 reports 200 hidden units for the comprehension model.
  • seq2seq GRU hidden sizes = encoder 800, decoder 1600
    Section 7.4 reports encoder hidden size 800 and decoder hidden size 800*2.
  • beam search width = 5
    Section 7.4 reports beam search with width 5 for inference.
assumptions (6)
  • domain assumption Geoquery and Restaurant gold logical forms are correct, and denotation match is a valid measure of query accuracy.
    The evaluation in Section 7.4 adopts the standard benchmark and the denotation-match metric from [37], without independent verification.
  • domain assumption Schema names and column values in the spatial databases align with the constants and keywords in the target logical forms.
    SpatialMapper in Algorithm 2 detects keywords and values by matching question tokens against database contents; if the database lexicon differs from the query language, detection fails.
  • domain assumption POI type for an ambiguous phrase is determined by the question context and can be learned from training data.
    The spatial comprehension model in Section 4.2 assumes context is sufficient to disambiguate type, and that the auxiliary training labels derived from gold logical forms are accurate.
  • domain assumption GloVe embedding similarity captures semantic relatedness between natural language words and schema keywords.
    Section 4.1 uses semantic_distance with GloVe to match for example 'place' to 'spot'; the paper offers no tuning or validation of this assumption beyond one example.
  • domain assumption Moving a prepositional phrase to the start or end of a sentence preserves the logical meaning of the question.
    Section 6 data augmentation shuffles PP position and treats the new sentence as equivalent to the original; no checks for changed word sense or grammaticality are described.
  • domain assumption The spatial comprehension model's train and test splits align with the parser's splits, so its labels do not leak test answers.
    Section 7.3 reports train/test accuracy but does not explicitly state the split correspondence; if misaligned, the reported end-to-end accuracy could be optimistic.
invented entities (1)
  • Special symbol tokens <k_i>, <v_i>, <eok>, <eov>, and <@>
    purpose: Mark detected keywords, values, and ambiguous phrases in the question before seq2seq translation and after recovery.
    These tokens are invented by the authors and have meaning only within the training set and recovery mapping; their behavior is not validated outside the two benchmarks.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SpatialNLI: A Spatial Domain Natural Language Interface to Databases Using Spatial Comprehension." pith.science (2026). https://pith.science/paper/2IM7LYNV

@misc{pith2026190810917,
  author       = {Pith},
  title        = {Pith review of: SpatialNLI: A Spatial Domain Natural Language Interface to Databases Using Spatial Comprehension},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2IM7LYNV}},
  note         = {Machine review of arXiv:1908.10917}
}
read the original abstract

A natural language interface (NLI) to databases is an interface that translates a natural language question to a structured query that is executable by database management systems (DBMS). However, an NLI that is trained in the general domain is hard to apply in the spatial domain due to the idiosyncrasy and expressiveness of the spatial questions. Inspired by the machine comprehension model, we propose a spatial comprehension model that is able to recognize the meaning of spatial entities based on the semantics of the context. The spatial semantics learned from the spatial comprehension model is then injected to the natural language question to ease the burden of capturing the spatial-specific semantics. With our spatial comprehension model and information injection, our NLI for the spatial domain, named SpatialNLI, is able to capture the semantic structure of the question and translate it to the corresponding syntax of an executable query accurately. We also experimentally ascertain that SpatialNLI outperforms state-of-the-art methods.

Figures

Figures reproduced from arXiv: 1908.10917 by the authors.

Figure 1
Figure 1. Three examples show that the spatial semantics is [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. SpatialNLI overview. 3 CHALLENGES OF SPATIAL NLI I. Sparse training data. Even though data augmentation is a fea￾sible solution to sparse training data, it is likely that the deep model will be forced to handle unseen questions that are not covered by data augmentation, and required to support transfer learning. II. Spatial semantics ambiguity. A unique feature of spatial ques￾tions is its expressiveness in the spat… view at source ↗
Figure 3
Figure 3. Spatial POI ambiguity Theoretically, a powerful data augmentation should be able to address the first challenge; however, such data augmentation strat￾egy is rare in practice. Moreover, a seq2seq model is designed to translate a sentence, it is reasonable that it fails to capture the con￾text precisely and infer the correct spatial semantics. For example, in the question “How many rivers does Mississippi have ?” (sh… view at source ↗
Figures from the paper (6 more)
Figure 5
Figure 5. Figure 5: Spatial Comprehension Model sparse training data, which relieves the burden of collecting large training sets. By our definition, Spatial Comprehension is spatial se￾mantic understanding using machine comprehension. Our strategy is to exploit pre-trained Glove embeddin…
Figure 6
Figure 6. Figure 6: Spatial Comprehension Model Training Samples [PITH_FULL_IMAGE:figures/full_fig_p005_6.png]
Figure 7
Figure 7. Figure 7: We will validate in the experiment section that our type [PITH_FULL_IMAGE:figures/full_fig_p006_7.png]
Figure 8
Figure 8. Figure 8: Two examples of data augmentation 7 EXPERIMENTAL VALIDATION 7.1 Experimental Settings Configuration All our experiments are conducted on a machine equipped with 2 Intel CPU E5-2670 v3 running at 2.3GHz with 256GB of RAM and 2 NVIDIA Tesla K80 GPUs. Dataset To evaluate …
Figure 9
Figure 9. Figure 9: Three examples of data augmentation is the name of a state based on the spatial database. Given this example, we change it to new questions, in which the word “Florida” is replaced by the name of other states in the database. Here, the word “Florida” is replaced by “Rh…
Figure 10
Figure 10. Figure 10: Spatial Comprehension Case Study. *+SC means using Spatial Comprehension, -SC means without. [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 45 canonical work pages

  1. [9]

    A Transfer-Learnable Natural Language Interface for Databases

    Wenlu Wang, Yingtao Tian, Hongyu Xiong, Haixun Wang, and Wei-Shinn Ku. A transfer-learnable natural language interface for databases. arXiv preprint arXiv:1809.02649, 2018

    work page Pith review arXiv 2018

  2. [1]

    Natural language interfaces to databases–an introduction

    Ion Androutsopoulos, Graeme D Ritchie, and Peter Thanisch. Natural language interfaces to databases–an introduction. Natural language engineering, 1(1):29–81, 1995

    work page 1995

  3. [2]

    Dataset for a neural natural language interface for databases (NNLIDB)

    Florin Brad, Radu Cristian Alexandru Iacob, Ionel-Alexandru Hosu, and Traian Rebedea. Dataset for a neural natural language interface for databases (NNLIDB). In Proceedings of the Eighth International Joint Conference on Natural Language Processing, IJCNLP, pages 906–914, 2017

    work page 2017

  4. [3]

    An End-to-end Neural Natural Language Interface for Databases

    Prasetya Utama, Nathaniel Weir, Fuat Basik, Carsten Binnig, Ugur Çetintemel, Benjamin Hättasch, Amir Ilkhechi, Shekar Ramaswamy, and Arif Usta. An end- to-end neural natural language interface for databases. CoRR, abs/1804.00401, 2018

    work page Pith review arXiv 2018

  5. [4]

    Nalir: an interactive natural language interface for querying relational databases

    Fei Li and Hosagrahar Visvesvaraya Jagadish. Nalir: an interactive natural language interface for querying relational databases. In International Conference on Management of Data, SIGMOD , pages 709–712, 2014

    work page 2014

  6. [5]

    Yunyao Li, Huahai Yang, and H. V. Jagadish. Nalix: an interactive natural language interface for querying XML. In Proceedings of the ACM SIGMOD International Conference on Management of Data , pages 900–902, 2005

    work page 2005

  7. [6]

    Ana-Maria Popescu, Oren Etzioni, and Henry A. Kautz. Towards a theory of natural language interfaces to databases. In Proceedings of the 8th International Conference on Intelligent User Interfaces , 2003

    work page 2003

  8. [7]

    Mittal, and Fatma Özcan

    Diptikalyan Saha, Avrilia Floratou, Karthik Sankaranarayanan, Umar Farooq Minhas, Ashish R. Mittal, and Fatma Özcan. ATHENA: an ontology-driven system for natural language querying over relational data stores. PVLDB, 9(12):1209– 1220, 2016

    work page 2016

Show all 48 references
  1. [8]

    Learning a neural semantic parser from user feedback

    Srinivasan Iyer, Ioannis Konstas, Alvin Cheung, Jayant Krishnamurthy, and Luke Zettlemoyer. Learning a neural semantic parser from user feedback. InProceedings of the 55th Annual Meeting of the Association for Computational Linguistics, ACL , pages 963–973, 2017

    2017

  2. [10]

    A cross-domain natural language interface to databases using adversarial text method

    Wenlu Wang. A cross-domain natural language interface to databases using adversarial text method. Database, 1:q2, 2019

    2019

  3. [11]

    Efficient parallel spatial skyline evaluation using mapreduce

    Wenlu Wang, Ji Zhang, M-T Sun, and W-S Ku. Efficient parallel spatial skyline evaluation using mapreduce. In Proceedings of the 20th international conference on extending database technology , 2017

    2017

  4. [12]

    A scalable spatial skyline evaluation system utilizing parallel independent region groups.The VLDB Journal The International Journal on Very Large Data Bases , 28(1):73–98, 2019

    Wenlu Wang, Ji Zhang, Min-Te Sun, and Wei-Shinn Ku. A scalable spatial skyline evaluation system utilizing parallel independent region groups.The VLDB Journal The International Journal on Very Large Data Bases , 28(1):73–98, 2019

    2019

  5. [13]

    Dynamic indoor navigation with bayesian filters

    Wenlu Wang and Wei-Shinn Ku. Dynamic indoor navigation with bayesian filters. SIGSPATIAL Special, 8(3):9–10, 2017

    2017

  6. [14]

    Recommendation-based smart indoor navigation

    Wenlu Wang and Wei-Shinn Ku. Recommendation-based smart indoor navigation. In Proceedings of the Second International Conference on Internet-of-Things Design and Implementation, pages 311–312. ACM, 2017

    2017

  7. [15]

    Spatial semantics

    Jordan Zlatev. Spatial semantics. The Oxford handbook of cognitive linguistics , pages 318–350, 2007

    2007

  8. [16]

    TRANX: A transition-based neural abstract syntax parser for semantic parsing and code generation

    Pengcheng Yin and Graham Neubig. TRANX: A transition-based neural abstract syntax parser for semantic parsing and code generation. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, EMNLP , pages 7–12, 2018

    2018

  9. [17]

    Bidirectional attention flow for machine comprehension

    Min Joon Seo, Aniruddha Kembhavi, Ali Farhadi, and Hannaneh Hajishirzi. Bidirectional attention flow for machine comprehension. In 5th International Conference on Learning Representations, ICLR , 2017

    2017

  10. [18]

    Machine comprehension using match-lstm and answer pointer

    Shuohang Wang and Jing Jiang. Machine comprehension using match-lstm and answer pointer. In 5th International Conference on Learning Representations, ICLR , 2017

    2017

  11. [19]

    Zettlemoyer and Michael Collins

    Luke S. Zettlemoyer and Michael Collins. Learning to map sentences to logical form: Structured classification with probabilistic categorial grammars. In UAI ’05, Proceedings of the 21st Conference in Uncertainty in Artificial Intelligence , 2005

    2005

  12. [20]

    Online learning of relaxed ccg grammars for parsing to logical form

    Luke Zettlemoyer and Michael Collins. Online learning of relaxed ccg grammars for parsing to logical form. InProceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL), 2007

    2007

  13. [21]

    Zettlemoyer, Sharon Goldwater, and Mark Steedman

    Tom Kwiatkowski, Luke S. Zettlemoyer, Sharon Goldwater, and Mark Steedman. Inducing probabilistic CCG grammars from logical form with higher-order uni- fication. In Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, EMNLP, pages 1223–1233, 2010

    2010

  14. [22]

    Zettlemoyer, Sharon Goldwater, and Mark Steedman

    Tom Kwiatkowski, Luke S. Zettlemoyer, Sharon Goldwater, and Mark Steedman. Lexical generalization in CCG grammar induction for semantic parsing. In Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, EMNLP, pages 1512–1523, 2011

    2011

  15. [23]

    Scaling semantic parsers with on-the-fly ontology matching

    Tom Kwiatkowski, Eunsol Choi, Yoav Artzi, and Luke Zettlemoyer. Scaling semantic parsers with on-the-fly ontology matching. In Proceedings of the 2013 conference on empirical methods in natural language processing , 2013

    2013

  16. [24]

    Morpho-syntactic lex- ical generalization for ccg semantic parsing

    Adrienne Wang, Tom Kwiatkowski, and Luke Zettlemoyer. Morpho-syntactic lex- ical generalization for ccg semantic parsing. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) , 2014

    2014

  17. [25]

    Type-driven incremental semantic parsing with polymorphism

    Kai Zhao and Liang Huang. Type-driven incremental semantic parsing with polymorphism. In NAACL HLT 2015, The 2015 Conference of the North Ameri- can Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1416–1421, 2015

    2015

  18. [26]

    Jordan, and Dan Klein

    Percy Liang, Michael I. Jordan, and Dan Klein. Learning dependency-based compositional semantics. Computational Linguistics, 39(2):389–446, 2013

    2013

  19. [27]

    Learning to parse database queries using inductive logic programming

    John M Zelle and Raymond J Mooney. Learning to parse database queries using inductive logic programming. InProceedings of the national conference on artificial intelligence, pages 1050–1055, 1996

    1996

  20. [28]

    Automated construction of database interfaces: Integrating statistical and relational learning for semantic parsing

    Lappoon R Tang and Raymond J Mooney. Automated construction of database interfaces: Integrating statistical and relational learning for semantic parsing. In Proceedings of the 2000 Joint SIGDAT conference on Empirical methods in natu- ral language processing and very large cor...

    2000

  21. [29]

    Using multiple clause constructors in inductive logic programming for semantic parsing

    Lappoon R Tang and Raymond J Mooney. Using multiple clause constructors in inductive logic programming for semantic parsing. In European Conference on Machine Learning, pages 466–477. Springer, 2001

    2001

  22. [30]

    A statistical semantic parser that integrates syntax and semantics

    Ruifang Ge and Raymond J Mooney. A statistical semantic parser that integrates syntax and semantics. In Proceedings of the ninth conference on computational natural language learning. Association for Computational Linguistics, 2005

    2005

  23. [31]

    Recurrent continuous translation models

    Nal Kalchbrenner and Phil Blunsom. Recurrent continuous translation models. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1700–1709, 2013

    2013

  24. [32]

    Sequence to sequence learning with neural networks

    Ilya Sutskever, Oriol Vinyals, and Quoc V Le. Sequence to sequence learning with neural networks. In Advances in neural information processing systems , pages 3104–3112, 2014

    2014

  25. [33]

    Learning phrase repre- sentations using rnn encoder-decoder for statistical machine translation

    Kyunghyun Cho, Bart Van Merriënboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. Learning phrase repre- sentations using rnn encoder-decoder for statistical machine translation. arXiv preprint arXiv:1406.1078, 2014

    2014 arXiv

  26. [34]

    Kummerfeld, Li Zhang, Karthik Ra- manathan, Sesh Sadasivam, Rui Zhang, and Dragomir R

    Catherine Finegan-Dollak, Jonathan K. Kummerfeld, Li Zhang, Karthik Ra- manathan, Sesh Sadasivam, Rui Zhang, and Dragomir R. Radev. Improving text-to-sql evaluation methodology. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics, ACL , p...

    2018

  27. [35]

    Abstract syntax networks for code generation and semantic parsing

    Maxim Rabinovich, Mitchell Stern, and Dan Klein. Abstract syntax networks for code generation and semantic parsing. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, ACL , pages 1139–1149, 2017

    2017

  28. [36]

    Language to logical form with neural attention

    Li Dong and Mirella Lapata. Language to logical form with neural attention. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, ACL, 2016

    2016

  29. [37]

    Data recombination for neural semantic parsing

    Robin Jia and Percy Liang. Data recombination for neural semantic parsing. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, ACL, 2016

    2016

  30. [38]

    Transfer learning for neural semantic parsing

    Xing Fan, Emilio Monti, Lambert Mathias, and Markus Dreyer. Transfer learning for neural semantic parsing. In Proceedings of the 2nd Workshop on Representation Learning for NLP, pages 48–56, 2017

    2017

  31. [39]

    Neural architectures for multilingual semantic parsing

    Raymond Hendy Susanto and Wei Lu. Neural architectures for multilingual semantic parsing. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 38–44, 2017

    2017

  32. [40]

    Neural semantic parsing over multiple knowledge-bases

    Jonathan Herzig and Jonathan Berant. Neural semantic parsing over multiple knowledge-bases. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, ACL, pages 623–628, 2017

    2017

  33. [41]

    Seq2sql: Generating struc- tured queries from natural language using reinforcement learning

    Victor Zhong, Caiming Xiong, and Richard Socher. Seq2sql: Generating struc- tured queries from natural language using reinforcement learning. CoRR, abs/1709.00103, 2017

    2017 arXiv

  34. [42]

    Annotation of spatial relations in natural language

    Qijun Shen, Xueying Zhang, and Wenming Jiang. Annotation of spatial relations in natural language. In 2009 International Conference on Environmental Science and Information Application Technology, volume 3, pages 418–421. IEEE, 2009

    2009

  35. [43]

    A linguis- tic ontology of space for natural language processing

    John A Bateman, Joana Hois, Robert Ross, and Thora Tenbrink. A linguis- tic ontology of space for natural language processing. Artificial Intelligence, 174(14):1027–1071, 2010

    2010

  36. [44]

    Relational learning for spatial relation extraction from natural language

    Parisa Kordjamshidi, Paolo Frasconi, Martijn Van Otterlo, Marie-Francine Moens, and Luc De Raedt. Relational learning for spatial relation extraction from natural language. In International Conference on Inductive Logic Programming , pages 204–220. Springer, 2011

    2011

  37. [45]

    Extracting spatial information from place descriptions

    Arbaz Khan, Maria Vasardani, and Stephan Winter. Extracting spatial information from place descriptions. In COMP@ SIGSPATIAL, page 62, 2013

    2013

  38. [46]

    Encoding spatial relations from natural language

    Tiago Ramalho, Tomáš Kocisk`y, Frederic Besse, SM Eslami, Gábor Melis, Fabio Viola, Phil Blunsom, and Karl Moritz Hermann. Encoding spatial relations from natural language. arXiv preprint arXiv:1807.01670, 2018

    2018 arXiv

  39. [47]

    Glove: Global vectors for word representation

    Jeffrey Pennington, Richard Socher, and Christopher Manning. Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP) , pages 1532–1543, 2014

    2014

  40. [48]

    Building a semantic parser overnight

    Yushi Wang, Jonathan Berant, and Percy Liang. Building a semantic parser overnight. In Proceedings of the 53rd Annual Meeting of the Association for Compu- tational Linguistics and the 7th International Joint Conference on Natural Language Processing, volume 1, pages 1332–1342, 2015

    2015

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.