Pith. sign in

REVIEW 3 major objections 5 minor 80 references

Semantic Hypergraphs

T0 review · 3 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 Five hypergraph patterns, drawn from a recursive sentence representation, outperform specialized open information extraction systems.

desk verdict The formalism and parser are worth a serious look, but the headline OIE rank is a post-hoc fit to a 57-sentence benchmark, so the empirical claim is unproven. read the letter →

arxiv 1908.10784 v2 pith:UCFALYYM submitted 2019-08-28 cs.IR cs.AIcs.CL

Telmo Menezes , Camille Roth This is my paper · ORCID
classification cs.IRcs.AIcs.CL
keywords naturallanguageunderstandingknowledgerepresentationinformationextractioninferencesystemsexplainableartificialintelligencehypergraphsrecursiveopen
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 aims to establish that natural language can be translated into a recursive hypergraph form, called a semantic hypergraph, in which every sentence becomes a single nested hyperedge, and that this representation makes hard extraction tasks easy: five hand-generalized hyperedge patterns are enough to rank first in a recent Open Information Extraction benchmark, ahead of specialized statistical systems. It also claims that the NL-to-SH parser produces completely correct hypergraphs for 79% of tested sentences across five text categories, with most errors being minor defects rather than total failures. A sympathetic reader would care because the representation is open and human-readable while still adaptive, using standard machine-learning components for the classification step and a symbolic search tree for the grouping step, a combination that could serve as a shared lingua franca between human analysts and computational systems. The deeper goal is to give social science and other corpus researchers an inspectable instrument for detecting claims, conflicts, and conceptual structure at scale.

What carries the argument

The load-bearing object is the recursive ordered hyperedge: an ordered list of elements that may themselves be hyperedges, so that statements can contain statements and concepts can be built from concepts. The formal syntax requires the first element of every non-atomic hyperedge to be a connector, and the type-inference rules determine the resulting hyperedge type. Parsing works by repeatedly matching these inference rules as unordered patterns over the sequence of typed atoms, using the heuristic $h$ to resolve competing matches by preferring directly dependency-linked tokens; the same hyperedge syntax then doubles as a pattern-matching language, with variables, wildcards, and replacement rules, which is what makes the five OIE patterns possible.

What would settle it

A decisive test: rerun the OIE benchmark with the grouping heuristic replaced by random selection among pattern matches while keeping token types unchanged. If the five patterns no longer beat the baselines, the dependency-connectivity heuristic carries the benchmark result; if they still do, the representation itself is sufficient.

Watch

Extended reading notes

Core claim

The central claim is that sentence meaning can be captured by a recursive ordered hyperedge whose first element is always a connector (predicate, modifier, builder, trigger, or conjunction), with a compact eight-type system and a small set of type-inference rules that close the syntax. Starting from tokens classified into atomic types, a search-tree algorithm applies those inference rules bottom-up, guided by a heuristic that prefers groupings whose tokens are directly linked in the dependency tree. The paper reports that this pipeline yields completely correct hyperedges for 79% of 500 held-out sentences, and that on a public Open Information Extraction benchmark the five most general patterns derived from the 50 most frequent hyperedge patterns rank first with an F1 of 0.365, edging out seven dedicated OIE systems.

Load-bearing premise

The parser's grouping stage assumes that tokens directly linked in the grammatical dependency tree belong together in the same semantic unit, so if grammatical connectivity stops being a reliable guide for longer or more ambiguous sentences, the grouping step fails even when every token is classified correctly.

Editorial extensions

If this is right

  • If the central claim holds, a transparent, human-readable representation can match specialized statistical systems on a standard OIE benchmark using only five patterns, so openness does not have to cost competitive performance.
  • The reported 79% complete-parse rate across fiction, nonfiction, news, science, and Wikipedia indicates that the representation is not tuned to a single text genre.
  • The conjunction decomposition rules and the co-occurrence based coreference method should allow corpus-scale actor-claim-topic analysis, as demonstrated on news headlines.
  • Because the type system is designed to cover all grammatical relations in Universal Dependencies, the same hyperedge syntax should transfer to other languages once a small amount of token-classification training data is provided.

Reading between the lines

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

  • The pattern-learning loop in Section 5.3 is human-in-the-loop; a direct extension would be to automate the generalization step with a search tree over wildcard substitutions and test whether the same five OIE patterns can be rediscovered without human feedback.
  • If the dependency-connectivity heuristic is the real bottleneck behind the 79% figure, replacing it with a learned grouping model could raise complete-parse accuracy while keeping the SH representation intact.
  • The Universal Dependencies completeness argument suggests a clean language-transfer experiment: annotate a few hundred sentences in a typologically distant language and check whether the same eight types suffice.
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

3 major / 5 minor

Summary. The paper introduces Semantic Hypergraphs (SH), a recursive, ordered, typed hypergraph representation of natural-language meaning, together with a two-stage parser (an α-stage token-type classifier and a β-stage search-tree structuring algorithm), a pattern language for knowledge inference, and a set of applications including conjunction decomposition, open information extraction, concept taxonomy inference, coreference resolution, and a news-corpus claim/conflict case study. The central empirical claims are that the NL-to-SH parser achieves 79% completely correct parses across five text categories (Section 4.3, Table 4) and that five simple SH hyperedge patterns suffice to rank first on the WiRe57 Open Information Extraction benchmark (Section 5.4, Tables 6 and 7). The paper also emphasizes the open-adaptive nature of the formalism, its human readability, and the availability of the Graphbrain open-source implementation.

Significance. If the empirical claims were properly supported, this would be a valuable demonstration that a small, human-readable symbolic representation can compete with specialized statistical systems on a standard extraction task, and that a knowledge representation of this kind is usable across several downstream text-analysis tasks. The paper is also notable for shipping an open-source implementation (Graphbrain), for defining a concrete type system with explicit inference rules, and for attempting an integrated social-science-oriented application on a large news-title corpus. These strengths make the work of potential interest to the NLP and computational social science communities. However, the significance currently rests on evaluations whose validity needs reinforcement, especially the OIE ranking result, so the contribution is promising rather than fully established.

major comments (3)
  1. [Section 5.4, Tables 6 and 7] The rank-first OIE claim is a post-hoc fit to the evaluation set, not a held-out result. The five patterns were obtained by taking the 50 most common patterns from a Wikipedia hypergraph, validating 36 of them against the benchmark's annotation guidelines, compressing them to 5, and then ordering them "by decreasing contribution to F1" with the ranks in Table 6 and the final rank in Table 7 all computed on WiRe57 itself. No split, cross-validation, or significance testing is reported. The margin over MinIE is 0.365 vs 0.358 F1, i.e. 0.007 on a 57-sentence benchmark, so a small number of extractions can flip the rank. The paper's disclaimer that it does not claim to be the best OIE system softens the wording but does not repair the inference, because this result is the main external quantitative evidence for the expressiveness of SH. The authors should either evaluate patterns selected on a training portion and tested on held-out sentences, or explicitly reframe the result as a post-hoc descriptive comparison without the rank-first claim.
  2. [Section 4.3, Table 4] The parser evaluation is based on 500 sentences rated by the authors themselves, with no inter-annotator agreement, no confidence intervals, and no release of the annotated judgments. The categories "completely correct", "defect", and "completely wrong" involve subjective semantic judgments, and the reported 79% complete-correct rate and the per-category rates would be considerably more convincing if at least a second annotator scored a subset and Cohen's kappa or a similar agreement measure were reported. This concern is secondary to the OIE issue, but it directly affects the reliability of the parser's headline accuracy and of every downstream experiment that consumes the parser output.
  3. [Section 4.2, Algorithm 1] The β-stage heuristic is the load-bearing assumption that grammatical connectivity in the dependency tree correlates with semantic grouping, but the paper does not quantify how often the heuristic selects a suboptimal grouping when the dependency parse is correct, nor how sensitive the final parse quality is to this choice. The statement in Section 4.3 that "a correct α classification and dependency parse tree always lead to the construction of an SH that preserves the meaning" is based on the same 500-sentence author-rated sample and is not supported by a controlled analysis of the heuristic's decisions. A small ablation or error analysis separating α errors, dependency-parse errors, and β heuristic errors would materially strengthen the parser validation.
minor comments (5)
  1. [Section 2] There is a typo in the related-work discussion: "we sill see how SHs can represent semantic information" should be "we will see".
  2. [Section 5.4] In the description of patterns 3, 4, and 5, the text says "REL is split into two or thee variables" and later "REL1, REL2 and REL2"; the first should be "three" and the second should likely be "REL1, REL2 and REL3".
  3. [Section 4.1] The phrase "one of thealpha categories" has a missing space; it should read "one of the α categories".
  4. [Figure 1] The right panel would benefit from explicit axis labels and error bars or confidence intervals, since the text makes claims about significant differences between text categories.
  5. [Section 3.2] The definition of the recursive hyperedge set E_V is stated compactly and would be clearer with an explicit base case distinguishing atoms from non-atomic hyperedges; currently the notation is ambiguous about whether atoms are included in E_V by definition or by the V ∪ E_V disjunct.

Circularity Check

1 steps flagged · score 6.0 of 10

OIE '5 patterns rank first' is benchmark-fitted: the patterns were ordered by F1 on WiRe57 and the rank is measured on the same benchmark.

  1. fitted input called prediction [Section 5.4, Table 6 caption and Table 7]
    "Open Information Extraction patterns, ordered by decreasing contribution to F1 (presented cumulatively). Ranks correspond to the rank achieved in the benchmark of Table 7 by using patterns up to the given line."

    The headline claim says that five simple hyperedge patterns are sufficient to rank first in the WiRe57 OIE benchmark. The table caption states that the patterns are ordered by decreasing contribution to F1 and that the ranks are the benchmark ranks obtained by using patterns up to each line. Thus the pattern set, its cardinality, and its order were all selected by evaluating on WiRe57 itself, and the final rank in Table 7 is computed on that same benchmark. The result is therefore a post-hoc fit to the test set rather than an independent prediction: the F1 values and rank used as the evidence were also the objective used for selection.

full rationale

Most of the paper's derivation chain is self-contained: the SH type system and syntax are defined formally; the α-stage classifier is trained on annotated sentences and tested on random splits; the β-stage is evaluated on 500 held-out sentences; and the concept, coreference, and claim/conflict applications are checked by human inspection or by comparison to the corpus. Self-citations are pointers to the released Graphbrain code, not load-bearing evidence for the formal claims. The one significant circular step is the OIE headline. In Section 5.4 the five patterns are obtained by generalizing common hyperedges, validating candidates against the benchmark's annotation guidelines, compressing the valid ones, and then ordering them 'by decreasing contribution to F1', with Table 6 explicitly tying each rank to the benchmark of Table 7. The same WiRe57 benchmark supplies the final F1 and rank, so the claim that five simple patterns can rank first is a test-set selection result, not a held-out prediction. This OIE comparison is the paper's main external quantitative evidence, so the circularity is load-bearing for that specific claim. Other parts of the paper remain independent, which keeps the overall score at 6 rather than higher.

Assumptions & free parameters 4 free parameters · 5 assumptions · 4 invented entities

The central contributions depend on the eight-type taxonomy, the Table 2 composition rules, the β heuristic, and a series of completeness claims about Universal Dependencies. These are introduced by the authors and validated through the paper's own experiments; no machine-checked proof or external benchmark of representation completeness is provided.

free parameters (4)
  • coreference threshold θ = 0.7
    Section 6.4: 'We set the threshold to the values θ = .7 and θ′ = .05, that we verified empirically to produce good results.' Chosen on the same corpus used for the reported coreference results.
  • coreference threshold θ′ = 0.05
    Same sentence in Section 6.4; controls minimum degree-to-deep-degree ratio for seed assignment, tuned empirically without held-out validation.
  • OIE pattern set = 5 patterns
    Section 5.4: from 50 common patterns, 36 were deemed valid using the benchmark annotation guidelines, then compressed to 5. This is a hand-selected structural choice made for the target benchmark and ordered by F1 contribution.
  • Conjunction decomposition rules = 3 rules
    Section 5.2 defines three manually written rules for conjunction resolution; they are used as a preprocessing step in OIE and claim analysis without separate evaluation.
assumptions (5)
  • domain assumption Universal Dependencies catalogs all grammatical constructs, and the Appendix A mapping covers each construct in SH.
    The paper argues SH completeness from a one-example-per-UD-relation table (Appendix A), not from a proof that every dependency tree produces a valid or meaning-preserving hyperedge.
  • domain assumption Dependency-tree connectivity is a reliable guide to semantic grouping in the β-stage heuristic.
    Algorithm 1 uses heuristic h that prefers groupings whose tokens are directly connected in the dependency tree; the paper calls this a hypothesis in Section 4.2, and no ablation tests it.
  • ad hoc to paper The type-inference rules in Table 2 are the correct composition rules for natural-language meaning.
    These rules define the SH formalism itself; they are introduced by the authors and evaluated only through downstream tasks, not derived from an independent theory.
  • domain assumption Manual 'correct/defect/wrong' ratings are reliable indicators of semantic equivalence.
    Section 4.3 reports parser quality based on author judgments of 500 sentences; no inter-annotator agreement or detailed rubric is provided.
  • domain assumption spaCy's POS, dependency and NER outputs are accurate enough to support the parser.
    The pipeline inherits spaCy's reported accuracies; the paper notes that badly-structured dependency trees significantly hurt β, but does not quantify this at scale.
invented entities (4)
  • Semantic Hypergraph formalism (recursive ordered typed hyperedges)
    purpose: Represent natural-language meaning in a human-readable, machine-processable structure
    The core invention of the paper; its value is supported only by the paper's own experiments, with no external falsifiable handle outside this work.
  • Special builder atom (+/B)
    purpose: Represent compound nouns and implicit concept construction
    Introduced in Section 3.3 to handle compound nouns; its semantics are defined by the authors and used in OIE patterns.
  • Implicit conjunction connector (:/J)
    purpose: Join loosely related constructs and serve as the fallback in the β-stage
    Introduced in Section 3.3 and used in Algorithm 1 when no type-inference pattern matches.
  • lemma/J connector
    purpose: Link surface word forms to lemmas for morphological inference
    Introduced in Section 6.1; used to connect atoms such as saying/P to say/P.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semantic Hypergraphs." pith.science (2026). https://pith.science/paper/UCFALYYM

@misc{pith2026190810784,
  author       = {Pith},
  title        = {Pith review of: Semantic Hypergraphs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UCFALYYM}},
  note         = {Machine review of arXiv:1908.10784}
}
read the original abstract

Approaches to Natural language processing (NLP) may be classified along a double dichotomy open/opaque - strict/adaptive. The former axis relates to the possibility of inspecting the underlying processing rules, the latter to the use of fixed or adaptive rules. We argue that many techniques fall into either the open-strict or opaque-adaptive categories. Our contribution takes steps in the open-adaptive direction, which we suggest is likely to provide key instruments for interdisciplinary research. The central idea of our approach is the Semantic Hypergraph (SH), a novel knowledge representation model that is intrinsically recursive and accommodates the natural hierarchical richness of natural language. The SH model is hybrid in two senses. First, it attempts to combine the strengths of ML and symbolic approaches. Second, it is a formal language representation that reduces but tolerates ambiguity and structural variability. We will see that SH enables simple yet powerful methods of pattern detection, and features a good compromise for intelligibility both for humans and machines. It also provides a semantically deep starting point (in terms of explicit meaning) for further algorithms to operate and collaborate on. We show how modern NLP ML-based building blocks can be used in combination with a random forest classifier and a simple search tree to parse NL to SH, and that this parser can achieve high precision in a diversity of text categories. We define a pattern language representable in SH itself, and a process to discover knowledge inference rules. We then illustrate the efficiency of the SH framework in a variety of tasks, including conjunction decomposition, open information extraction, concept taxonomy inference and co-reference resolution, and an applied example of claim and conflict analysis in a news corpus.

Figures

Figures reproduced from arXiv: 1908.10784 by the authors.

Figure 1
Figure 1. Left: accuracy of the α-classifier, comparing several feature sets; all includes all features, GA a features set obtained with a genetic algorithm, F3 is the outcome of iterative ablation with p < .005 and F5 with p < .05. Right: accuracy by source text category using F5. and preserving the order of the remainder of the hyper￾edge i.e., “(of/B capital/C germany/C)”, which conforms to the second inference rule of tab… view at source ↗
Figure 2
Figure 2. (a) Dependency parse tree with dependency labels (green) and fine grained part-of-speech tags (red). (b) α-stage classification of atom types. (c) β-stage structuring of sentence by iterative application of the patterns from table 2. A non-selected pattern is greyed-out. pattern (M C), and the sequence (capital/C, of/B, ger￾many/C) matches the pattern (B C C+). We thus rely on the above-mentioned heuristic function,… view at source ↗
Figure 3
Figure 3. Pattern learning template and example with two passes. At the end of the second pass, the pattern [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Example of coreference resolution. On the left panel we can see the co-occurrence graph and its compo [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]
Figure 5
Figure 5. Figure 5: Two examples of relations starting with either a claim or a conflict predicate. [PITH_FULL_IMAGE:figures/full_fig_p022_5.png]
Figure 6
Figure 6. Figure 6: Network of conflicts between actors over the topic “Syria”. Arrows point from the originator of the conflict [PITH_FULL_IMAGE:figures/full_fig_p027_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

80 extracted references · 74 canonical work pages

  1. [1]

    RDFa in XHTML: Syntax and processing

    Adida, B., Birbeck, M., McCarron, S., Pemberton, S., 2008. RDFa in XHTML: Syntax and processing. Recommenda- tion, W3C 7

    work page 2008

  2. [2]

    Agirre, E., Soroa, A., 2009. Personalizing pagerank for word sense disambiguation, in: Proceedings of the 12th Conference of the European Chapter of the Association for Computational Linguistics, Association for Computa- tional Linguistics. pp. 33–41

    work page 2009

  3. [3]

    What’ s in a semantic net- work?, in: Proceedings of the 20th annual meeting on As- sociation for Computational Linguistics, Association for Computational Linguistics

    Allen, J.F ., Frisch, A.M., 1982. What’ s in a semantic net- work?, in: Proceedings of the 20th annual meeting on As- sociation for Computational Linguistics, Association for Computational Linguistics. pp. 19–27

    work page 1982

  4. [4]

    Lever- aging linguistic structure for open domain information extraction, in: Proc

    Angeli, G., Premkumar, M.J.J., Manning, C.D., 2015. Lever- aging linguistic structure for open domain information extraction, in: Proc. 53rd Annual Meeting of the Associa- tion for Computational Linguistics and 7th Intl. Joint Con- ference on Natural Language Processing, pp. 344–354

    work page 2015

  5. [5]

    Top2vec: Distributed representations of topics

    Angelov, D., 2020. Top2vec: Distributed representations of topics. arXiv preprint arXiv:2008.09470

    arXiv 2020

  6. [6]

    Dbpedia: A nucleus for a web of open data, in: The semantic web

    Auer, S., Bizer, C., Kobilarov, G., Lehmann, J., Cyganiak, R., Ives, Z., 2007. Dbpedia: A nucleus for a web of open data, in: The semantic web. Springer, pp. 722–735

    work page 2007

  7. [7]

    Abstract meaning representation for sembanking, in: Proc

    Banarescu, L., Bonial, C., Cai, S., Georgescu, M., Grif- fitt, K., Hermjakob, U., Knight, K., Koehn, P ., Palmer, M., Schneider, N., 2013. Abstract meaning representation for sembanking, in: Proc. 7th linguistic annotation workshop and interoperability with discourse, pp. 178–186

    work page 2013

  8. [8]

    Networks beyond pairwise interactions: structure and dynamics

    Battiston, F ., Cencetti, G., Iacopini, I., Latora, V ., Lucas, M., Patania, A., Young, J.G., Petri, G., 2020. Networks beyond pairwise interactions: structure and dynamics. Physics Reports 874, 1–92

    work page 2020

Show all 80 references
  1. [9]

    Hypergraphs: combinatorics of finite sets

    Berge, C., 1984. Hypergraphs: combinatorics of finite sets. volume 45 ofNorth-Holland mathematical library. North- Holland, Amsterdam

    1984

  2. [10]

    Publishing on the se- mantic web

    Berners-Lee, T ., Hendler, J., 2001. Publishing on the se- mantic web. Nature 410, 1023–1024

    2001

  3. [11]

    Probabilistic topic models

    Blei, D.M., 2012. Probabilistic topic models. Communica- tions of the ACM 55, 77–84

    2012

  4. [12]

    Latent dirichlet allocation

    Blei, D.M., Ng, A.Y., Jordan, M.I., 2003. Latent dirichlet allocation. Journal of machine Learning research 3, 993– 1022

    2003

  5. [13]

    Directed recursive labelnode hyper- graphs: A new representation-language

    Boley, H., 1977. Directed recursive labelnode hyper- graphs: A new representation-language. Artificial Intel- ligence 9, 49–85

    1977

  6. [14]

    Comet: Commonsense transform- ers for automatic knowledge graph construction

    Bosselut, A., Rashkin, H., Sap, M., Malaviya, C., Celikyil- maz, A., Choi, Y., 2019. Comet: Commonsense transform- ers for automatic knowledge graph construction. arXiv 1906.05317

    2019 arXiv

  7. [15]

    Language models are few-shot learners

    Brown, T .B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P ., Neelakantan, A., Shyam, P ., Sastry, G., Askell, A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T ., Child, R., Ramesh, A., Ziegler, D.M., Wu, J., Winter, C., Hesse, C., Chen, M., Sigler, E.,...

    2020 arXiv

  8. [16]

    Net- work properties of folksonomies

    Cattuto, C., Schmitz, C., Baldassarri, A., Servedio, V .D., Loreto, V ., Hotho, A., Grahl, M., Stumme, G., 2007. Net- work properties of folksonomies. Ai Communications 20, 245–262

    2007

  9. [17]

    Chavalarias, D., Wallach, J.D., Li, A.H.T ., Ioannidis, J.P .,

  10. [18]

    It depends: Depen- dency parser comparison using a web-based evaluation tool, in: Proc

    Choi, J.D., Tetreault, J., Stent, A., 2015. It depends: Depen- dency parser comparison using a web-based evaluation tool, in: Proc. 53rd Annual Meeting of the Association for Computational Linguistics and 7th Intl. Joint Conference on Natural Language Processing, pp. 387–396

    2015

  11. [19]

    A unified architecture for natural language processing: Deep neural networks with multitask learning, in: Proceedings of the 25th interna- tional conference on Machine learning, pp

    Collobert, R., Weston, J., 2008. A unified architecture for natural language processing: Deep neural networks with multitask learning, in: Proceedings of the 25th interna- tional conference on Machine learning, pp. 160–167

    2008

  12. [20]

    Clausie: clause-based open information extraction, in: Proceedings of the 22nd international conference on World Wide Web, pp

    Del Corro, L., Gemulla, R., 2013. Clausie: clause-based open information extraction, in: Proceedings of the 22nd international conference on World Wide Web, pp. 355– 366

    2013

  13. [21]

    BERT: Pre-training of deep bidirectional transformers for language understanding, in: Proc

    Devlin, J., Chang, M.W ., Lee, K., Toutanova, K., 2019. BERT: Pre-training of deep bidirectional transformers for language understanding, in: Proc. 2019 Conference of the North American Chapter of the Association for Computa- tional Linguistics: Human Language Technologies, AC...

    2019

  14. [22]

    Revealing social structure from texts: Meta-matrix text analysis as a novel method for network text analysis, in: Causal mapping for research in information technology

    Diesner, J., Carley, K., 2005. Revealing social structure from texts: Meta-matrix text analysis as a novel method for network text analysis, in: Causal mapping for research in information technology. IGI Global, pp. 81–108

    2005

  15. [23]

    Some properties of ordered hypergraphs

    Eslahchi, C., Rahimi, A., 2007. Some properties of ordered hypergraphs. Matematiˇ cki vesnik 59, 9–13. 28

    2007

  16. [24]

    Open information extraction from the web

    Etzioni, O., Banko, M., Soderland, S., Weld, D.S., 2008. Open information extraction from the web. Communi- cations of the ACM 51, 68–74

    2008

  17. [25]

    Identifying re- lations for open information extraction, in: Proc

    Fader, A., Soderland, S., Etzioni, O., 2011. Identifying re- lations for open information extraction, in: Proc. Conf. on empirical methods in natural language processing, ACL. pp. 1535–1545

    2011

  18. [26]

    Minie: minimizing facts in open information extraction, in: Proc

    Gashteovski, K., Gemulla, R., Del Corro, L., 2017. Minie: minimizing facts in open information extraction, in: Proc. of the 2017 Conf. on Empirical Methods in Natural Lan- guage Processing, Association for Computational Linguis- tics. p. 2620–2630

    2017

  19. [27]

    A network approach to topic models

    Gerlach, M., Peixoto, T .P ., Altmann, E.G., 2018. A network approach to topic models. Science advances 4, eaaq1360

    2018

  20. [28]

    Patterns, hypergraphs and embod- ied general intelligence, in: IJCNN’06 International Joint Conference on Neural Networks, IEEE

    Goertzel, B., 2006. Patterns, hypergraphs and embod- ied general intelligence, in: IJCNN’06 International Joint Conference on Neural Networks, IEEE. pp. 451–458

    2006

  21. [29]

    Text as data: The promise and pitfalls of automatic content analysis meth- ods for political texts

    Grimmer, J., Stewart, B.M., 2013. Text as data: The promise and pitfalls of automatic content analysis meth- ods for political texts. Political Analysis 21, 267–297

    2013

  22. [30]

    Opencog: A software frame- work for integrative artificial general intelligence, in: Arti- ficial General Intelligence, IOS Press

    Hart, D., Goertzel, B., 2008. Opencog: A software frame- work for integrative artificial general intelligence, in: Arti- ficial General Intelligence, IOS Press. pp. 468–472

    2008

  23. [31]

    An improved non- monotonic transition system for dependency parsing, in: EMNLP’15 Proc

    Honnibal, M., Johnson, M., et al., 2015. An improved non- monotonic transition system for dependency parsing, in: EMNLP’15 Proc. of the 2015 Conf, on Empirical Methods in Natural Language Processing, pp. 1373–1378

    2015

  24. [32]

    HyperGraphDB: A generalized graph database, in: Shen, H.T ., Pei, J., Özsu, M.T ., Zou, L., Lu, J., Ling, T .W ., Yu, G., Zhuang, Y., Shao, J

    Iordanov, B., 2010. HyperGraphDB: A generalized graph database, in: Shen, H.T ., Pei, J., Özsu, M.T ., Zou, L., Lu, J., Ling, T .W ., Yu, G., Zhuang, Y., Shao, J. (Eds.), Web- Age Information Management, Springer Berlin Heidel- berg, Berlin, Heidelberg. pp. 25–36

    2010

  25. [33]

    Distributed representations of sentences and documents, in: International conference on machine learning, PMLR

    Le, Q., Mikolov, T ., 2014. Distributed representations of sentences and documents, in: International conference on machine learning, PMLR. pp. 1188–1196

    2014

  26. [34]

    Wire57 : A fine- grained benchmark for open information extraction, in: Friedrich, A., Zeyrek, D., Hoek, J

    Léchelle, W ., Gotti, F ., Langlais, P ., 2019. Wire57 : A fine- grained benchmark for open information extraction, in: Friedrich, A., Zeyrek, D., Hoek, J. (Eds.), Proc. of the 13th Linguistic Annotation Workshop, LAW at ACL 2019, Flo- rence, Italy, August 1, 2019, Association ...

    2019

  27. [35]

    Resources for the open information extraction benchmark WiRe57, com- panion to Léchelle et al., 2019

    Léchelle, W ., Gotti, F ., Langlais, P ., 2020. Resources for the open information extraction benchmark WiRe57, com- panion to Léchelle et al., 2019. URL: https://github. com/rali-udem/WiRe57

    2020

  28. [36]

    Cyc: toward programs with common sense

    Lenat, D.B., Guha, R.V ., Pittman, K., Pratt, D., Shepherd, M., 1990. Cyc: toward programs with common sense. Communications of the ACM 33, 30–49

    1990

  29. [37]

    Co-word maps and topic modeling: A comparison using small and medium- sized corpora (n < 1,000)

    Leydesdorff, L., Nerghes, A., 2017. Co-word maps and topic modeling: A comparison using small and medium- sized corpora (n < 1,000). Journal of the American Society for Information Science and Technology 68, 1024–1035

    2017

  30. [38]

    Argumentation mining: State of the art and emerging trends

    Lippi, M., Torroni, P ., 2016. Argumentation mining: State of the art and emerging trends. ACM Transactions on In- ternet Technology (TOIT) 16, 10

    2016

  31. [39]

    Understanding wordscores

    Lowe, W ., 2008. Understanding wordscores. Political Anal- ysis 16, 356–371

    2008

  32. [40]

    Mausam, Schmitz, M., Soderland, S., Bart, R., Etzioni, O.,

  33. [41]

    Open information extraction sys- tems and downstream applications, in: Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence, pp

    Mausam, M., 2016. Open information extraction sys- tems and downstream applications, in: Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence, pp. 4074–4077

    2016

  34. [42]

    Recursive functions of symbolic ex- pressions and their computation by machine, part i

    McCarthy, J., 1960. Recursive functions of symbolic ex- pressions and their computation by machine, part i. Com- munications of the ACM 3, 184–195

    1960

  35. [43]

    Mihalcea, R., 2005. Unsupervised large-vocabulary word sense disambiguation with graph-based algorithms for sequence data labeling, in: Proceedings of Human Lan- guage Technology Conference and Conference on Empiri- cal Methods in Natural Language Processing, pp. 411–418

    2005

  36. [44]

    Textrank: Bringing order into text, in: EMNLP’04 Proc

    Mihalcea, R., Tarau, P ., 2004. Textrank: Bringing order into text, in: EMNLP’04 Proc. 2004 Conf. on Empirical Meth- ods in Natural Language Processing, pp. 404–411

    2004

  37. [45]

    Efficient estimation of word representations in vector space

    Mikolov, T ., Chen, K., Corrado, G., Dean, J., 2013. Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781

    2013 arXiv

  38. [46]

    Wordnet: a lexical database for english

    Miller, G.A., 1995. Wordnet: a lexical database for english. Communications of the ACM 38, 39–41

    1995

  39. [47]

    Fightin’words: Lexical feature selection and evalua- tion for identifying the content of political conflict

    Monroe, B.L., Colaresi, M.P ., Quinn, K.M., 2008. Fightin’words: Lexical feature selection and evalua- tion for identifying the content of political conflict. Political Analysis 16, 372–403

    2008

  40. [48]

    Distant reading

    Moretti, F ., 2013. Distant reading. Verso Books

    2013

  41. [49]

    Statement map: reducing web information credibility noise through opinion classification, in: Pro- ceedings of the fourth workshop on Analytics for noisy unstructured text data, ACM

    Murakami, K., Nichols, E., Mizuno, J., Watanabe, Y., Ma- suda, S., Goto, H., Ohki, M., Sao, C., Matsuyoshi, S., Inui, K., et al., 2010. Statement map: reducing web information credibility noise through opinion classification, in: Pro- ceedings of the fourth workshop on Analytic...

    2010

  42. [50]

    A survey of named en- tity recognition and classification

    Nadeau, D., Sekine, S., 2007. A survey of named en- tity recognition and classification. Lingvisticae Investiga- tiones 30, 3–26

    2007

  43. [51]

    Graph connectivity mea- sures for unsupervised word sense disambiguation., in: IJCAI, pp

    Navigli, R., Lapata, M., 2007. Graph connectivity mea- sures for unsupervised word sense disambiguation., in: IJCAI, pp. 1683–1688

    2007

  44. [52]

    Univer- sal dependencies v1: A multilingual treebank collection, in: Proceedings of the Tenth International Conference on 29 Language Resources and Evaluation (LREC’16), pp

    Nivre, J., De Marneffe, M.C., Ginter, F ., Goldberg, Y., Ha- jic, J., Manning, C.D., McDonald, R., Petrov, S., Pyysalo, S., Silveira, N., Tsarfaty, R., Zeman, D., 2016. Univer- sal dependencies v1: A multilingual treebank collection, in: Proceedings of the Tenth International ...

    2016

  45. [53]

    The proposi- tion bank: An annotated corpus of semantic roles

    Palmer, M., Gildea, D., Kingsbury, P ., 2005. The proposi- tion bank: An annotated corpus of semantic roles. Com- putational linguistics 31, 71–106

    2005

  46. [54]

    Opinion mining and sen- timent analysis

    Pang, B., Lee, L., et al., 2008. Opinion mining and sen- timent analysis. Foundations and Trends in Information Retrieval 2, 1–135

    2008

  47. [55]

    Deep contextualized word representations

    Peters, M.E., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., Zettlemoyer, L., 2018. Deep contextualized word representations. arXiv 1802.05365

    2018 arXiv

  48. [56]

    Transforming de- pendency structures to logical forms for semantic pars- ing

    Reddy, S., Täckström, O., Collins, M., Kwiatkowski, T ., Das, D., Steedman, M., Lapata, M., 2016. Transforming de- pendency structures to logical forms for semantic pars- ing. Transactions of the Association for Computational Linguistics 4, 127–140

    2016

  49. [57]

    Ritter, A., Clark, S., Etzioni, O., et al., 2011. Named entity recognition in tweets: an experimental study, in: Proceed- ings of the Conference on Empirical Methods in Natural Language Processing, Association for Computational Lin- guistics. pp. 1524–1534

    2011

  50. [58]

    Socio-semantic frameworks

    Roth, C., 2013. Socio-semantic frameworks. Advances in Complex Systems 16, 1350013

    2013

  51. [59]

    Ruiz, P ., Plancq, C., Poibeau, T ., 2016. More than word cooccurrence: Exploring support and opposition in inter- national climate negotiations with semantic parsing, in: LREC: The 10th Language Resources and Evaluation Con- ference, pp. 1902–1907

    2016

  52. [60]

    Term-weighting approaches in automatic text retrieval

    Salton, G., Buckley, C., 1988. Term-weighting approaches in automatic text retrieval. Information processing & management 24, 513–523

    1988

  53. [61]

    Atomic: An atlas of machine commonsense for if-then reasoning, in: Proceedings of the AAAI Conference on Ar- tificial Intelligence, pp

    Sap, M., Le Bras, R., Allaway, E., Bhagavatula, C., Lourie, N., Rashkin, H., Roof, B., Smith, N.A., Choi, Y., 2019. Atomic: An atlas of machine commonsense for if-then reasoning, in: Proceedings of the AAAI Conference on Ar- tificial Intelligence, pp. 3027–3035

    2019

  54. [62]

    The semantic web revisited

    Shadbolt, N., Berners-Lee, T ., Hall, W ., 2006. The semantic web revisited. IEEE intelligent systems 21, 96–101

    2006

  55. [63]

    Mea- suring ideological proportions in political speeches, in: EMLP’13 Proc

    Sim, Y., Acree, B.D., Gross, J.H., Smith, N.A., 2013. Mea- suring ideological proportions in political speeches, in: EMLP’13 Proc. 2013 Conf. on Empirical Methods in Nat- ural Language Processing, pp. 91–101

    2013

  56. [64]

    Modern information retrieval: A brief overview

    Singhal, A., 2001. Modern information retrieval: A brief overview. IEEE Data Eng. Bull. 24, 35–43

    2001

  57. [65]

    A machine learn- ing approach to coreference resolution of noun phrases

    Soon, W .M., Ng, H.T ., Lim, D.C.Y., 2001. A machine learn- ing approach to coreference resolution of noun phrases. Computational linguistics 27, 521–544

    2001

  58. [66]

    Principles of semantic networks: Explo- rations in the representation of knowledge

    Sowa, J.F ., 2014. Principles of semantic networks: Explo- rations in the representation of knowledge. Morgan Kauf- mann

    2014

  59. [67]

    Models documentation

    spaCy, 2020. Models documentation. URL: https:// spacy.io/models/en

    2020

  60. [68]

    Conceptnet 5.5: An open multilingual graph of general knowledge, in: Pro- ceedings of the AAAI Conference on Artificial Intelligence, pp

    Speer, R., Chin, J., Havasi, C., 2017. Conceptnet 5.5: An open multilingual graph of general knowledge, in: Pro- ceedings of the AAAI Conference on Artificial Intelligence, pp. 4444–4451

    2017

  61. [69]

    Text mining: Classifi- cation, clustering, and applications

    Srivastava, A.N., Sahami, M., 2009. Text mining: Classifi- cation, clustering, and applications. CRC Press

    2009

  62. [70]

    Handbook on ontologies

    Staab, S., Studer, R., 2010. Handbook on ontologies. Springer Science & Business Media

    2010

  63. [71]

    Get- ting more out of syntax with props

    Stanovsky, G., Ficler, J., Dagan, I., Goldberg, Y., 2016. Get- ting more out of syntax with props. arXiv 1603.01648

    2016 arXiv

  64. [72]

    Parliamentarization of popular contention in great britain, 1758-1834

    Tilly, C., 1997. Parliamentarization of popular contention in great britain, 1758-1834. Theory and Society 26, 245– 273

    1997

  65. [73]

    Parsing, semantic networks, and political authority using syntactic analysis to extract semantic relations from dutch newspaper articles

    Van Atteveldt, W ., Kleinnijenhuis, J., Ruigrok, N., 2008. Parsing, semantic networks, and political authority using syntactic analysis to extract semantic relations from dutch newspaper articles. Political Analysis 16, 428–446

    2008

  66. [74]

    Clause analysis: using syntactic information to automatically extract source, subject, and predicate from texts with an application to the 2008–2009 Gaza War

    Van Atteveldt, W ., Sheafer, T ., Shenhav, S.R., Fogel-Dror, Y., 2017. Clause analysis: using syntactic information to automatically extract source, subject, and predicate from texts with an application to the 2008–2009 Gaza War. Po- litical Analysis 25, 207–222

    2017

  67. [75]

    Wikidata: A new platform for collab- orative data collection, in: Proceedings of the 21st inter- national conference on World Wide Web, pp

    Vrandeˇ ci´ c, D., 2012. Wikidata: A new platform for collab- orative data collection, in: Proceedings of the 21st inter- national conference on World Wide Web, pp. 1063–1064

    2012

  68. [76]

    Wang, H., Can, D., Kazemzadeh, A., Bar, F ., Narayanan, S.,

  69. [77]

    Large-scale computerized text analysis in political science: Opportunities and chal- lenges

    Wilkerson, J., Casas, A., 2017. Large-scale computerized text analysis in political science: Opportunities and chal- lenges. Annual Review of Political Science 20, 529–544. A Mapping Universal Stanford De- pendencies to hyperedges We used the Universal Stanford Dependencies [5...

    2017

  70. [79]

    A system for real-time twitter sentiment analysis of 2012 US presidential election cycle, in: Proceedings of the ACL 2012 System Demonstrations, Association for Com- putational Linguistics. pp. 115–120

    2012

  71. [2012]

    Open language learning for information extraction, in: Proceedings of the 2012 joint conference on empiri- cal methods in natural language processing and compu- tational natural language learning, pp. 523–534

    2012

  72. [2016]

    Jama 315, 1141–1148

    Evolution of reporting p values in the biomedical literature, 1990-2015. Jama 315, 1141–1148

    1990

Pith tools

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