Simfish/InquilineKea's Thoughts


more on discernment
March 30, 2008, 10:28 pm
Filed under: Uncategorized

It’s kind of like when you get to know people well, they oftentimes grant you privileges special to you and not to others (this is often a result of signal value – “competent + reliable” people for any field are a mere subset of all possible people who could be paired with the field, but with the exception of a population where people are inclined to be enthusiastic about things they suck at, the granting of such privileges is usually a response to a signal that carries correlative meaning (since the enthusiastic are more likely than average to be “competent/reliable” for the field).

(this may be true for all fields where appreciation is proportional to time spent – or fields where people tend to be internally motivated [oftentimes those where recognition is uncommon] – this may not apply for fields like political offices where many people are enthusiastic and where enthusiasm may be far less correlative with “competency”.

Advertisements


March 30, 2008, 10:17 pm
Filed under: Uncategorized

So commenting on the below post, I think some profs in some fields would be convinced (that it would be best for your own education) if you asked them for solution manuals for problems to do on your own as long as you kept them to yourself [in cases of perfect discernment when you can perfectly discern between when to do a problem and when you probably should look up the solution]

==

Another thing: gaps in knowledge. A lot of people use a person’s ability to recollect a certain fact as representative of the person’s knowledge of other areas of the field. Recollections of some facts are sometimes sufficient enough to show that you understand something – for example – recollections of the theorems of vector calculus are usually sufficient to show that you have a basic understanding of vector calculus (unless you happened to be a “rare in this population” individual who happened to memorize without comprehension).  Sometimes they’re also necessary – an inability to do calculus betrays an inability to do a lot of fields (although there are some amazing counterexamples – for example – dyscalculia – a difficulty with arithmetic – doesn’t always come with diminished intelligence – the authors of “origins of mathematics” recollect a patient who was able to do physical chemistry without knowing how to add!)

When it comes to specific facts, the signal value of whether you know “fact” or not is often dependent on the percentage of the population who go through a particular educational system/curriculum (and also dependent on a small percentage of people who self-study out of non-traditional books).  Of course most facts are related to others and so an inability to recite one fact will usually betray an inability to recite other facts in the area (although this is just probability – there will be gaps in every person’s knowledge, and some people will have gaps totally different from those of others)



Ideal discernment with perfect knowledge
March 30, 2008, 9:56 pm
Filed under: Uncategorized

Is of course never attained.

If you had perfect knowledge of your beliefs, then artificial constraints are useless. Artificial constraints include retirement accounts that don’t accumulate interest and the prevention of student access to full solution manuals and teaching materials.

It is probably true that a subset of students will learn better if they had access to both teaching materials and full solution manuals. In fact, access to full solution manuals is the basis behind a lot of self-study programs (and it’s quite possibly true that a number of students with access to them do learn better with them than without them). But policies are directed towards the vast majority of students and the variability of the behavior of them during the periods they’re most likely to pursue the subject of interest. It is obvious that a person’s impulse control varies throughout the day and that a person at consistently peak impulse control is probably able to discern between what’s best for himself and what isn’t best for himself given perfect knowledge of what he finds perfectly appropriate and what he doesn’t find perfectly appropriate. But consistently perfect impulse control is rare and so people, even with perfect knowledge of the long-term benefit functions of their various actions, are oftentimes physically unable to select what’s best for them at all times.

(there is a difference between the keywords “select” and “discern”.) “select” in this context implies perfect knowledge with failures of impulse control. “discern” in this context implies imperfect knowledge.

So for example, most people realize that they have to save for retirement. They intellectually are able to discern between desirable long-term savings behavior and undesirable long-term savings behavior. Yet they cannot always select what’s best for them due to failures of impulse control and so artificial policies are sometimes needed in the context of perfect knowledge.

(but here what is perfect knowledge? Perfect knowledge at every give time implies that one pursues the action most conducive to one’s own sustainable welfare). But we can at least say that people can have perfect knowledge at peak moments of impulse control but imperfect knowledge when their impulse control fails (as in, “I think I’ll be happiest for the long-run if I just buy this one more thing/I’ll look at this solution to this one problem” during a impulse control failure.)

word count: 389



research priorities
March 20, 2008, 11:23 pm
Filed under: Uncategorized

It so happens that there is a potentially infinite amount of research you can conduct on any subject. Yes, this includes subjects such as football games, computer games, reading habits, library-using habits, blogging habits, and so on. It just happens that there are some issues that academic researchers find more interesting than others at the particular time (such “interesting” issues are socially determined, although they may not be completely arbitrary as at any given time there are some activities that influence the behaviors of a substantial number of people and other activities that do not influence such behaviors).But we can apply the “natural” vs. “experiential” analysis again here. Oftentimes the easiest research to do is the research that is on subjects that have already been analyzed. After all, comparisons and verifications are easier on such subjects. However, this alone does not make such subjects as more worthwhile research pursuits than other subjects. There are some fields that are more “stable” than others, stability being determined by the history and depth of research in the field.It should also be noted that this applies to math and science research. A lot of subjects actually publish papers on theoretical math – but different fields of theoretical math. Even psychology professors write papers on theoretical math – just on equations they find more applicable to their immediate problems than theoretical math professors do (after all, there are a potentially infinite number of functions one can analyze). Much theoretical math progress has come out of those alternate fields. Physicists with string theory, population geneticists with Fischer tests, etc.


what’s “naturally” optimal vs. what’s “experientially” optimal
March 20, 2008, 11:15 pm
Filed under: Uncategorized

Goal: (Basically this addresses the question “why do you choose action X rather than action Y when you haven’t given equal consideration to both of them?” action Y may be more optimal for you in the long run than action X, even if action X happens to be more locally optimal for you)

As it stands, most people are fairly more or less aware of their strategies optimal for a given goal. However, they frequently fail to distinguish between what’s naturally optimal for them as compared to what’s experientially optimal for them. There is quite a huge distinction between the two. Sometimes the differences between what’s naturally optimal and what’s experientially optimal aren’t that great – strategies are not inflexible and so people’s range of behavior is variable even within the context of each strategy. Nonetheless, strategies are often built on the result of “stable states” that come out of environmentally-influenced bifurcations.

Human nature and environmental constraints are not infinitely flexible and so there are usually strategies that are naturally optimal for realistic phenotypes and environments, even if they may not be optimal for extreme environments or completely different phenotypes.

For example, many people are overly socialized in such a way that they believe that they learn best from lectures as compared to self-study. What often happens, though, is that they don’t even pre-study before lectures. And what often comes out of pre-study is the realization that one can continue pre-studying and then one discovers that one’s lectures are actually useless beyond preparing for ad hoc exams. In this way, what’s “naturally” optimal for a lot of people isself-study, but what’s “experientially” optimal for them happens to be lecture-based learning – as they’ve been socialized to learning from lectures and so they realize their maximum benefit per unit of time [at any given time] by continuing to learn from lectures (as it takes time to develop experience through self-study). “Natural” optimality is only possible to measure if people have equal opportunity to be exposed to both styles of learning. Even then, one has to consider that the resources of one’s youth are different from the resources of one’s later ages. Self-study is a more sustainable learning style than lecture-based learning since you can self-study anything at any time.

Similarly, many people are socialized to type fastest on QWERTY keyboards rather than Dvorak ones. However, Dvorak keyboards are customized to ensure fast typing speeds, whereas Qwerty keyboards are not customized to ensure fast typing speeds. Yet it takes time to get used to type on Qwerty keyboards – a lot of time. And so what’s experientially optimal for most people is to continue typing on Qwerty keyboards even though it’s naturally optimal for them to type on Dvorak ones. One must also consider that in the time being, qwerty typing ability is more sustainable than Dvorak typing ability since one must use qwerty keyboards away from home.

In the same case, we’re socialized to do a lot of things that aren’t necessary (we’re socialized to do what’s “locally” optimal rather than what’s “naturally” optimal). Such as…

– Eating meat and refined grains (when there are plenty of vegetarian foods that taste great). You don’t need to be an animal rights activist to realize that meat is extremely inefficient resource-wise.

– Sleeping on beds mounted on bedframes. Like seriously, you can just put the mattress on the floor

– Communicating academic information by the spoken form rather than the written form. Yes personal information is oftentimes communicated better with body language. But body language is useless when it comes to academic content – especially content delivered for lectures. Written information is archivable, distributable, and retrievable in the future.

– Showering daily. While some people must shower daily (for the purpose of smelling “clean”), others can go for days without showering and still smell “clean”).

– Participating in price-habituated activities that replace price-sensitized one
(read Rachlin’s “Science of Self-Control” for explanation). In fact my theory is pretty much another version of his primrose path (although it’s on a more global level)



test2 for google reader
March 9, 2008, 7:00 pm
Filed under: Uncategorized

i deleted test1, it still appears on google reader. so i want to do test2



list of abstract quantifiers
March 6, 2008, 11:16 pm
Filed under: Uncategorized

scientific intuition

overloading

mathematical maturity

http://en.wikipedia.org/wiki/Time_horizon

maximization given constraints (useful in both math and RL)

auto-parse