Grounding human vocabulary in robot perception through interaction

This thesis addresses the problem of word learning in computational agents. The motivation behind this work lies in the need to support language-based communication between service robots and their human users, as well as grounded reasoning using symbols relevant for the assigned tasks. The research focuses on the problem of grounding human vocabulary in robotic agent’s sensori-motor perception. Words have to be grounded in bodily experiences, which emphasizes the role of appropriate embodiments. On the other hand, language is a cultural product created and acquired through social interactions. This emphasizes the role of society as a source of linguistic input. Taking these aspects into account, an experimental scenario is set up where a human instructor teaches a robotic agent the names of the objects present in a visually shared environment. The agent grounds the names of these objects in visual perception. Word learning is an open-ended problem. Therefore, the learning architecture of the agent will have to be able to acquire words and categories in an openended manner. In this work, four learning architectures were designed that can be used by robotic agents for long-term and open-ended word and category acquisition. The learning methods used in these architectures are designed for incrementally scaling-up to larger sets of words and categories. A novel experimental evaluation methodology, that takes into account the openended nature of word learning, is proposed and applied. This methodology is based on the realization that a robot’s vocabulary will be limited by its discriminatory capacity which, in turn, depends on its sensors and perceptual capabilities. An extensive set of systematic experiments, in multiple experimental settings, was carried out to thoroughly evaluate the described learning approaches. The results indicate that all approaches were able to incrementally acquire new words and categories. Although some of the approaches could not scale-up to larger vocabularies, one approach was shown to learn up to 293 categories, with potential for learning many more.

Mining biomedical information from scientific literature

The rapid evolution and proliferation of a world-wide computerized network, the Internet, resulted in an overwhelming and constantly growing amount of publicly available data and information, a fact that was also verified in biomedicine. However, the lack of structure of textual data inhibits its direct processing by computational solutions. Information extraction is the task of text mining that intends to automatically collect information from unstructured text data sources. The goal of the work described in this thesis was to build innovative solutions for biomedical information extraction from scientific literature, through the development of simple software artifacts for developers and biocurators, delivering more accurate, usable and faster results. We started by tackling named entity recognition - a crucial initial task - with the development of Gimli, a machine-learning-based solution that follows an incremental approach to optimize extracted linguistic characteristics for each concept type. Afterwards, Totum was built to harmonize concept names provided by heterogeneous systems, delivering a robust solution with improved performance results. Such approach takes advantage of heterogenous corpora to deliver cross-corpus harmonization that is not constrained to specific characteristics. Since previous solutions do not provide links to knowledge bases, Neji was built to streamline the development of complex and custom solutions for biomedical concept name recognition and normalization. This was achieved through a modular and flexible framework focused on speed and performance, integrating a large amount of processing modules optimized for the biomedical domain. To offer on-demand heterogenous biomedical concept identification, we developed BeCAS, a web application, service and widget. We also tackled relation mining by developing TrigNER, a machine-learning-based solution for biomedical event trigger recognition, which applies an automatic algorithm to obtain the best linguistic features and model parameters for each event type. Finally, in order to assist biocurators, Egas was developed to support rapid, interactive and real-time collaborative curation of biomedical documents, through manual and automatic in-line annotation of concepts and relations. Overall, the research work presented in this thesis contributed to a more accurate update of current biomedical knowledge bases, towards improved hypothesis generation and knowledge discovery.

PRESENTE, um projeto de futuro sobre o passado

No Monumento Nacional aos Combatentes do Ultramar, em Belém, encontram-se dispostos por ano e ordem alfabética os nomes dos militares mortos nesse conflito que durou treze anos. Este enunciado é o ponto de partida para um projeto artístico que, não sendo construído fisicamente a partir de fontes documentais ou de artefactos relacionados com os factos históricos, se irá desenvolver com base em premissas conceptuais no sentido de despoletar a partilha dessa memória. Este projeto artístico é, em si, a criação de um novo documento que olha o passado e o procura projetar no futuro com base no momento “PRESENTE”. Nesta comunicação propomo-nos, metodologicamente, discutir o processo de construção de um projeto artístico que, com a atribuição do prémio Bolsa Estação Imagem | Mora 2014 dará origem a uma exposição pública e à publicação de um livro relacionando-o com um conjunto de possibilidades que questionam as potencialidades que a área da criação artística dispõe para contaminar as questões da musealização de forma a contribuir com o despontar de novas abordagens e narrativas nas práticas da materialização de exposições como médium e lugar de criação artística. Através da consideração processual deste projeto procuramos atingir o significado da memória nos processos de mediação artística onde as imagens renunciando à possibilidade de serem simulacro ou fantasmagoria, simbolizam cada coisa e o seu contrário, abeirando-se da não representação e, neste limite, qual o papel do museu nessas práticas de mediação.

Vaness: DNS for nomadic users in vehicular networks

Vehicular networks, also known as VANETs, are an ad-hoc network formed by vehicles and road-side units. Nowadays they have been attracting big interest both from researchers as from the automotive industry. With the upcoming of automotive specific operating systems and self-driving cars, the use of applications on vehicles and the integration with common mobile devices is becoming a big part of VANETs. Although many advances have been made on this field, there is still a big discrepancy between the communication layer services provided by VANETs and the user level services, namely those accessible through mobile applications on other networks and technologies. Users and developers are accustomed to user-to-user or user-tobusiness communication without explicit concerns related with the available communication transport layer. Such is not possible in VANETs since people may use more than one vehicle. However, to send a message to a specific user in these networks, there is a need to know the ID of the vehicle where the user is, meaning that there is a lack of services that map each individual user to VANETs endpoint (vehicle identification). This dissertation work proposes VANESS, a naming service as a resource to support user-to-user communication within a heterogeneous scenario comprising typical ISP scenario and VANETs focused on mobile devices. The proposed system is able to map the user to an end point either locally (i.e. there is not internet connection at all), online (i.e. system is not in a vehicular network but has direct internet connection) and using a gateway (i.e. the system is in a vehicular network where some of the nodes have internet access and will act as a gateway). VANESS was fully implemented on android OS with results proving his viability, and partially on iOS showing its multiplatform capabilities.