Radu Luchian(ov): Projects: CogSci
Research Interests

This document was part of my application to
the Carleton University PhD in Cognitive Science program

My academic interests relevant to this application to Carleton's Ph.D. program in Cognitive Science, as detailed below, are: learning and knowledge acquisition, language acquisition, object-oriented programming, database management systems, communication and IT, and gaming; I will address all of these areas below. As far as work is concerned, I have experimental laboratory experience and I have assisted professors in research and teaching both at the undergraduate and the graduate level.

I'm afraid that my ambitions are pretty tough to satisfy. Cognitive Science is just one step on the way I had set out for myself. I am interested in the application of my research and I am not the kind of person who'd spend a lifetime on 'teoria sine praxis'. I'm interested in engineering, physics, chemistry, communication, social sciences, literature - in short, I'd like to have a plausible answer (since we can never get true ones), to any of the questions my children might ask me. And the first step toward that goal that I could think of, is self-directed Knowledge Engineering. I can see that the human mind is totally buried under the avalanche of discovery that it created. Normal human cognitive capacity is already exceeded by the knowledge necessary for proper handling of only one area of expertise, not to mention the span that I set for myself as a target. So, the solution I'm working on is two-fold:
(1) to find a model of learning that would allow the kind of folding and unfolding (generalization and analysis) that only the human mind is capable of, and
(2) to provide that model with a consistency and span not available in the biological base of the human mind, but potentially available in the digital world.
I am looking for the mechanisms (as few and simple as possible), which allow the mind to create and deal with "abstract concepts" as well as with "the concrete data relayed by the senses". The model I am working on, MneMonic (see my dissertation topic proposal), complements the natural simplicity of a neural network with the consistency and compositionality of a symbolic network and stores the semantics of every concept, word or particle as a distribution of weights from (and to) all the other concept/word/particles that the network had experienced together (as part of the same context).

As a programmer I do not restrict myself to one specific programming language or environment. Before starting a task I consider the advantages and disadvantages of the available tools and strive to choose the right tool for the right task. Even so, I've experienced cases when the scope of the project enlarged outside the boundaries of a language or environment (as the end users seem never to be sure of what it is they need in the initial phases of project development.) In those cases I had to port projects from an environment to another or even from one programming language to another. Also, several times it helped to do a prototype of what the project was to become in a simple environment - usually a high-level one (e.g. text, sketches, Prolog) - and later on, for the sake of optimization and refinement, I continued in a lower level environment (e.g. imaging or animation programs, VisualStudio), or a more suited programming language (e.g. various macro languages, javaScript). My quite extensive survey of the major classes of programming languages and having experienced using several of them in intermediate to large projects (see my curriculum vitae for highlights), helps a lot in those kind of considerations.

The area of Communications and Information Technology is another one of my academic interests in which I had out-of class experience. Since I was 12 - in the 5th grade - I dabbled now-and-then in Journalism. I started by 'publishing' my own sci-fi stories, in manuscript and with hand-drawn and cropped-and-glued illustration and collages, for my classmates. Later on I participated in different literary and journalism clubs, and more recently published in the American University in Bulgaria (AUBG) student papers. I created database-supported and WWW-supported information systems (ISes): simple data acquisition and processing databases written in Pascal for Journalism and Sociology courses, a complex dedicated IS for archaeological excavations complete with detailed job descriptions, information flow and core database, an efficient distance-management system for general projects management (developed for the web-development company studioVISIA).

I was wondering whether to include gaming in the category of 'academic' interests. I decided to do so, because, as I never read anything passively, as a pass-time, I also never play a game without trying to analyze its strategy, it's 'Markov chain'-like progress and the ways in which the physical appearance and the rules (abstract appearance) of the game conduct the players to make each decision they make, in close analogy with their previous experience. I always try to compare my decisions as a novice to the ones that I make later, as I gain experience. In my research I try to find engaging ways to present stimuli to the experiment subjects in order to escape the 'scientific' environment which most probably influences the results of the experience and thus, I try to render them less artificial; to increase the ecological factor of the experiment. And what better way to do that than using games? By staging experiments in the midst of gaming experiences, and factoring out the parameters that differ from a real situation - like the reduced anxiety to being harmed, or the modality-reduced input - we get a far more ecological environment than that of a conventional lab. Thus a far more complete response spectrum, an increase attention/cooperation level, reduced cost for experimental subjects, a larger amount of data per subject, which all participates to increased accuracy of the experiment.

My experience in the Experimental Psychology Research Lab of the New Bulgarian University consists in consulting for the design of experiments, actual programming of on-line experiments, using Carnegie-Mellon University's PsyScope, audio and visual stimuli selection and presentation, data formatting, data management, and statistical analysis. I directed my own experiments toward verifying some of the assumptions I had made while implementing MneMonic (see my dissertation): Strategies in Learning to Use an Abstract Domain, Frequency Effects on Lexical Access as a Cultural Factor. Since I also work -for pleasure and money- as a webmaster, I do apply what I learn to that domain too, like in the Application of a Multidimensional Scaling Method to Web Design applied experiment tool. To give an idea of the spectrum of domains I'm familiar with, I will list the experiments on which I collaborated: Context Effects on Choice in Decision-Making and Context Effects on Evaluation (both supervised by Assist. Prof. Boicho Kokinov), Effects of Gender Agreement on Picture Naming (supervised by Assist. Prof. Elena Andonova), Norming of Picture Naming (international effort led by Elizabeth Bates of UCSD), Lexical Access of Nouns and Lexical Access of Verbs and Adjectives, (both part of a study supervised by Assist. Prof. Elena Andonova), Imagery Use in Understanding Idioms Cross-Culturally (leading to the defense of the Master's thesis of Armina Djanian), The Role of Gender and Number Markers on Language Processing in Romanian and Bulgarian Monolinguals and Bilinguals (leading to the defense of the Master's thesis of Mugur Badarau), Saturation Effects on Lexical Access (supervised by Armina Djanian). To further automate the more time-consuming portions of the process of running an experiment, I condensed my expertise in the object of my current Master's Thesis, MoStaCon (Monicsoft's STAtistics and data CONversion tool), an assistant in experiment design, data formatting and statistic analysis. But my involvement in experimental studies did not start when I arrived at NBU: while at the American University in Bulgaria, I have programmed an integrated, object-oriented experiment environment (PresentaTHOR), containing a 3-dimensinal object editor, a renderer and a 'slide' manager with reaction time capture for replicating on-line and exploring Meizer and Shepard's (1974) experiment on mental rotations for Prof. Anthony Birch's Ph.D. work.

As far as teaching is concerned, I look at it not only as a means of spreading and refining knowledge, but as a source for my own development in the area as well. My first experiences in this area date back to my high school years, when I used to help colleagues with their homework or prepare them for exams. Then, the new freedom that the Internet had brought to the distribution of information, prompted me to write WWW-technology tutorials and later a course in Multimedia technologies, which I used as training materials for the Electronic Publishing Team at AUBG; my advisor needed to advertise and publish materials and we needed people skilled in handling traditional media as well as virtual media - printed, taped and digital text, images and sound. Later on, at the New Bulgarian University, as a research assistant and as a teaching assistant, I adapted pieces of my Multimedia tutorials to helping fellow students, the other research assistants and the operators in the Experimental Psychology Research Lab in creating on-line experiments and dealing with the large amounts of data generated by these, as well as off-line experiments. I can also mention the teaching - again applied Multimedia skills and data organization - that I had to do as part of the task of optimizing the work of the team members assigned to me for various projects at the world-wide-web-development firm, studioVISIA, where I worked shortly as a Project Leader.

I hope that I have outlined already what I am looking forward to as I apply to Carleton's Ph.D. in Cognitive Science: to further my research in the modeling of language development. Here's a list of the courses I would be interested in: 95.506 and 95.555 Natural Language Processing, 95.510 Topics in Artificial Intelligence, 95.520 Cerebral Computation, 95.581 Topics in Machine Learning. However, my previous experience in Machine Learning left me wanting for a less pre-programmed, and a more emerging-based approach to learning algorithms. I would also be interested in: 95.510 Cognition and Artificial Cognitive Systems, 49.530 Perceptual Processes, 49.570 Experimental Research in Cognition, 49.624 or 49.625 Neuroscience Techniques, maybe 29.561 Cognition and Language and 32.534 or 32.535 Tutorial in Selected Problems of Philosophy. And, of course, I hope the seminars will be useful as debates on integration and hybrid approaches: 04.680T2 Proseminar in Cognitive Science, 04.690T2 Research Seminar in Cognitive Science.

As for my language skills, I am quite comfortable with English (653 TOEFL November 1995), since for the last 8 years it became the primary language I speak, read and write in - since my studies at AUBG and NBU were in English and my wife is Bulgarian with an American degree in Journalism. My French is a little rusty since I didn't have much chance to practice it, but it is the language in which I learned how to write; my godmother and her sister used to address me primarily in French. I also had translation certificates from and to French and English from the Romanian Ministry of Culture and I did extensive translation from these two languages (plus some Italian) for the Romanian Television.

In conclusion, given my background in DBMS, OOP and Web technologies, I could contribute to the work and research in the Language Analysis and Knowledge Engineering Research lab. Also, I hope that my education will benefit from interaction with the members of the Knowledge Acquisition and Machine Learning research group. I believe that Carleton's program in Cognitive Science is the right place for me to continue my research.

[ Home ][ Current Projects ][ Portfolio ][ Pastime ][ Pseudonyms ][ CV ]