MétaCan
Menu
Back to cohort
Record W1196909593 · doi:10.20381/ruor-12790

Semantic relations across syntactic levels

2004· dissertation· en· W1196909593 on OpenAlexaff
Viviana A. Nastase

Bibliographic record

VenueuO Research (University of Ottawa) · 2004
Typedissertation
Languageen
FieldComputer Science
TopicNatural Language Processing Techniques
Canadian institutionsUniversity of Ottawa
Fundersnot available
KeywordsUtteranceVerbRelation (database)LinguisticsPoint (geometry)Computer scienceOrder (exchange)Natural language processingArtificial intelligenceMathematicsPhilosophy

Abstract

fetched live from OpenAlex

In order to make sense of a message conveyed to us via a spoken or written utterance, we understand what things are talked about, and how they are connected. From this point of view, do these sentences convey different messages? (1) I will arrive at 11 am. and I will arrive when you arrive. (2) I will meet you in the office. and I will meet you where we met last time. (3) Sweets before dinner spoil your appetite. and (4) Eating sweets before dinner spoils your appetite. I will arrive at a certain point in time: at 11 am., or when you arrive. I will meet you at a certain place: in the office, or where we met last time. We can talk about sweets and mean eating sweets. Literature review suggests that the relations exemplified by these pairs of sentences are different, because they connect different types of syntactic units. The first relation in each pair connects a verb and one of its arguments, the second---two clauses. Such distinctions are artificial. Semantic relations link concepts, and will surface on the syntactic level on which the concepts they connect surface. We aim to give an account of semantic relations that does not depend on syntactic levels. We will justify a unified view of semantic relations across syntactic levels. Such a view has a positive effect on text analysis. It will allow us to gather evidence for a particular semantic relation from all levels at which it appears. Having such information that is not separated according to syntactic levels will allow a text analysis and knowledge acquisition system to use at each processing step, all the evidence previously gathered. We will show that this translates into faster learning and better results. We can take semantic relation analysis onto another level. We can look for descriptions of concepts connected by a specific semantic relation to find what characteristics or features of the concepts connected make them interact in this way. (1) blue book, happy person, interesting study; (2) paper bag, wooden chair, iron gate; (3) oak tree, cumulus cloud, flounder fish. Blue, happy, interesting are properties, and paper, wood, iron are materials. Oak is a specific type of tree, cumulus is a type of cloud, and flounder is a type of fish. We will use ontologies to find similarities between concepts that explain or give us indications about the semantic relations in which they are involved. All these aspects we explore serve to improve text analysis. We propose a uniform processing of texts that allows us to extracts pairs of concepts that interact, and to describe this interaction through semantic relations.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.001
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Theoretical or conceptual · Consensus signal: Theoretical or conceptual
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.524
Threshold uncertainty score0.883

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0010.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0010.001
Science and technology studies0.0010.000
Scholarly communication0.0000.001
Open science0.0030.000
Research integrity0.0000.001
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.048
GPT teacher head0.363
Teacher spread0.315 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; a candidate call from one teacher head, not a consensus.

The models applied no category: nothing in the taxonomy fit this work.
Study designTheoretical or conceptual
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations6
Published2004
Admission routes1
Has abstractyes

Explore more

Same venueuO Research (University of Ottawa)Same topicNatural Language Processing TechniquesFrench-language works237,207