Let ε' < 0

Protected: zz

lowai — Thu, 27 Jan 2011 00:44:29 +0000

This post is password protected. You must visit the website and enter the password to continue reading.

Increasing stupidity

lowai — Tue, 13 Apr 2010 04:23:12 +0000

Had been wondering about an oscillatory integral of the first kind for over an hour. At last, I suddenly realised that I mistook the following inequality for the opposite direction: x^{-1}' title='x^{-1/2} > x^{-1}' class='latex' /> for 1' title='x>1' class='latex' />. Damn it!

A Parody

lowai — Mon, 05 Apr 2010 01:20:42 +0000

I had a sudden impulse to write a parody of All Kinds of Everything (youtube). So difficult to make it rhyme… =’( It’s also difficult to find appropriate words since the end of each line is a strong syllable…

Shock waves and fluid flows, convection of heat,
Laplace and Neumann, things of PDE,
Boundary value, initial value, solve in ,
All kinds of everything remind me of you

Neither fish nor fowl

lowai — Thu, 07 Jan 2010 01:55:35 +0000

Yes, that’s the situation of my research. My supervisor introduced a new problem, which is, in my opinion, purely application and physics. Suppose there are transmitters at unknown positions and receivers whose positions are known. We want to determine the positions of the transmitters using the signals received by the receivers. Actually, there can be possible positions of the transmitters, and we need only to specify out of them. Laborious computation will convert this to a sparse recovery problem, where the measurement matrix may not satisfy the classical requirement of pairwise independence but is small, say, bounded above by . Then the usual techniques will be applied (although I think the mathematical foundation for this more generic case is not solid — I’d rather work on this aspect). He listed some possible improvements, all of which were more complicated phyiscal models. Obviously this is not theoretical computer science at all, but is it applied mathematics?

(I admit that it interests me — I don’t object applications — analysing physical models can be fun and may involve a lot of skills too, however, I need to care about publication, which will directly and closely affect my future job.)

First two formants

lowai — Mon, 28 Dec 2009 15:50:33 +0000

The means of specimen formants (F1 and F2) of the sounds are from a paper by J. C. Wells in 1962. Standard deviation can be found at his website.

Sound	Specimen Formants	My Formants
i: (see)	200 2300	231 2500
ɪ (city)	360 2100	381 2090
e (bed)	570 1970	530 1893
æ (cat)	750 1750	726 1526
ɑ: (father)	680 1100	719 926
ɒ (hot)	600 900	605 872
ɔ: (law)	450 740	524 752
u: (soon)	300 940	257 732
ʊ (book)	380 950	394 959
ɐ (run)	720 1240	708 1475
ɜ: (bird)	580 1380	581 1325

The ESL teacher asked me to say the word not, and said I had a British vowel. She said the British vowel was slightly higher in the vowel chart than the American counterpart — I never thought about formants, to me, the difference was just roundness — It’s amazing that she could tell it!

A Stanford Parser

lowai — Fri, 18 Dec 2009 21:38:55 +0000

The parser is http://josie.stanford.edu:8080/parser/. And I used it to parse the following sentence

No assurance security or payment which may be avoided under any enactment relating to bankruptcy or under section 127, 238 to 245 (inclusive) of the Insolvency Act 1986 or any of such sections and no release settlement discharge or arrangement which may have been given or made on the faith of any such assurance security or payment shall prejudice or affect your right to recover from the undersigned to the full extent of this Guarantee as if such assurance security payment release settlement discharge or arrangement (as the case may be) had never been granted given or made.

It ran about half a minute and returned a mistaken answer, which read `No assurance security or payment … shall prejudice or affect’ and `any of (such sections and no release settlment discharge or arrangement…)’. Obviously the correct structure is `No assurance security or payment … and no release settlement discharge or arrangment … shall prejudice or affect …’

The following is the returned structure by the parser.

(ROOT
  (S
    (NP
      (NP (DT No) (NN assurance) (NN security)
        (CC or)
        (NN payment))
      (SBAR
        (WHNP (WDT which))
        (S
          (VP (MD may)
            (VP (VB be)
              (VP (VBN avoided)
                (PP
                  (PP (IN under)
                    (NP
                      (NP (DT any) (NN enactment))
                      (VP (VBG relating)
                        (PP (TO to)
                          (NP (NN bankruptcy))))))
                  (CC or)
                  (PP (IN under)
                    (NP
                      (NP (NN section) (CD 127))
                      (, ,)
                      (NP
                        (NP
                          (QP (CD 238) (TO to) (CD 245)))
                        (PRN (-LRB- -LRB-)
                          (NP (NN inclusive))
                          (-RRB- -RRB-))
                        (PP (IN of)
                          (NP
                            (NP (DT the) (NNP Insolvency) (NNP Act) (CD 1986))
                            (CC or)
                            (NP
                              (NP (DT any))
                              (PP (IN of)
                                (NP
                                  (NP (JJ such) (NNS sections))
                                  (CC and)
                                  (NP (DT no) (NN release) (NN settlement) (NN discharge)
                                    (CC or)
                                    (NN arrangement))))
                              (SBAR
                                (WHNP (WDT which))
                                (S
                                  (VP (MD may)
                                    (VP (VB have)
                                      (VP (VBN been)
                                        (VP (VBN given)
                                          (CC or)
                                          (VBN made)
                                          (PP (IN on)
                                            (NP
                                              (NP (DT the) (NN faith))
                                              (PP (IN of)
                                                (NP
                                                  (NP (DT any) (JJ such) (NN assurance) (NN security))
                                                  (CC or)
                                                  (NP (NN payment)))))))))))))))))))))))))
    (VP (MD shall)
      (VP (VB prejudice)
        (CC or)
        (VB affect)
        (NP (PRP$ your) (NN right))
        (S
          (VP (TO to)
            (VP (VB recover)
              (PP (IN from)
                (NP (DT the) (NN undersigned)))
              (PP (TO to)
                (NP
                  (NP (DT the) (JJ full) (NN extent))
                  (PP (IN of)
                    (NP (DT this) (NN Guarantee)))))
              (SBAR (RB as) (IN if)
                (S
                  (NP
                    (NP (JJ such) (NN assurance) (NN security) (NN payment) (NN release) (NN settlement) (NN discharge)
                      (CC or)
                      (NN arrangement))
                    (PRN (-LRB- -LRB-)
                      (SBAR (IN as)
                        (S
                          (NP (DT the) (NN case))
                          (VP (MD may)
                            (VP (VB be)))))
                      (-RRB- -RRB-)))
                  (VP (VBD had)
                    (ADVP (RB never))
                    (VP (VBN been)
                      (VP (VBN granted)
                        (VP (VBN given)
                          (CC or)
                          (VBN made))))))))))))
    (. .)))

Dental Amalgam

lowai — Sun, 13 Dec 2009 20:26:03 +0000

It is very disturbing that some dentists still prefer dental amalgam (which has been banned in Norway, Sweden and Denmark) because of the traditional impression of better durability over composite filling. There has been a rapidly growing concern of toxicity of mercury contained in the amalgam — the mercury vapour inhaled into human body and accumulated in brain, liver, kidney, intestine, &c. It does not necessarily cause any severe medical conditions but it is entirely possible to change your body a little — memory loss, difficulty in concentration, constipation, &c. Three weeks after I had the first amalgam filling, I began to feel constipation (it really is rare on me) — Not sure if there is a relation though. I think I’d better have a toxin examination some time in the future.

TA application sort of rejected

lowai — Sat, 12 Dec 2009 03:27:50 +0000

My supervisor informed me today that the department wanted him to appoint me as an RA unless I had a good reason for being a TA. Possibly there are excessive applicants and my English is miserably poor…

Silver Bells, Silver Bells

lowai — Sun, 29 Nov 2009 06:29:47 +0000

Well, not the original song, but the parody by Dick Lee: http://www.youtube.com/watch?v=kP1kcCpvYEk (0:30 to 3:45) And I hope that you help me to get all the lyrics – there are some words that I am not sure and marked by `?’ below.

City sidewalks, busy sidewalks
Dressed in holiday style
In the air
There’s a feeling of Christmas
Children laughing
People passing
Meeting smile after smile
And on every street corner you’ll see

Traffic jams, traffic jams
So many cars in the city
Hear them horn, dust to dorm(?)
Soon it won’t be Christmas Day

From Dynasty to Marina
People walk to and fro
In the air
There are camera flashers
Everybody shooting photo
Why ??? take I don’t know
Cut the picture from newspaper lah

Coloured lights, flashing lights
Too many lights in the city
See them shine
They’ll make you blind
And which ones are traffic lights

If you plan a celebration
Bring a card preferably gold
Suddenly all the prices have doubled
You’ll be served by ???
Saying `mister why not try our beer ??? buy one for one’
Then you will wish you’ve stayed home

Endless lists of Christmas gifts
But for the best Christmas shoppin’
Why don’t you
Just go to
KL or Johor Bahru

Christmas Day, oh, go away
It’s just one more reason for spending
If I try close my eyes
Will Santa Claus pass me by, oh no,
But soon it’ll be kong si fa tsai

Waterloo

lowai — Sat, 21 Nov 2009 17:12:58 +0000

I really was astonished when I saw the mark of my midterm exam of Continuous Optimisation — the exam was essentially multivariable calculus! My mark was even below the average and median — What a shame! It is very uncommon for a Chinese to perform so poorly in a math exam. It reminds me of what a Chinese professor at University of California, Irvine said. He said his university enrolled two math students, who had excellent transcripts, from SJTU, and both of them got C in some course. Everyone was surprised, as he said, because it is already abnormal if a Chinese student does not get an A, let alone a grade of C! Thus the image has formed that SJTU students are bad at maths. It is really sad, indeed, that a new instance further confirms his statement.

I made so many miscalculations in the midterm that I couldn’t believe them myself. They are just differentiation – yes, differentiation — an operation for which you don’t need to use your brain. But I made mistakes such as (I didn’t know I wrote that! Actually I wrote this at the beginning of the solution, so I lost almost all points of that whole problem. The exam marker wrote on my answer sheet `Your methods are correct, but be careful with your calculation’) and (I really don’t know why I didn’t write down the square in the denominator) It was utterly a mess.

The lecturer is kind enough though, who has announced the policy to boost the grades: Anyone who gives a twenty-minute presentation will get his score promoted by one level, i.e., B- to B, C to C+, and so on. This applies to those with scores below average only. Later the weights of exams were announced to be adjusted. For those who get scores below average in the midterm exam, the original weight 40-40 may be adjusted to 30-50 even 20-60, depending on how much progress is made in the final exam. It sounds relieving. It would be very shameful if I can not get an A- for such a stupid math course.

Meanwhile, to maintain my GPA, I definitely need an A+, and the only possibility is from the Stochastic Processes course. My midterm result does not look bad. I thought I messed up the problem of discrete option pricing, whilst it turned out that I got two points deducted only.