Quantum-inspired meaning modeling of collections of documents

    Activiteit: Talk or presentation at a conference

    Description

    Our century is witnessing unprecedented efforts to understand the underlying struc-
    ture of how meaning is built and processed by human minds. For some, the ultimate

    goal of such efforts is artificial intelligence.
    In that respect, it seems that we are still in a sort of “alchemical period” for
    the study of meaning representation and processing: several intuitions about how
    meaning can be structured have combined with our capacity to store and process
    vast amounts of data, producing a myriad of approaches to handle certain particular
    tasks. However, no one appears to have a clear idea of how meaning is formed at
    a more fundamental level. Therefore, similarly to how the alchemical period has

    been superseded by modern scientific theories, our efforts in what concerns mean-
    ing representation should also be that of developing well-principled theories of how

    meaning is represented and processed, beyond the good performance of individual
    specific tasks. Following some of the recent developments in quantum cognition,
    precisely underlying such attempt to find a well-principled theory of how concepts
    are formed and decisions are made by human minds, we present a formalism that
    promises to be able to handle at a fundamental level the complexities of meaning
    representation for a collection of documents (such as those forming the World Wide
    Web). Namely, we associate a quantum-like ‘entity of meaning’ to a ‘language
    entity formed by printed documents’. A language entity is hence conceived as a
    collection of traces left by the entity of meaning, resulting from a “measurement
    interaction” between a “user” (which does not have to be necessarily a human) and
    the collection of documents. Therefore, a collection of documents, like the Web, is
    described as the space of manifestation of a more complex entity – the QWeb – which
    is the object of our modeling. More precisely, we explain how to build a QWeb that
    accounts for the correlations between words appearing in printed documents, e.g.,
    co-occurrences, and we show that both ‘quantum context and interference effects’
    are required to explain the probabilities calculated by counting the relative number
    of documents containing certain words and their co-ocurrrences. We also discuss
    the potential applications of our formalism to other types of media, such as images
    and videos.
    Periode10 okt 2018
    EvenementstitelQuantum Cagliari 2018
    EvenementstypeConference
    LocatieCagliary, Italy
    Mate van erkenningInternational