• Contenu principal
  • Menu
OpenEdition Books
  • Accueil
  • Catalogue de 15479 livres
  • Éditeurs
  • Auteurs
  • Facebook
  • X
  • Partager
    • Facebook

    • X

    • Accueil
    • Catalogue de 15479 livres
    • Éditeurs
    • Auteurs
  • Ressources numériques en sciences humaines et sociales

    • OpenEdition
  • Nos plateformes

    • OpenEdition Books
    • OpenEdition Journals
    • Hypothèses
    • Calenda
  • Bibliothèques

    • OpenEdition Freemium
  • Suivez-nous

  • Lettre d’information
OpenEdition Search

Redirection vers OpenEdition Search.

À quel endroit ?
  • Accademia University Press
  • ›
  • Collana dell'Associazione Italiana di Li...
  • ›
  • Proceedings of the Second Italian Confer...
  • ›
  • Learning Grasping Possibilities for Arti...
  • Accademia University Press
  • Accademia University Press
    Accademia University Press
    Informations sur la couverture
    Table des matières
    Liens vers le livre
    Informations sur la couverture
    Table des matières
    Formats de lecture

    Plan

    Plan détaillé Texte intégral 1. Introduction 2. Manual Annotation of Grasping Possibilities 3. Semantic and physical knowledge about artifacts: guessing grasping possibilities 4. Conclusions and Future Works Bibliographie Auteurs

    Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

    Ce livre est recensé par

    Précédent Suivant
    Table des matières

    Learning Grasping Possibilities for Artifacts: Dimensions, Weights and Distributional Semantics

    Irene Russo et Irene De Felice

    p. 241-245

    Résumés

    In this paper we want to test how grasping possibilities for concrete objects can be automatically classified. To discriminate between objects that can be manipulated with one hand and the ones that require two hands, we combine conceptual knowledge about the situational properties of the objects, which can be modeled with distributional semantic methodologies, and physical properties of the objects (i.e. their dimensions and their weights), which can be found on the web through crawling.

    In questo articolo vogliamo testare come le possibilità di manipolazione degli oggetti concreti possano essere classificate automaticamente. Per distinguere tra oggetti che possono essere manipolati con una mano e oggetti che richiedono due mani, combiniamo conoscenza concettuale sulle proprietà situazionali dell’oggetto - rappresentandola secondo il paradigma della semantica disribuzionale - con le proprietà fisiche degli oggetti (le loro dimensioni e il loro peso) estratte dal web mediante crawling.

    Texte intégral 1. Introduction 2. Manual Annotation of Grasping Possibilities 3. Semantic and physical knowledge about artifacts: guessing grasping possibilities 3.1 Classification Experiment 4. Conclusions and Future Works Bibliographie Auteurs

    Texte intégral

    1. Introduction

    1Distributional semantic models of word meanings are based on representations that want to be cognitively plausible and that, as a matter of fact, have been tested to produce results correlated with human judgments when concepts similarity and automatic conceptual categorizations are the aim of the experiment (Erk, 2012; Turney and Pantel, 2010).

    2These approaches share the idea that two nominal concepts are similar and can be clustered in the same group if the corresponding lexemes occur in comparable linguistic contexts.

    3Their success is also due to the expectations of the Natural Language Processing (henceforth NLP) community: both for count and predictive models of distributional semantics (Baroni et al. 2014), the core idea is that encyclopedic knowledge packed in a big corpus can improve the performance in tasks such as word sense disambiguation.

    4However, purely textual representations turn out to be incomplete because in language learning and processing human beings are exposed to perceptual stimuli paired with linguistic ones: the old AI dream to ground language in the world requires the mapping between these two sources of knowledge. One of the aim of this paper is to understand how much physical knowledge can be retrieved in language. Can distributional representations of concrete nouns be helpful for the automatic classification of objects, when grasping possibilities are the focus? Could they help to discriminate between objects that can be manipulated with one hand and the ones that require two hands? More generally, how much knowledge about the physical world can be found in language?

    5Inspired by the cognitive psychology literature on the topic, in this paper artifactual categories are theorized as situated conceptualization where physical and situational properties meet (Barsalou 2002). These situational properties describe a physical setting or event in which the target object occurs (as grocery store, fruit basket, slicing, picnic for apple). In an action-based categorization of objects, these kinds of properties function as a complex relational system, which links the physical structure of the object, its use, the background settings, and the design history (Chaigneau et al. 2004). Situational properties can be derived from distributional semantic models, where each co-occurrence vector approximates the encyclopedic knowledge about its referent.

    6A complementary, but more action-oriented idea, is the psychological notion of affordance as the possibilities for actions that every environmental object offers (Gibson 1979). Conceptual information concerning objects affordances can be partially acquired through language, considering verb-direct object pairs as the linguistic realizations of the relations between the actions that can be performed by an agent, and the objects involved in those actions. Affordance verbs, intended as verbs that select a distinctive action for a specific object, can be discovered through statistical measures in corpora (Russo et al. 2013).

    7The main assumption of this paper is that the primary affordance for grasping of an artifact largely depends on its physical properties, in particular dimensions and weight. Such features are found in e-commerce websites. Extracting these values for many similar items, for example for all instances of “plate”, may help to automatically represent average dimensions for that object. However, combining this knowledge with situational properties of objects modeled as distributional semantics vectors can help understanding if they can be combined. This issue is relevant for the implementation of a module that automatically classifies grasping possibilities for objects in embodied robotics.

    8The paper is structured as follow: section 2 reports on the manual annotation of grasping possibilities for a set of 143 artifacts, discussing the definition of the gold standard that will be the dataset for classification experiments in section 3. Section 4 presents conclusions and ideas for future work.

    2. Manual Annotation of Grasping Possibilities

    9Concerning grasping possibilities for concrete objects, we expect as relevant several features. First of all, objects dimensions strongly influence the type of grasp afforded by objects. For instance, we are likely to grasp a tennis ball with a whole hand, but a soccer ball with two hands: the difference between the two spheres clearly is in their diameter.

    10Heavy objects require a type of grasp different from the one required by the light ones. Apart from these features, we should also consider more subjective factors, such as culture, past experience with objects, or intentions. This is particularly evident for artifacts and tools, that are the kind of objects most typically involved in manipulation and grasping and that often have a part that is specifically designed (or more suited than others) for grasping, for its shape and conformation, such as a handle (which we may call affording parts; cf. De Felice, 2015; in press). However, such parts (e.g. the handle of a cup) are usually grasped when the agents intention is to use the object for its canonical function (e.g. to drink from the cup), whereas in other cases it may be ignored and a different grasp could be performed (e.g. the whole cup might be taken from the above if we simply wanted to displace it).

    11Therefore, we can individuate at least four different grasp types afforded by concrete entities (cf. infra): the undifferentiated one-handed or twohanded grasps; a grasp by part, i.e. directed to a specific part of the object; a grasp with instrument, for substances, aggregates or every sort of things usually manipulated with some other object. In order to obtain a gold standard annotation of artifacts grasping possibilities, we first searched WordNet 3.0 for all the nouns that have artifact as hyperonym, obtaining a list of 1510 synsets. From this list, we chose the nouns that have enough pictures as products sold on amazon.com, since it was our intention to extract objects dimensions from this website for classification experiments (cf. 3). We selected the nouns for which at least 15 pages about that object sold on amazon.com were homogeneous - i.e. they contain objects of the same type- reducing noise caused by the crawling strategy. We obtained a total number of 143 nouns. Then, for each of these nouns, we manually annotated the type of grasp afforded by the object, according to the following classes:

    • One-handed grasp: this kind of grasp is forobjects that have no handles or protruding parts suited for the grasp, and that can be grasped by using only one hand. The size of two of the objects dimensions (length, width or thickness) usually does not exceed the maximum span of a hand with at least two fingers bent in order to grasp and hold something. E.g.: bowl, bottle, candle, shell, necklace, clothes peg.

    • Two-handed grasp: this kind of grasp is forobjects that have no handles or protruding parts suited for the grasp, and that are usually grasped with two hands, because their size exceeds the maximum span of a single hand. E.g.: board, soccer ball, player piano, table, computer.

    • Grasp by part: this kind of grasp is for: (i)small or large objects that have a part specifically designed for the grasping; (ii) entities that have a well identifiable part that, even if it is not specifically designed for this specific purpose, is more suited than others for the grasping thanks to its shape and conformation. E.g. knife, jug, axe, trolley, bag.

    • Grasp with instrument: this kind of grasp ismainly for substances, aggregates, and entities which cannot be (or are usually not) controlled without using some other object (an instrument, generally a container). E.g. water, broth, flour, bran, sand.

    Table 1: Number of items per classes in the gold standard.

    class

    #nouns

    onehand

    43

    onehandORpart

    1

    oneORtwohand

    25

    part

    23

    twohand

    73

    twohandORpart

    3

    12For several objects more than one grasping possibility is plausible, depending on the size (a plate can be small or big) or on the availability of a container (sand can be grasped by hand).

    13The dataset of 143 nouns have been annotated by two annotators and the inter-annotator agreement was 0.66. Since we need a gold standard for experiments, we managed disagreements reaching a consensus on every noun. The gold standard contains items assigned to 6 classes, distributed as in Table 1.

    3. Semantic and physical knowledge about artifacts: guessing grasping possibilities

    14The way humans can grasp an object can be designed as a function that depends on multiple variables, such as the presence of affording parts (i.e. handle for bag), its shape, its dimensions, its weight and the final aim of the action of grasping, modeled here as part of the situational properties. In this paper we want to test which one of these features can help in classifying artifacts that have been manually annotated according to 6 categories (see par. 2). In particular we experiment with a combination of 4 features provided for each noun:

    • distributional semantics information fromtwo corpora (GoogleNews and instructables.com) obtained with word2vec toolkit (Mikolov et al. 2013);

    • averagedepth) for each object, obtained crawling atdimensions (height, length and least 15 pages per object from amazon.com;

    • average weight for each object, obtainedcrawling at least 15 pages per object from amazon.com;

    • co-occurrence matrix in the corpus instructa-bles.com with nouns that are affording parts, extracting the syntactic pattern AFFORDING PART NOUN of ARTIFACT (e.g. ”handle of the bag”).

    15Because all the big corpora available contain in general news or web crawled texts that don’t mention concrete actions and concrete objects so often, we choose to build a smaller but coherent corpus of do-it-yourself instructions, with the assumption that it will contain frequent instances of concrete language.

    16We crawled from the website instructables.com all the titles and descriptions for the projects available online in six categories (e.g. technologies, workshop, living, food, play, outside). Cleaned of the html code, the instructables.com corpus has 17M tokens; each project was parsed with the Stanford parser (de Marneffe and Manning 2008). To test if a do-it-yourself instructions corpus is useful with respect to a generic one, we represent each noun in the following experiment as a vector extracted from GoogleNews with word2vec toolkit (Mikolov et al. 2013) but also as a vector extracted from the instructables.com corpus trained with the same toolkit. These are the purely textual representations we experimented with; to complete this knowledge we added extracted information about dimensions, weight and affording parts for 143 objects.

    17The list of objects’ parts that afford grasping and are component of the pattern extracted for the feature “affording parts” has been derived with a psycholinguistic test (De Felice 2015). Thirty students of the University of Pisa were interviewed and presented with 42 images of graspable entities. For each picture, they were asked to describe in the most detailed way how they would have grasped the object represented. Among the objects depicted, there were 31 artefacts. From the interviews recorded for these artefacts, we extracted all nouns denoting objects’ parts that were named as possible target of the grasp (e.g. the handle for the bag, the cup or the ladle). The list of 78 nouns was then translated in English.

    Table 2: Precision and recall for 8 combinations of features for the 6 classes dataset.

    features

    Precision

    Recall

    instructables.com

    0.113

    0.336

    GoogleNews

    0.113

    0.336

    weight

    0.364

    0.406

    dimensions

    0.413

    0.517

    dimensions+weight

    0.561

    0.531

    affording parts

    0.25

    0.399

    instructables.com + all

    0.443

    0.552

    GoogleNews + all

    0.458

    0.559

    3.1 Classification Experiment

    18The experiment is based on a multi-label classification, since our dataset consists of 143 nouns denoting artifacts, annotated according to 6 categories. The implementation of Support Vector Multi-Classification is based on LibSVM software (Chang and Lin 2001) in WEKA with 10 fold cross-validation. Table 2 reports the results in terms of precision and recall. The best perfomance depends on information about average dimensions and weight of the objects. Distributional semantics vectors seems useless.

    19The overall performance is influenced by the fact that some classes are small in the gold standard. For this reason, we experimented with the same features including just the 91 nouns that belong to the “onehand” or “twohand” classes. In Table 3, results show again that dimensions and dimensions plus weight produce good results (with “dimensions” as the best feature), even if they do not improve the performance when combined with distributional vectors that in this case are useful per se. Again, affording parts co-occurences produce the worst performance, mainly because the list of affording parts was originally derived for only 31 artefacts, and not for all the objects considered in our experiment.

    Table 3: Precision and recall for 8 combinations of features on two-classes dataset (“onehand” VS “twohand”).

    features

    Precision

    Recall

    GoogleNews

    0.846

    0.846

    weight

    0.715

    0.714

    dimensions

    0.851

    0.846

    dimensions+weight

    0.831

    0.802

    affording parts

    0.63

    0.615

    GoogleNews + all

    0.846

    0.846

    4. Conclusions and Future Works

    20In this paper we test how distributional representations of nouns denoting artifacts can be combined with physical information about their dimensions and weights automatically extracted from an ecommerce website and with co-occurrence information about their affording parts as found in a corpus of do-it-yourself instructions. The starting hypothesis - concerning grasping possibilities as basic manipulative actions for object - was that they are conceptually a combination of situational and physical properties.

    21As a consequence, we expect the best performance from a mixed features models. This hypothesis is not confirmed; for the two-classes dataset (“onehand” VS “twohand”) both physical knowledge and distributional semantics vectors give good resultsbut they don’t improve the classifier’s performance when combined.

    22These results are in line with the current trend to mix textual and visual features from computer vision algorithms (Bruni et al. 2012) in order to go beyond the limitations of purely textual semantic representations that cannot encode information about colors, dimensions, shapes etc. As future work we plan to integrate the features used for the experiment in this paper with representations of words as bag of visual words derived from the scale-invariant feature transform (SIFT) algorithm (Lowe 1999) that in computer vision helps to detect and describe local features in images.

    Bibliographie

    Barsalou, L.W. 2002. Being there conceptually: simulating categories in preparation for situated action. Representation, Memory, and Development: Essays in Honor of Jean Mandler,1–15.

    Baroni, M., Dinu, G. and Kruszewski, G. 2014. Don’t count, predict! A systematic comparison of contextcounting vs. context-predicting semantic vectors. Proceedings of ACL 2014 (52nd Annual Meeting of the Association for Computational Linguistics), East Stroudsburg PA: ACL, 238-247.

    Ashok K. Chandra, Dexter C. Kozen, and Larry J. Stockmeyer. 2012. Distributional semantics with eyes: Using image analysis to improve computational representations of word meaning. Proceedings of ACM Multimedia, 1219-1228.

    Chaigneau, S.E., Barsalou, L.W., and Sloman, S. 2004. Assessing the causal structure of function. Journal of Experimental Psychology: General, 133: 601625.

    Chih-Chung Chang and Chih-Jen Lin. 2011. LIBSVM: A library for support vector machines. ACM Transactions on Intelligent Systems and Technology, 2:27:1-27:27. Software available at http://www.csie.ntu.edu.tw/cjlin/libsvm.

    Katrin Erk. 2012. Vector Space Models of Word Meaning and Phrase Meaning: A Survey. Language and Linguistics Compass, 6(10):635-653.

    De Felice, I. in press. Objects’ parts afford action: evidence from an action description task. In V. Torrens (ed.), Language Processing and Disorders. Newcastle: Cambridge Scholars Publishing.

    De Felice, I. 2015. Language and Affordances. PhD thesis, University of Pisa, Italy.

    Gibson, J. J. 1979. The Ecological Approach to Visual Perception. Boston: Houghton Mifflin.

    Lowe, D.G. 1999. Object recognition from local scale-invariant features. International Conference on Computer Vision, pp. 1150-1157.

    Marie-Catherine de Marneffe and Christopher D. Manning. 2008. The Stanford typed dependencies representation. In COLING Workshop on Crossframework and Cross-domain Parser Evaluation.

    Russo, I., De Felice, I., Frontini, F., Khan, F., and Monachini, M. 2013. (Fore)seeing actions in objects. Acquiring distinctive affordances from language. In B. Sharp, and M. Zock (eds.), Proceedings of The 10th International Workshop on Natural Language Processing and Cognitive Science - NLPCS 2013 (Marseille, France, 15-17/10/2013), 151-161.

    Peter D. Turney and Patrick Pantel. 2010. From Frequency to Meaning: vector space models of semantics. J. Artif. Int. Res., 37(1):141-188, January.

    Auteurs

    • Irene Russo

      Istituto di linguistica Computazionale “A . Zampolli” CNR - irene.russo@ilc.cnr.it

    • Irene De Felice

      University of Pisa - irene.defelice@ilc.cnr.it

    Précédent Suivant
    Table des matières

    Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0

    Le texte seul est utilisable sous licence Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

    Voir plus de livres
    Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

    Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

    3-4 December 2015, Trento

    Cristina Bosco, Sara Tonelli et Fabio Massimo Zanzotto (dir.)

    2015

    Proceedings of the Third Italian Conference on Computational Linguistics CLiC-it 2016

    Proceedings of the Third Italian Conference on Computational Linguistics CLiC-it 2016

    5-6 December 2016, Napoli

    Anna Corazza, Simonetta Montemagni et Giovanni Semeraro (dir.)

    2016

    EVALITA. Evaluation of NLP and Speech Tools for Italian

    EVALITA. Evaluation of NLP and Speech Tools for Italian

    Proceedings of the Final Workshop 7 December 2016, Naples

    Pierpaolo Basile, Franco Cutugno, Malvina Nissim et al. (dir.)

    2016

    Proceedings of the Fourth Italian Conference on Computational Linguistics CLiC-it 2017

    Proceedings of the Fourth Italian Conference on Computational Linguistics CLiC-it 2017

    11-12 December 2017, Rome

    Roberto Basili, Malvina Nissim et Giorgio Satta (dir.)

    2017

    Proceedings of the Fifth Italian Conference on Computational Linguistics CLiC-it 2018

    Proceedings of the Fifth Italian Conference on Computational Linguistics CLiC-it 2018

    10-12 December 2018, Torino

    Elena Cabrio, Alessandro Mazzei et Fabio Tamburini (dir.)

    2018

    EVALITA Evaluation of NLP and Speech Tools for Italian

    EVALITA Evaluation of NLP and Speech Tools for Italian

    Proceedings of the Final Workshop 12-13 December 2018, Naples

    Tommaso Caselli, Nicole Novielli, Viviana Patti et al. (dir.)

    2018

    EVALITA Evaluation of NLP and Speech Tools for Italian - December 17th, 2020

    EVALITA Evaluation of NLP and Speech Tools for Italian - December 17th, 2020

    Proceedings of the Seventh Evaluation Campaign of Natural Language Processing and Speech Tools for Italian Final Workshop

    Valerio Basile, Danilo Croce, Maria Maro et al. (dir.)

    2020

    Proceedings of the Seventh Italian Conference on Computational Linguistics CLiC-it 2020

    Proceedings of the Seventh Italian Conference on Computational Linguistics CLiC-it 2020

    Bologna, Italy, March 1-3, 2021

    Felice Dell'Orletta, Johanna Monti et Fabio Tamburini (dir.)

    2020

    Proceedings of the Eighth Italian Conference on Computational Linguistics CliC-it 2021

    Proceedings of the Eighth Italian Conference on Computational Linguistics CliC-it 2021

    Milan, Italy, 26-28 January, 2022

    Elisabetta Fersini, Marco Passarotti et Viviana Patti (dir.)

    2022

    Voir plus de livres
    1 / 9
    Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

    Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

    3-4 December 2015, Trento

    Cristina Bosco, Sara Tonelli et Fabio Massimo Zanzotto (dir.)

    2015

    Proceedings of the Third Italian Conference on Computational Linguistics CLiC-it 2016

    Proceedings of the Third Italian Conference on Computational Linguistics CLiC-it 2016

    5-6 December 2016, Napoli

    Anna Corazza, Simonetta Montemagni et Giovanni Semeraro (dir.)

    2016

    EVALITA. Evaluation of NLP and Speech Tools for Italian

    EVALITA. Evaluation of NLP and Speech Tools for Italian

    Proceedings of the Final Workshop 7 December 2016, Naples

    Pierpaolo Basile, Franco Cutugno, Malvina Nissim et al. (dir.)

    2016

    Proceedings of the Fourth Italian Conference on Computational Linguistics CLiC-it 2017

    Proceedings of the Fourth Italian Conference on Computational Linguistics CLiC-it 2017

    11-12 December 2017, Rome

    Roberto Basili, Malvina Nissim et Giorgio Satta (dir.)

    2017

    Proceedings of the Fifth Italian Conference on Computational Linguistics CLiC-it 2018

    Proceedings of the Fifth Italian Conference on Computational Linguistics CLiC-it 2018

    10-12 December 2018, Torino

    Elena Cabrio, Alessandro Mazzei et Fabio Tamburini (dir.)

    2018

    EVALITA Evaluation of NLP and Speech Tools for Italian

    EVALITA Evaluation of NLP and Speech Tools for Italian

    Proceedings of the Final Workshop 12-13 December 2018, Naples

    Tommaso Caselli, Nicole Novielli, Viviana Patti et al. (dir.)

    2018

    EVALITA Evaluation of NLP and Speech Tools for Italian - December 17th, 2020

    EVALITA Evaluation of NLP and Speech Tools for Italian - December 17th, 2020

    Proceedings of the Seventh Evaluation Campaign of Natural Language Processing and Speech Tools for Italian Final Workshop

    Valerio Basile, Danilo Croce, Maria Maro et al. (dir.)

    2020

    Proceedings of the Seventh Italian Conference on Computational Linguistics CLiC-it 2020

    Proceedings of the Seventh Italian Conference on Computational Linguistics CLiC-it 2020

    Bologna, Italy, March 1-3, 2021

    Felice Dell'Orletta, Johanna Monti et Fabio Tamburini (dir.)

    2020

    Proceedings of the Eighth Italian Conference on Computational Linguistics CliC-it 2021

    Proceedings of the Eighth Italian Conference on Computational Linguistics CliC-it 2021

    Milan, Italy, 26-28 January, 2022

    Elisabetta Fersini, Marco Passarotti et Viviana Patti (dir.)

    2022

    Voir plus de chapitres

    Sardinian on Facebook: Analysing Diatopic Varieties through Translated Lexical Lists

    Irene Russo, Simone Pisano et Claudia Soria

    Samskara

    Minimal structural features for detecting subjectivity and polarity in Italian tweets

    Irene Russo et Monica Monachini

    Italian in the Trenches: Linguistic Annotation and Analysis of Texts of the Great War

    Irene De Felice, Felice Dell’Orletta, Giulia Venturi et al.

    Visualising Italian Language Resources: a Snapshot

    Riccardo Del Gratta, Francesca Frontini, Monica Monachini et al.

    EnetCollect in Italy

    Lionel Nicolas, Verena Lyding, Luisa Bentivogli et al.

    HaSpeeDe 2 @ EVALITA2020: Overview of the EVALITA 2020 Hate Speech Detection Task

    Manuela Sanguinetti, Gloria Comandini, Elisa di Nuovo et al.

    Voir plus de chapitres
    1 / 6

    Sardinian on Facebook: Analysing Diatopic Varieties through Translated Lexical Lists

    Irene Russo, Simone Pisano et Claudia Soria

    Samskara

    Minimal structural features for detecting subjectivity and polarity in Italian tweets

    Irene Russo et Monica Monachini

    Italian in the Trenches: Linguistic Annotation and Analysis of Texts of the Great War

    Irene De Felice, Felice Dell’Orletta, Giulia Venturi et al.

    Visualising Italian Language Resources: a Snapshot

    Riccardo Del Gratta, Francesca Frontini, Monica Monachini et al.

    EnetCollect in Italy

    Lionel Nicolas, Verena Lyding, Luisa Bentivogli et al.

    HaSpeeDe 2 @ EVALITA2020: Overview of the EVALITA 2020 Hate Speech Detection Task

    Manuela Sanguinetti, Gloria Comandini, Elisa di Nuovo et al.

    Accès ouvert

    Accès ouvert

    ePub

    PDF

    PDF du chapitre

    Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

    X Facebook Email

    Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

    Vous pouvez vous connecter à votre bibliothèque à l’adresse suivante : https://freemium.openedition.org/oebooks

    Suggérer l’acquisition à votre bibliothèque

    Si vous avez des questions, vous pouvez nous écrire à access[at]openedition.org

    Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015

    Vérifiez si votre bibliothèque a déjà acquis ce livre : authentifiez-vous à OpenEdition Freemium for Books.

    Vous pouvez suggérer à votre bibliothèque d’acquérir un ou plusieurs livres publiés sur OpenEdition Books. N’hésitez pas à lui indiquer nos coordonnées : access[at]openedition.org

    Vous pouvez également nous indiquer, à l’aide du formulaire suivant, les coordonnées de votre bibliothèque afin que nous la contactions pour lui suggérer l’achat de ce livre. Les champs suivis de (*) sont obligatoires.

    Veuillez, s’il vous plaît, remplir tous les champs.

    La syntaxe de l’email est incorrecte.

    Référence numérique du chapitre

    Format

    Russo, I., & De Felice, I. (2015). Learning Grasping Possibilities for Artifacts: Dimensions, Weights and Distributional Semantics. In C. Bosco, S. Tonelli, & F. M. Zanzotto (éds.), Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015 (1‑). Accademia University Press. https://doi.org/10.4000/books.aaccademia.1546
    Russo, Irene, et Irene De Felice. « Learning Grasping Possibilities for Artifacts: Dimensions, Weights and Distributional Semantics ». In Proceedings of the Second Italian Conference on Computational Linguistics CLiC-It 2015, édité par Cristina Bosco, Sara Tonelli, et Fabio Massimo Zanzotto. Torino: Accademia University Press, 2015. https://doi.org/10.4000/books.aaccademia.1546.
    Russo, Irene, et Irene De Felice. « Learning Grasping Possibilities for Artifacts: Dimensions, Weights and Distributional Semantics ». Proceedings of the Second Italian Conference on Computational Linguistics CLiC-It 2015, édité par Cristina Bosco et al., Accademia University Press, 2015, https://doi.org/10.4000/books.aaccademia.1546.

    Référence numérique du livre

    Format

    Bosco, C., Tonelli, S., & Zanzotto, F. M. (éds.). (2015). Proceedings of the Second Italian Conference on Computational Linguistics CLiC-it 2015 (1‑). Accademia University Press. https://doi.org/10.4000/books.aaccademia.1277
    Bosco, Cristina, Sara Tonelli, et Fabio Massimo Zanzotto, éd. Proceedings of the Second Italian Conference on Computational Linguistics CLiC-It 2015. Torino: Accademia University Press, 2015. https://doi.org/10.4000/books.aaccademia.1277.
    Bosco, Cristina, et al., éditeurs. Proceedings of the Second Italian Conference on Computational Linguistics CLiC-It 2015. Accademia University Press, 2015, https://doi.org/10.4000/books.aaccademia.1277.
    Compatible avec Zotero Zotero

    1 / 3

    Accademia University Press

    Accademia University Press

    • Plan du site
    • Se connecter

    Suivez-nous

    • Facebook
    • Flux RSS

    URL : http://www.aaccademia.it/

    Email : info@aaccademia.it

    Adresse :

    Accademia University Press

    Via Carlo Alberto 55

    I‐10123

    Torino

    Italia

    OpenEdition
    • Candidater à OpenEdition Books
    • Connaître le programme OpenEdition Freemium
    • Commander des livres
    • S’abonner à la lettre d’OpenEdition
    • CGU d’OpenEdition Books
    • Accessibilité : partiellement conforme
    • Données personnelles
    • Gestion des cookies
    • Système de signalement