Short bio
I am a medical doctor (MD) trained in General Medicine. I have a keen interest in computing that led me to a PhD in Medical Informatics. In my doctoral dissertation "A Deterministic Dynamic Associative Memory (DDAM) Model for Concept Space Representation," I explored research topics in the areas of knowledge discovery and representation, with relevance to Case-Based Reasoning, Natural Language Processing and Information Retrieval.
I have a special interest in uncovering fundamental aspects of Medical Informatics and medical decision-making and, in my quests for answers, I favour the creation of computational models of information processing. This approach which I like to call "applied informatics" is different than the usual "powerpoint informatics" where ideas and models are being proposed, presented and communicated without due consideration for their realization or implementation in an actual working system. My thinking about fundamental concepts has also resulted in the proposal of novel concepts such as the “knowledge spectrum”, the “usability axiom of medical information systems,” and the notion of algorithmic significance specifically aimed at complementing statistical significance, which is the traditional approach to judging the significance of clinical evidence. These theoretical notions helped me convey fundamental aspects of Medical Informatics to my students and to other audiences. Most importantly, they allowed me to view Medical Informatics as "context-dependent processing of medical information," a view fundamentally linked to such research areas as Case Based Reasoning, Information Retrieval and Algorithmic Information Theory. In particular, the concept of algorithmic significance as an alternative to judging significance of data is based on the detection/existence of ‘‘statistically rare (hence statistically insignificant) but algorithmically significant patterns” that posses algorithmic properties (e.g., description lengths) that are more important than their counts (i.e., statistics). That is to say that the existence of a pattern (or repetition) with a sufficiently long algorithmic description length that is discovered in two independent data sources (e.g., two individual clinical case descriptions) is significant for decision making and therefore a good example of algorithmic significance. Most importantly, this notion of algorithmic significance explains better the importance of clinical evidence available in Electronic Health Records (EHRs) where data may possess algorithmic properties are more significant than their statistical properties (e.g., frequency of occurrence). From a practical standpoint this view also calls for the re-evaluation of what is currently accepted as high-quality or best evidence in Evidence-based Medicine (EBM).
My teaching approaches have very strong applied components that complement and support understanding of theoretical concepts. In his view, this approach to teaching is difficult to separate from the kind of applied research aimed at the exploration of new technologies and their applications to healthcare problems. A relevant example in this regard is the development of a simple prototype, The Health Monitoring Station, that helps demonstrating typical properties of Health Informatics applications to non-technical audiences. One other such project is the Android Garage Door Opener, a simple proof of concept idea that helps with the exploration of distributed information systems architectures for sensor data capture, processing, analysis, communication, display and decision support.
Teaching
I taught/teach quite a wide a variety of courses on a variety of topics ranging from anatomy and physiology of the human body to research methods and epidemiology. Some courses are very technical (e.g., computer programming and database courses), some are higher level, more theoretical (e.g., health informatics, decision support, research methods) and/or applied (e.g., clinical informatics).
Hobbies
I love writing computer software. I get a kick out of every line of code I write. I do not know how it happens, but it seems that whatever I do, professionally or otherwise, I end up needing to write some piece of software. Most are prototypes that are not ready for use by others. I use a variety of tools and languages (C#, JavaScript, C) but I am partial to Linux and FreePascal compiler and tools. I've started with daydreaming about small Basic programs, followed by pen and paper, well before I got my first computer in '89. After that, I focused on Z80 assembler and tried whatever tools and languages I could try on my Z80 clone. Did not take long to come across TurboPascal in the early '90s once I got access to a PC.
I have working knowledge of electronics. This allows me to repair broken things, to design my own circuits and create my own printed circuit boards. But this has also been the cause of many a mishap that resulted in broken devices and/or release of smoke. While in medschool I've built a 3 channel ECG device that worked, for the most part. When I'll get the time, I will post a few images of it. Whenever I have the time and strength to get to it, I am usually busy devising new ways to automate things around the house. I managed to make lights turn on/off according to precise dusk/dawn calculations and to program the garage door to text me if I left it open longer than 15 min. Funny story about that, while shovelling some snow for longer than 15 minutes it kept bugging me with texts. Which made me invent a simple override button in an app, which may lead to a more sophisticated, 1 hour, temporary override button...which may need adjustable time, etc., you get the idea... which begs for some smart AI assistant to take care of the door on my behalf. Right now I have my eye on the darn water softener that runs out of salt and keeps too quiet about it.
I sometimes play music. I am at a fair level in guitar but usually willing to try whatever instrument comes my way. These days I often pound on the electronic keyboard we've originally bought for our daughter. One of these days I might record something and post it here...