Patrizia Grifoni

A survey of grammatical inference methods for natural language learning

The high complexity of natural language and the huge amount of human and temporal resources necessary for producing the grammars lead several researchers in the area of Natural Language Processing to investigate various solutions for automating grammar generation and updating processes. Many algorithms for Context-Free Grammar inference have been developed in the literature. This paper provides a survey of the methodologies for inferring context-free grammars from examples, developed by researchers in the last decade. After introducing some preliminary definitions and notations concerning learning and inductive inference, some of the most relevant existing grammatical inference methods for Natural Language are described and classified according to the kind of presentation (if text or informant) and the type of information (if supervised, unsupervised, or semi-supervised). Moreover, the state of the art of the strategies for evaluation and comparison of different grammar inference methods is presented. The goal of the paper is to provide a reader with introduction to major concepts and current approaches in Natural Language Learning research.

An Overview of Methods for Virtual Social Networks Analysis

The increasing achievement of the Web has led people to exploit collaborative technologies in order to encourage partnerships among different groups. The cooperation can be achieved by Virtual Social Networks that facilitate people’s social interaction and enable them to remain in touch with friends exploiting the pervasive nature of information devices and services. The interest in analysing Virtual Social Networks has grown massively in recent years, and it involves researches from different fields. This led to the development of different methods to study relationships between people, groups, organisations- and other knowledge-processing entities on the Web. This chapter classifies these methods in two categories. The first category concerns methods used for the network data collection while the second category deals with methods used for the network data visualisation. The chapter gives an example of application of these methods to analyse the Virtual Social Network LinkedIn.

A multimodal pervasive framework for ambient assisted living

The increasing need to access information everywhere and at any time leads us to believe that future user interfaces, through which users interact with pervasive computing systems, must address both device and modality independence. The pervasive computing paradigm sees almost every object in the everyday environment as a system able to communicate with users and other systems in their own languages. The interaction between users and systems is therefore typically multimodal. The main challenge of multimodal interaction, that is also the main topic of this paper, lies in developing a framework that is able to process information derived from whatever input modalities, giving these inputs an appropriate representation and integrating these individual representations into a joint semantic interpretation. A description of this multimodal pervasive framework will be given in this paper, along with some details of its application in Ambient Assisted Living and the usability test that was implemented to validate its effectiveness.

New forms of social and professional digital relationships: the case of Facebook

In recent years social media are becoming ubiquitous and important for social networking and for sharing content and knowledge; they are increasingly used for professional purposes beyond personal and home use. Social networks are spreading so much that they now represent one of the most relevant and interesting social phenomena. The study presented in this paper aims to analyse how Facebook is used by researchers, by means of a quantitative method, and the experiences and perceptions of changes and trends of professionals and researchers towards professional use, by means of a qualitative method. In particular, the study analysed: (i) how and why researchers of the National Research Council (CNR) in Italy in the Information and Communication Technology (ICT) and Cultural Identity (IC) Departments use social networks, and more specifically Facebook (ii) their attitude towards risks and privacy concerns (iii) professional use and perceptions by CNR researchers and professionals of changes over time in using Facebook for professional purposes.

Multimodal interaction systems: information and time features

Multimodal interaction systems combine visual information (involving images, text, sketches and so on) with voice, gestures and other modalities to provide flexible and powerful dialogue approaches, enabling users to choose one or more of the multiple interaction modalities. They break down the barriers in adopting mobile devices for value-added services and the use of integrated multiple input modes enables users to benefit from the natural approach used in human communication. This paper deals with the main features of multimodal interaction and systems, starting from the definition of visual language given in Bottoni et al. (1995) and extending it to multimodality. Modal/multimodal message, interpretation and materialisation functions and multimodal sentence are defined. This paper introduces and formally defines the different classes of cooperation between different modes, introducing the time relationships among the involved modalities and the relationships between chunks of information connected with these modalities.

Generating Multimodal Grammars for Multimodal Dialogue Processing

This paper presents a new multimodal grammar generation system (MGGS) that allows defining a multimodal grammar in a very easy and intuitive way, overcoming the difficulties arising from the textual description of grammar production rules. The novelty of the proposed approach relies in adopting a “by example” paradigm to define a multimodal grammar. This paradigm consists in providing concrete examples of multimodal sentences, and enabling a grammar inference algorithm to automatically generate the grammar rules to parse those sentences. The main contribution of this approach is that it is general enough to be applicable for whatever modalities and in whichever domains. Moreover, the use of the grammar inference algorithm for automating the grammar writing and updating processes, reduces costs of the grammar development and maintenance. In the first part of this paper, we present the MGGS, describing in detail the methodology we have implemented in this system. More specifically, the multimodal attribute grammar and the grammar inference algorithm are illustrated. The second part describes the experiment aimed at observing participants while interacting with the system in order to provide some real data about the usability of the system. Results of this experiment showed that the proposed system facilitates the grammar definition and updating, and it is more suitable also for nonexpert people as it does not require the learning of the grammar notation.

A Learning Algorithm for Multimodal Grammar Inference

The high costs of development and maintenance of multimodal grammars in integrating and understanding input in multimodal interfaces lead to the investigation of novel algorithmic solutions in automating grammar generation and in updating processes. Many algorithms for context-free grammar inference have been developed in the natural language processing literature. An extension of these algorithms toward the inference of multimodal grammars is necessary for multimodal input processing. In this paper, we propose a novel grammar inference mechanism that allows us to learn a multimodal grammar from its positive samples of multimodal sentences. The algorithm first generates the multimodal grammar that is able to parse the positive samples of sentences and, afterward, makes use of two learning operators and the minimum description length metrics in improving the grammar description and in avoiding the over-generalization problem. The experimental results highlight the acceptable performances of the algorithm proposed in this paper since it has a very high probability of parsing valid sentences.

InteSe: An Integrated Model for Resolving Ambiguities in Multimodal Sentences

The pervasiveness of ambiguity in communication processes suggests addressing the problem of semantic and syntactic ambiguities in multimodal interaction languages. This paper presents an integrated model based on layered, hierarchical, and hidden Markov models for dealing with the complex process of multimodal ambiguity resolution. The proposed model consists of different levels, from the terminals of a multimodal language (terminal elements) to the level of multimodal sentences. A software module implemented the model that has been evaluated in terms of accuracy and robustness. The experimental results show good levels of accuracy and robustness compared with other existing approaches.

Multidimensional Visualization System for Travel Social Networks

In this paper we provide a multidimensional visualization system for travel social network. This system allows to analyse the structure and the social aggregation of travel networks providing different perspectives according to three dimensions: the space that defines locations connected to the members; the time that defines temporal evolution about tourist interests of the members; and coordinates involving classes of interests that define typologies of tourism chosen by members. Furthermore the system offers location-based tourist services providing a spatial visualization of the services connected to the specific place and sharing information about specific place. Finally in order to evaluate the proposed method, interviews to tourists and stakeholders were carried out after testing of the system.

Approximating Geographical Queries

This article proposes a graph-theoretic methodology for query approximation in Geographic Information Systems, enabling the relaxation of three kinds of query constraints: topological, semantic and structural. An approximate query is associated with a value corresponding to the degree of similarity with the original query. Such a value is computed for topological constraints on the basis of the topological distance between configurations, for semantic constraints using the information content approach, and for structural constraints revisiting the maximum weighted matching problem in bipartite graphs. Finally, the high correlation of our proposal with human judgment is demonstrated by an experiment.