A virtual salesperson

This chapter describes the use of a graphical humane interface - a Virtual Salesperson. The face of the Virtual Salesperson is a generic Facial Animation Engine developed at the University of Genova in Italy and uses a 3-D computer graphics model based on the MPEG-4 standard supplemented by Cyberware scans for facial detail. The appearance of the head may be modified by Facial Definition Parameters to more accurately model the required visage allowing one model to represent many different Talking Heads. The “brain” of the Virtual Salesperson, developed at Curtin University, integrates natural language parsing, text to speech synthesis, and artificial intelligence systems to produce a “bot” capable of helping a user through a question/answer sales enquiry. The Virtual Salesperson is a specific example of a generic Human Computer Interface - a Talking Head.

Scaling connectionist compositional representations

The Recursive Auto-Associative Memory (RAAM) has come to dominate connectionist investigations into representing compositional structure. Although an adequate model when dealing with limited data, the capacity of RAAM to scale-up to real-world tasks has been frequently questioned. RAAM networks are difficult to train (due to the moving target effect) and as such training times can be lengthy. Investigations into RAAM have produced many variants in an attempt to overcome such limitations. We outline how one such model ((S)RAAM) is able to quickly produce context-sensitive representations that may be used to aid a deterministic parsing process. By substituting a symbolic stack in an existing hybrid parser, we show that (S)RAAM is more than capable of encoding the real-world data sets employed. We conclude by suggesting that models such as (S)RAAM offer valuable insights into the features of connectionist compositional representations.

Information retrieval in documents using semantic criteria

The central problem of automatic retrieval from unformatted text is that computational devices are not adequately trained to look for associated information. However for complete understanding and information retrieval, a complete artificial intelligence would have to be built. This paper describes a method for achieving significant information retrieval by using a semantic search engine. The underlying semantic information is stored in a network of clarified words, linked by logical connections. We employ simple scoring techniques on collections of paths in this network to establish a degree of relevance between a document and a clarified search criterion. This technique has been applied with success to test examples and can be easily scaled up to search large documents.

The effects of controlled language processing on listening comprehension and recall

This study seeks to determine the possible interactions between listening proficiency and the state of strategic self-awareness; second, and more importantly, to investigate the effects of learned strategies on listening comprehension and recall; and finally to describe the most common real-time listening comprehension problems faced by EFL learners and to compare the differences between learners with different listening abilities. After ten training sessions, an assessment was made to see whether or not well-learned strategies could provide students with ample opportunity to practice the comprehension and recall processes. The analyses of the data revealed the causes of ineffective low-level processing and provided insights to solve the problems of parsing. Moreover, the study reveals that explicit instruction of cognitive and metacognitive strategies is needed if a syllabus wishes to help learners improve their listening comprehension and become more-proficient at directing their own learning and development as L2 listeners.

Object oriented regular expressions

Regular expressions are used to parse textual data to match patterns and extract variables. They have been implemented in a vast number of programming languages with a significant quantity of research devoted to improving their operational efficiency. However, regular expressions are limited to finding linear matches. Little research has been done in the field of object-oriented results which would allow textual or binary data to be converted to multi-layered objects. This is significantly relevant as many of todaypsilas data formats are object-based. This paper extends our previous work by detailing an algorithmic approach to perform object-oriented parsing, and provides an initial study of benchmarks of the algorithms of our contribution