VerbAtlas 1.0 - Semantic Role Labeling and beyond! c/o Roberto Navigli

VerbAtlas 1.0 - Semantic Role Labeling and beyond!

We are proud to announce that VerbAtlas 1.0 (http://verbatlas.org) is finally available for download at http://verbatlas.org/download. Developed in the Sapienza NLP group (http://nlp.uniroma1.it), the multilingual Natural Language Processing research team at the Sapienza University of Rome, VerbAtlas is a novel large-scale manually-crafted semantic resource for wide-coverage, intelligible and scalable Semantic Role Labeling. The goal of VerbAtlas is to manually cluster WordNet synsets that share similar semantics into sets of semantically-coherent frames. The main features are:

• 466 semantically-coherent frames using 26 cross-frame VerbNet-inspired semantic roles for their argument structure.
• Available both for download and via RESTful API.
• Full coverage of WordNet 3.0 verb synsets (13,000+).
• Complete linkage to BabelNet 4.0, which supports 280+ languages (new version to come later this year!).
• Manual mapping to PropBank of all CoNLL-2009 and CoNLL-2012 dataset occurrences (5000+ mappings).
• Selectional preferences: the superconcept most probably associated with a semantic role in a frame (e.g. food for the patient role of the EAT frame).
• Default/shadow arguments: arguments logically implied or already incorporated into a verb.
• Implicit arguments: arguments that are implicit in the argument structure of a verb.

And much more! Please check out our EMNLP 2019 paper:

A. Di Fabio, S. Conia, R. Navigli “VerbAtlas: a Novel Large-Scale Verbal Semantic Resource and Its Application to Semantic Role Labeling”, EMNLP-IJCNLP 2019, pp. 627-637

and http://verbatlas.org for more details!

A state-of-the-art span-based SRL system which uses VerbAtlas and PropBank jointly is also available from the same website (just type a sentence!).

VerbAtlas is an output of the MOUSSE ERC Consolidator Grant n. 726487 and the ELEXIS project no.731015. Babelscape proudly developed the online interface and API, provides the infrastructure for maintaining the service and works on the improvement of the resource. 

Enjoy!

The Sapienza NLP group

=====================================
Roberto Navigli

Dipartimento di Informatica
Sapienza University of Rome
Viale Regina Elena 295b (building G, second floor)
00161 Roma Italy
Home Page: http://wwwusers.di.uniroma1.it/~navigli

Sapienza NLP Group: http://nlp.uniroma1.it
My Source:
The Semantic-web list @w3.org



What are / should be the current core verbs allocated to each of the care (knowledge) domains of Hodges' model?

individual

|

INTERPERSONAL : SCIENCES

humanistic ----------------------------------------------- mechanistic

SOCIOLOGY : POLITICAL

|

group - population

*personalise	scientify
socialise	politicise

*(intra- inter-)personalise?


From above introduction to VerbAtlas:

"A state-of-the-art span-based SRL system which uses VerbAtlas and PropBank jointly is also available from the same website (just type a sentence!)."

Sentence: "Nurses wear certified personal protective equipment to care for patients with Coronavirus." in VerbAtlas 1.0

See also:

For those working at the intersection between #NLP and #COVID19 here is a possibly interesting venue:

NLP COVID-19 Workshop at ACL 2020 @aclmeeting https://t.co/hLFqljTVJC pic.twitter.com/0bOdcU0qBV

— Emilio Ferrara (@emilio__ferrara) April 4, 2020

Reflections [III] Conceptual Spaces At Work: Lund University, Sweden, May 2012

The next session - Event Structure, Conceptual Spaces and the Semantics of Verbs Peter Gärdenfors, Massimo Warglien & Matthijs Westera, was presented by Peter Gärdenfors whose book Conceptual Spaces (2000) I bought in Sedbergh in 2008 (talk about serendipity!).

Just reading the title back then was enough...

There was a starter question: Why are there word classes (nouns, verbs, adverbs...)?

It became evident that first morning that conceptual space as a representational form is very much still in active development. The abstract begins:

The aim of this paper is to integrate spatial cognition with lexical semantics. We develop cognitive models of actions and events based on conceptual spaces and vectors on them. The models are then used to present a semantic theory of verbs. ...

Word classes were related (as in the book) to conceptual space elements, e.g.

nouns - categories
adjectives - properties

In addressing verbs this paper seeks a two-vector model of an event.

Agent (Force) -> Patient (Result)   - result verbs describe the change in the object.

The Professor said that cognitive and linguistic aspects are often confused. Two theses were also provided - thesis 2: single domain constraint.

In addition talk of several spaces might be troubling (adding complexity), but to me the prospects of action space, category space and physical space makes the potential for modelling more explicit and accessible to Hodges' model.

I scribbled down 'agent not always necessary (fall, growth?...)' and pondered about self-care, relatives acting as a proxy.

There was a lot of detail on language, linguistics: static verbs, intentional verbs -> physical action; telic verbs. ... Rather than feel (completely!) confused by this specialist talk, I felt enthused for what is to follow.