Oxenfurt lecture hall

+
Oxenfurt lecture hall

Welcome to the University of Oxenfurt. Here we will discuss all fields of academic thought, including but not limited to natural science, mathematics and computing, engineering, humanities and social science.

A personal area of interest are problems at the edge of experimental approaches. That is, the very small and the very large: astrophysics, nanoscience, neuroscience. Another topic worth pursuing is computational science. For instance, the above areas would hardly exist without mathematical modeling and computer simulations.

My area of research is theoretical computer science, artificial intelligence, cognitive science and computational cognitive modeling.

Some principles everyone must observe:

1) There is no discipline inherently better than another. They are all aspects of human knowledge. Some may be better suited to explain phenomena, but be respectful of each other.

2) If you make a strong statement, back it up with proper, reputable sources. This is no place for rumours, although intelligent speculation is welcome.

3) Treat everyone like your equal. Others may not be experienced in your topic so don't disrespect them or undermine their desire to learn.

4) Have fun and don't bring up discussions explicitly prohibited by the forum.

Certain topics to begin with:

- Robot assistants and challenges of artificial embodied cognition.
- In numbers we trust. Scientific computing guides our hand.
- Human and animal experimentation. Behavioral and invasive neural and chemical approaches.
- A discussion on which formal methods are actually used in social and clinical sciences.

And anything else you like. From Aristotle to Galileo to the recent Nobel price for work in computational chemistry.
 
I think you should start off with a statement with which we can initiate the discussion. Why did you not quote your conversation about osychology here? :)
 
Think I may have to force myself away from this topic, laptop isn't working and I can't gather my mindblowing sources together on my phone >:[
 
Ok let's start defining the methods and object of study of areas of interest.

Since it came up in the other thread, lets discuss psychology.

From personal experience psychology seems to concern itself with human behaviour, which is externally visible and measurable, and less with internal cognitive processes such as memory, learning, perception, and so on. Of course these are related, but cognitive psychologists do work on these topics as well as mental representations (Johnson-Laird for instance), decision making (Tversky), heuristics (Gigerenzer) and even computational modeling (Edelman).

The field is very diverse and therefore there are many different approaches and methods, some more formal than others. The point of a formal language is that:

1. It uses clear, non ambitions symbols.
2. Has clearly defined grammar and inference rules.
3. Is finite.

For instance, first order logic, arithmetic, and many mathematical systems.

What tools do psychologists use to represent theories and results, in a clearly reproducible, non ambiguous manner?

The one problem I personally find with many psychologists is the tendency to name and informally describe phenomena, and fail to provide either a formal explanation or establish a causal relationship between observations and results.

Many psychologists have provided great contributions to many fields, including artificial intelligence. The problem is what they study is generally difficult to approach, which results in inadequate methods for the construction of explanations. Also they need to learn basic math :) (look at Shimon Edelman's work).
@pandoricai

We would love to have your input. Please take your time and post, the thread isn't going anywhere and my knowledge of your field is insufficient.
 
Isn't psychology mainly a field of empirical studies, where you then try to form results from these? The main problem opposed to physics for example is the reliability of your results and the ability to reproduce them. Since you are working with people and their experiences of themselves, it is all much more subjective.
This doesn't mean though that it is not important or necessarily wrong, but since we in most sciences form theories with a certain confidence, I feel that in most natural sciences we can be very confident in our theories until we find a better one. In psychology for example, I feel, the confidence in your theories is much lower, since you are lacking many of the scientific tools to measure input and output. All of your results come from a statistical background and as the saying goes: Don't trust a statistic you haven't forged yourself. :D
 
Quick thing before I work my way up to a full post - psychology isn't just behaviours and underlying processes as such, it is not always a vague theory or observation. I was involved in a study where the researcher was using something I can't remember the name of, but it basically fired a magnetic pulse at very specific areas of the brain to measure time perception and reaction speed. Just a basic example, but I'd say this sort of thing, as well as manipulation of hormones and neurotransmitters etc is replicable and non-subjective as some things HAVE and continue to be replicated and proven reliable. I think the problem is that psychology covers lots of things - we do study behaviourism which I would say is a social science, however the main modules are psychobiology and neuroscience. Long standing and useful results have been gained from many studies in those fields and I think it would be ridiculous for people to look at, for example, behavioural conditioning studies and say "look they're just watching that rat, there is no quantitative data, it's not science." (Though I'd disagree even on that basis) People are exposed to a lot of pop psychology and seem to forget that psychologists conduct research using biological and chemical methods - but since the context is in behaviour or emotions or mental illnesses etc., it is a 'thing' of its own...I had something else to say and that probably didn't make sense but I feel pretty ill at the moment, so I apologise :p

(Also I'm better at statistics than behavioural psychology :p)

EDIT: @Mataresa, we do come up with our own statistics 0.o
 
See, I was under the impression that psychology applies specifically to behavioural and emotional analysis and the relation between those two. I would have placed those experiments you mention @pandoricai more in the realm of neurobiology or physiology of the central nervous system. So, I would have expected those experiments to be performed by someone who studied either biology or medicine, for example, rather than someone who studied psychology. It´s always good to learn about one´s misconceptions!

Maybe my mistake arises from the different angles psychological studies can take, and how it is taught here. Maybe the angle they teach here in Argentina is more directed to behavioural studies rather than neurophysiological ones.
I know a couple of psychologists, so I could ask them about it...
 
I was under the same impression @gedierond.
@pandoricai I never accused anyone of making up statistics (if that was what you meant with your response), the saying rather refers to the fact that how statistics are presented can influence the conclusions you take from them a lot and a lot of the time how you obtained this data influences your results as well. Often measuring data happens with a certain wished result in mind so it is not that easy to conclude from statistics. This applies to all fields of science which measure data.
 
The field of Psychology seems swamped with nomenclature, it needs a holiday or a course of medication to sort itself out. Theres an awful lot of interesting and useful work & knowledge out there under that umbrella, but since no ones mentioned the ugly siblings Psychoanalysis & Psychiatry i'll throw them in there as a possible reason for the misunderstandings (I can't abide those high subjectivity low evidence "sciences" myself).

Nice thread Volsung, I'm sure i'll make a proper contribution some other time ;)
 
I'm not an an expert in the field, probably have very little of worth to add, but I am eager to chime in :ice: Having only 10 minutes of a formal psychology education from when sat in the wrong class, let us proceed :p I did however read and enjoy Carol Gilligans's In a Different Voice recently, and since the matter of "wished results" was mentioned , its a great analysis of what this fixed mindset can do. It discusses Lawrence Kohlberg's work with children in which he gave them a sort of "moral aptitude" test to examine the moral maturity of males vs females. Kohlberg reasoned males were more morally mature, based on the answers given to his questions and how they measured on his scale.

A brief and famous example: A man enters a drug store to buy medicine for his sick wife. He cannot afford it. Should he steal it?

The boy reasons that he should. Life is a more valuable currency than money. The girl disagrees. She asks should something happen to the man, say that he is arrested, what is to happen to his wife then? Like Kohlberg, I dismissed this remark at first, thinking she missed the point of the question, but as I read further, I realized I was being very short sighted and that she was answering "whether the man should steal the drug", whereas the boy was answering, "whether the man should steal the drug".

Its very good and a very accessible for non psychologists like myself :yes
 
There are neuropsychologists who do perform more of the verifiable type of experiments, but in the neural level there are still unsolved issues to explain the full process of neural computation.

Take for instance a common method: neuro imaging. Highly dubious as it depends largely on an expertly designed protocol. Very sensitive to errors. And not fully conclusive. You can only establish associations but not full causal relations (ie. It WAS caused by this input and not anything else, as Mataresa said). But it's a useful non invasive tool, and it has to be used until something better is designed.

The way Pandoricai explained it, psychology is more of a clinical than a social science, similar to psychiatry. There are other, mathematically oriented psychologists who have provided very satisfying models. I mentioned some, definitely not pop psychology.

Many experimental psychologists measure response times, for instance in human decision making tasks. The problem is working with humans (not fully understood complex systems). You can measure externally visible behavior and try to isolate internal factors (such as neurotransmitters). But ultimately you'll be working with a semi black box until you can explain its elements and the causal relations governing the systems dynamics that clearly and non ambiguously explain interactions between different levels of abstractions.

Psychology is a very broad field and I think what Pandoricai explained definitely gives it a scientific face. There are also psychologists who simply name things, and then the psychoanalists...

I think as sciences mature they (have to) become more mathematical and computational. What do you think?
 
I agree, so far as it is possible for the sciences that deal with wetware (life in general, or humans in any way). The signs of maturity as a science include the development of a common lexicon, though not always a formal language; and the use of mathematics, even when its use is limited to measurement and statistics.

The basic problem with wetware is that it obeys the "Harvard Law," "Under controlled conditions of light, temperature, humidity, and nutrition, the organism will do as it damn well pleases." [Larry Wall's version]
 
See, I was under the impression that psychology applies specifically to behavioural and emotional analysis and the relation between those two. I would have placed those experiments you mention @pandoricai more in the realm of neurobiology or physiology of the central nervous system. So, I would have expected those experiments to be performed by someone who studied either biology or medicine, for example, rather than someone who studied psychology. It´s always good to learn about one´s misconceptions!

Maybe my mistake arises from the different angles psychological studies can take, and how it is taught here. Maybe the angle they teach here in Argentina is more directed to behavioural studies rather than neurophysiological ones.
I know a couple of psychologists, so I could ask them about it...


Well considering my degree is a BSc in psychology and they are parts of CORE modules which means in England, to call yourself a psychology graduate you must have a fairly good background in such types of experiment...they come under psychology, though perhaps not EXCLUSIVELY ;) As a fairly 'young' science it's not surprising that there's confusion about it and its sub-divisions, particularly between countries...but here, you could easily do a degree in psychology with minimal behavioural psychology included if you so wished - but you couldn't skip the type that I was talking about

(Oh, and here the lecturers positively poke FUN at Psychoanalysts ;) )
 
Last edited:
I have a lot of respect for psychoanalysis. It might be philosophical, but it still has merits in practice, which makes it relevant. Besides philosophy is wonderful. Exploring the thoughts of the unknown and trying to push the limit of your own mind to realms that elude your senses and ability to perceive or understand.
 
Alright @pandoricai. You've read what we said, now it's your turn to tell us your perspective on psychological research :) For instance, open problems, methods, experiment design, analysis, etc.

I once met a group of developmental psychologists who needed some computational processing to automatically interpolate missing data on a longitudinal study. The problem was they were missing like 80℅ :p Not a very trustworthy study...

What area is next? Biology? Chemistry? Or a favorite of mine, computer science vs computer engineering vs software engineering. Or computational neuroscience (biology + physics + computer science & applied math).

Edit: got it. Somebody brought it up elsewhere. "Is engineering science?".
 
Last edited:
Very briefly, you can say engineering is the science of making things practical. Whether an activity is engineering is properly judged by whether it does two things:

1. It uses scientific methods and principles (not just the strict-sense "scientific method", but any activity that is in the sphere of science).

2. It has as its objective the creation, use, or improvement of artifacts (used inclusively, meaning everything from structures to processes to algorithms).

It's understood that these artifacts must be useful (productive, economic, safe); otherwise the difference between engineering and self-flagellation becomes one of degree and not of kind.

I'd say most of what passes for computer science is also engineering, because it is intended to discern ways to use computers to do useful things. There is a standing debate within the ACM over whether theoretical computer science (in the strict sense, meaning it is done to advance theory, and engineering results are at most useful side effects) exists at all. Since a key test of a theory in computer science is whether you can make a computer do that, the boundary between theory and artifact is blurred like the back line of the batter's box.

It's more profitable to distinguish between computer science and engineering on the theories, language, and methods employed. Computer science relies more on mathematics, formal language, and abstract models; engineering relies more on concrete language, hardware, and process.

A psychologist conducted an intelligence test. His subjects were a physicist, a chemist, and an engineer. He gave each a red rubber ball and told them, "Tell me the mass of this red rubber ball."

The physicist measured the circumference and oblateness of the ball to determine its volume, dropped it from a height and measured the rebound, deduced the density of the rubber, and computed the mass from the volume and density.

The chemist submerged the ball in a graduated cylinder and used Archimedes' Principle to compute the mass.

The engineer took out his red rubber ball catalog and looked up the mass.
 
Last edited:
@pandoricai

Wanna give some closure comments on the topic of psychology?

@Guy N'wah

That's a very good point. Engineering is concerned with making things work, but sometimes engineers have to develop theory first (see computational science). I think the objective is what sets things apart.

Computer science is a very broad field (ACM has like 18 subareas contained within it). When it lies next to engineering, I'd say it is concerned with making computers solve problems. When it sits next to math, it is concerned with models for general, sometimes abstract problem solving with solutions responding to some theory of computation (solvability, tractability, complexity). Many computational models are very satisfying mathematically but highly impractical. For instance Partially Observable Markov Decision Processes are a solid frameworks to model human decision making under uncertainty but their analytical solutions lead to very complex math and they usually become intractable very quickly. A typical research paper in theoretical comp sci would be something like "lower bounds for X-type graph Y-type search". Add some computing paradigm (not architecture) such as parallel computing and prove the method converges to an optimal value in order f(n) steps where f(n) is asymptotically better than all previously known alternatives. No programming involved, doesn't solve anyone's problem directly :p (in the long term yes). I'd say computing is science when it asks the questions and develops a strong theory to try to answer them, making it essentially math (applied to problem solving). And it is engineering when one such question is answered in a context in which the only restrictions are efficacy and efficiency, not elegance or plausibility. For instance, modeling an unstudied phenomenon vs. writing a program to solve P applying X's method because it works.

There are some fields where engineers usually also become scientists (eg. Materials science and engineering) and where computer scientists become engineers (eg. Computing cluster admin).

As usual lines are blurry though :)

Edit: wild mess there, posting on my phone.
 
Last edited:
The engineer took out his red rubber ball catalog and looked up the mass.
Ha, good one.

Just popped in to say that I find the topic very interesting. My training was in German literature with specialization in medieval. I've tried very hard to maintain a connection to historical research, despite the high cost of scholarly books and independent access to databases. My reading since college has mostly been late antiquity/ early medieval history. However, I don't have enough in depth knowledge to call myself a scholar of it. The day job takes too much out of me.
 
Hey @veleda,

If you want to share some of your knowledge with us you're welcome to!

I once took a linguistics graduate course and was kind of fascinated at how phonological analysis uses formal systems to describe transformations and derivations. This has allowed some linguists to trace these changes and map them historically and geographically, leading to strong theories of migration and cultural exchanges. Can literature be used as a similar tool?

(we can manage two topics at once, right? ;))
 
Top Bottom