Jie 的个人资料Sunset (hhj's public hom...照片日志列表更多 工具 帮助

日志


2007/11/25

Calling for Suggestions

I have some questions about Economics research. 
 
I am currently working on a project about Dynamics in Human Society. After have read a series of papers, I think I can fairly conclude that the research method they (my supervisor's team) were using is basically to simulate the society using Multi Agent Simulation technique, and then trying to extract interesting features from the simulate society. In this method, each individual is simulated as an agent, with a few internal parameters reflecting his or her mental state. The entire society is then simulated as a population of agents plus a set of rules of interaction. For instance, agent can choose to talk or fight to others. The decision of which interaction to take is based on agent's mental state. Meanwhile, each type of interaction also has different feedback effects on the agent's internal paramters. As time goes on, the simulated society would show some emergent society-level behaviour from the pre-defined rules on the agent. Most research was conducted so that different causal relation between individual's psychological feature and macro-scale society behaviour can be discovered or examined. For example, one research is about how culture was emerged because of sub-consciousness and consciousness in human mind. Another research is about how social norm was transmitted because of shame and pride. So on and so forth.
 
After the above background introduction, let us go back to my own question about Economics. Most part of my supervisor's research in this direction is closely related to sociology rather than Economics. However, I think (and my supervisor also agrees) that Economics would also be a perfect field to adopt this method. For instance, we have Micro Economics which deals with a single market or a single manufacture, with external factors being treated as 'environment'. Meanwhile, we also have Macro Economics which deals the emergent behaviour of the economy of entire country or the globe. However, it is noticed that the economy is much more complex than the scenario described above, for the fact is that every micro part's environment is actually formed by other parts of the system. As a highly inter-connect (rather than layered) system like this, it is now largely unclear that how micro-scale features of the system contributes to macro-scale phenomena. To fill the gap, simulation can be used. Another possible scene where this method can be used is dealing with complex situation which has weaker assumption on the economy. For instance, how would the market evolve if universal knowledge is not accessible? Or, how would the economy evolve if individual does not always perform rationally (and we can even further adopt some mathematical model in psychology in this case)? Pure mathematical analysis is hard, and mostly impossible (due to our insufficient knowledge of non-linear systems and/or chaotic systems). In this situation, Multi Agent Simulation can be a very effective alternative.
 
Therefore, the theme of my project report would be a survey of the application of this method in Economics research, and try to connect them to current research in complexity science and the field of sociology that I explained above. However, I found nothing really interesting so far. I have tried some combinations of key words, but searching outcome was rare, and mostly irrelevant. In other words, I think I failed to find a start point, a portal, a gateway to this whole matter. That is why I am calling for help.

Many thanks in advance to any suggestion!
 
(Today's photo: Kui Men, Three Gorges, shot on 01/Jun/2007)
DSC01236
2007/11/22

Thanksgiving Day, why so popular?

Technically, Thanksgiving is solely an Amerian festival, which has no direct or indirect relation with any other countries, not to mention China. For instance, British people do not celebrate this festival at all. However, why there are so many Chinese people have written something on this particular day?

I reason that there are two major reasons. First, the influence of the famousTV series 'Family Album USA' can not be overlooked. Indeed, this famous TV series used to be the only multimedia English learning material in China, especially for people like me, aging from 20~25. Even I, who hated learning English very much during my early days, remember the episode of Thanksgiving Day in this series. Learning a language implies more than accepting the language itself. It also means to get understood to their culture, their society and their belief. Unfortunately, British people did not have the wisdom of  occupying this frontline with their own learning materials. Otherwise, Guy Fawkes's day would probably become a well-known festival instead of Thanksgiving.

Another reason lies inside the nature of our own culture. Many people claim that eastern culture is different from westerner's in that we tend to prevent speaking out our own feeling to others, even it is appreciation. By contrast, appreciating others by action rather than words is highly recommended. From another persctive, it can also be explained that easterners tend to present their feeling in more inductive way. Therefore, to have a chance that one can formally appreciate others without being regared as 'insincere' may seems attractive.

To conclude, the popularity of Thanksgiving Day among Chinese people is not a coincedence but a deducible consequence of various factors.

(Is that like a standard IELTS style essay?)

Should I also give some of my appreciation verbally on this particular day? OK, Thank you all for visiting!

(Today's photo: West Lake of Hangzhou, shot in May, 2005, during my last semester at ZJU)

20050524

2007/11/21

Miscellaneous about this week

1. Reply to Sunseraphic: Thanks for asking! I am now studying at Imperial Colleg London, as a Master student in Advanced Computing. Finally, I have 'betrayed' ISEE, haha. Well, where are you now? Still at ZJU?

2. Which is the second most widely distributed book ever? Ironically, I got this answer from a little booklet persuading people to believe in Bible. In the booklet, it says that Holy Bible is the most widely distributed book ever, and it is 5 times 'popular' than the second one in this list. However, what is the second one? haha, 'Quotation of Chairman Mao'!

3. I am going to skip a class for the first time in this semester. Sadly, I have chosen a seminar course at the wrong time, which clashes with one of my regular course. Although the regular course is 'more exspensive', I will not be able to maintain my position in Research Stream if I failed to attend this particular seminar course, which makes it overwhelmingly important...

4. Gloomy days are coming. I am not talking about courseworks, hehe. Indeed, I have got plenty of courseworks to do by now, but that is not the reason of feeling 'gloomy'. Actually, it can never become the reason, since they are all interesting tasks to fulfill. It is the weather. Raining all the days, making everywhere wet and deadly cold. Well, I do not particularly hate raining weather, but I still do not think that there is anyone who could bare this kind of weather for weeks with full heart of appreciation. The raining days in winter also made me to recall those of my years spent in Hangzhou, with all sorts of joyful and not yet so pleasent memories. Nevertheless, the good thing is that I have heater this time, which is indeed a vital difference.

5. Dynamics in Human Society. After months of preliminary reading, I think that I have finally came up with some ideas that what this subjects is really about. However, I am not going to expand this topic here. Probably I will write a short article about it when I can find some free time for about half a day, but not now.

6 My first experience of evocuation. The fire alarm rang when I was in the central library yesterday, at about 14:00, and all people were forced to evocuate from Sheffield Building immediately. It is certain by now that it was a false alarm. However, evocuating from the main building and going through all sorts of emergency routines with more than a thousand of students and staffs would be a experience that can not be esily forgotten.

7. Being a 'Computer Scientist' at Imperial, so far, so good.

(Today's photo: busy sky near Heathrew, shot in my room)
DSC02475

2007/11/8

“华为事件”说明了什么

本篇日志用中文写,实在不想让老外知道这种恶心事。

    由于在这边上中国国内的网站普遍慢的不可思议,我现在也只能在新华网等等专门有“对外宣传”考虑的的网站上看看中国国内的消息了。大家都清楚,像新华网这种党的喉舌,除了歌颂祖国和无关痛痒的科技、体育新闻以外,一般是不会有什么很“刺激”的主题的。这就如同你不会指望在新闻联播中看到有关农村教育改革问题(注意,是“问题”)的深度报道一样。不过咱们也不应该把新华网说的就这么差。毕竟人家也是国内的新闻门户站,消息还算比较全面(只是有意思的消息不给你放到首页上就是了),偶尔还是能看到一些看过之后有想说两句话的新闻的。

    今天看到的这条新闻就属于前面说的这么一种。“华为事件”,哈哈,华为可真是有名,什么事都能让它赶上,怪不得同为中国通信行业巨头的大唐和中兴知名度远不及它。这是个什么事呢?说实话我刚看的时候真是不敢相信:它居然要求所有工作超过8年以上的员工主动辞职(注意,是“主动”)!莫非华为看员工们死的太多,决定做点善事关门大吉好让大家多活几年了?原来这事是为了应对两个月之后将要开始实施的新劳动法。根据新劳动法规定,企业必须与在该单位工作10年以上的员工签订“不限制具体期限的劳动合同”,简单的说,就是终身聘用制。当然了,这个终身聘用制并不是说就是铁饭碗,假如你工作干的不好,人家还是照样可以请走你,只是它以后再也不能无缘无故的请你去再就业了。华为就为了规避这一条,居然能做出这么大一件丧心病狂的事情来,实在另我对这个公司的魄力景仰有加啊。当然了,华为也不傻,大家都辞职了,它总不会去做光杆司令吧。到时候再跟你签个新合同,该剥削照样剥削,反正你的“工龄”就算是归零了,这样以后它啥时候看你不爽,等短期合同一过期您就回家去抱孩子吧,管你什么国家法律!管你什么劳动者权利!高,实在是高啊!

    “要求”员工“主动”辞职,这简直是我听说过最具创造性的语言!先不说别的,就只是从语言上看,这也是彻头扯尾的自相矛盾——您这倒底是“要求”呢,还是“主动”呢?这种话要是放在中学生作文里面,早就成全年级反面典型了。但是人家华为就能这么说,这是什么样的精神!(是弱智吗?)其次一点,诸位里面有谁听说过占员工便宜恶心到这么地步的?想要员工辞职,居然还要求你自己提出申请!日本的公司狠,人家裁员起码还说是我开了你,可是咱们华为可真实完美的重现了中国封建文化的最WS的部分:就算是老子逼你,你也得假装是自愿的。便宜要占,好人也要做,这就是咱们牛B的华为。当然了,咱们也得理解华为管理层的难处——只有员工自动辞职“工龄”才能归零,我们也没办法啊。狗急了还跳墙呢,何况是咱们华为呢。什么“为天下人所不齿”之类的事情通通不用管,反正老子全国最牛B,你不齿又能把我怎么着?我华为可是连把员工累死都没事的牛B企业!

    这就是咱们国家通信产业的领头羊吗?这就是咱们国家模范企业能做出来的事吗?

    骂过了,同志们爽了吧?但是如果就这么完了,我岂不是彻底退化到FQ的水平了。于是现在我们来从理性的角度来小小分析一下这个事情说明了什么。

    首先,“华为事件”不能被当作一个孤立的事件来看待,它不仅是华为的事,也是中国的事。第一,华为的管理层也是人,就算他们再有创造力,假如不是受到万万千伟大的“先行者”的启发,大概也不是能够一次就想出这么绝的对策吧。第二,华为可是国内IT业的带头大哥之一,他这么一做,我就不信其它企业中就没有跟风的。有了第一个吃螃蟹的,后来者的理由就充分多了:你看,华为都这么干了,大家都这么干,我有啥办法。其三,我们还必须要考虑到这次事件对于未来的影响——今天我能“要求”员工“自愿”辞职,明天我还有什么干不出来?其四,这件事对社会心理的影响亦不能被忽略:往轻了说,这是蔑视法律,往重了说,这是挑战道德底线。 因此,看“华为事件”,绝不能只看到一个华为,还要看到它的广泛而深远的外部性。由此可见,华为这一票干的,用我老家的口头禅来说,简直“坏麻了”。

    其次,华为为什么能干出这样丧心病狂的事来,这是一个我们不得思考的问题。依我之见,原因有二。

    一,社会道德缺失。就不说是什么原因了,我一直认为中国社会缺少道德观念,因此才会有这么多耸人听闻的事情发生。诚然,道德是软的,法律是硬的,你不能靠道德来完全控制住个人的行为,但是从统计意义上说,道德作为一种普遍的、自动的、平等的约束机制,对整个社会的影响力是显而易见的。

    二,市场不完善。华为为什么敢那么牛B?因为它知道,就算你们恨我,你们还不是照样挤破脑袋想让我剥削。中国的劳动力就算再不值钱,也没那么不值钱吧。那么为什么会这样呢?因为劳动力市场的结构性失调和市场公平保障的不完备(几乎不存在)。中国历来都缺少高级技术人才,现在这个情况并没有得到根本性的改善。人才市场总体供不应求,按理说应该是要素提供者为大才对。但是为什么现在像华为这样的IT企业还能这么牛?答:结构失调。什么叫结构失调?简单的说就是“我要的你没有,我不要的你给的太多”。比如最近几年,高级技工(数控机床操作工之类的)紧缺,浙江省一些企业开出6000/月(05年数据)的高薪都招不来人,而与之相对的,在与计算机/电子/通信领域,即使是如清华、浙大、北航、东南这样的老牌工科名校的毕业生,都开始“零工资就业”了(我不否认“零工资就业”不是价格均衡的结果,但是如果既有高薪,又有经验积累,你想想大家会选哪个?),这就是典型的供需结构失调导致的部门间失衡。结构性失调的直接后果就是各部门间不能共享均衡工资,从而使得在有些产业劳动者是大爷,而在另一些产业(比如华为所在的IT业)企业是大爷。本来应该总体均衡的市场,现在却处处是大爷,这能健康吗?关于上述情况的成因和解决方案,我不想展开讨论了,总之我们可以明确三点:一,这个现象绝对是不好的(好多大爷);二,这个问题是难以解决的(从你选择专业到找工作有四年的时间差,信息超级不完全啊,典型的搏奕问题);三,这个问题是可以改善的(比如赋予大学更多的关于开设专业自主权,从而可以利用看不见的手帮助优化资源分配。

    好,现在再来说说什么叫“市场公平保障的不完备”。两个字,垄断!大家可不要认为只有邮政、铁路这样的有国家撑腰的企业才叫垄断。准确的说,那叫行政垄断,只是垄断的一种。一个行业部门,如果其主要市场份额被少数几家大公司瓜分,那么这个部门就存在垄断。华为自然是典型的垄断企业。大家都知道,垄断是不好的。为什么呢?原因很简单:垄断企业可以控制市场,从而使资源分配无效率。比如华为,全国我最大,你鄙视我吧,你恨我吧,但是我做的再恶心,大家还是得从我这买产品,我想订高价就订高价,你不买就没地方买——这是控制产品市场价格从而获得超额的垄断利润。同时,因为国内的IT业的大企业就那么几家,你鄙视我吧,你恨我吧,但是我做的再恶心,你们还是得到我这讨饭吃,我想怎么剥削你就怎么剥削你,要不你就别吃——这是控制要素市场价格从而实现超低的垄断成本。特别说一下,中国的通信产业是寡头垄断,这个情况要比我刚才说的模型复杂的多,所以市场的真实情况并不完全是那样,也没那么直接(比如华为虽然给的工资比其它通信企业高,但其实仍然是压低后的垄断工资),但是垄断利润和垄断成本确实都是真实存在的。完全竞争市场在无监督的情况下一般会发展成为垄断竞争市场进而是垄断市场,这是大家普遍承认的共识(世界上总会存在Chicago学派的,他们不算“大家”)。市场会失调,所以国家要干预,因此要有反垄断法。但是中国没有(最近好像通过了一个,至少以前确实没有),反而还经常支持垄断(比如在新闻联播里面我们经常可以听到表扬某某垄断企业)。这个情况并不是完全没道理的,垄断企业因为可以自由控制资源分配,因为可以在短期(尤其是行业起步阶段)实现更高的局部效率。但是从长远看,垄断必须被控制,否则明天华为这样的企业可能就要决定取消员工工资了。

(未完,一会继续)

(今天没照片,想不好贴什么,抱歉)

2007/11/6

Permenant Head Damage

Whether to take a PhD is not a question to me. The decision has been made many years ago and it still goes to be my first choice. However, what I can not make clear so far is to where and to do what.
 
Should I continue to study in UK, particularly, at Imperial, or should I take another year to struggle with GRE and then bet my luck on application to an ivy league university in USA? The former path seems to be much easier and more practical, and yet also a very attractive choice. Nevertheless, I still can not convince myself to give up my last chance of fufilling my childhood's dream of studying the most advanced technology in USA.
 
To follow which direction is another vital question. When choosing the major of this MSc, I knew that whatever subject I would follow during this period, I would still have the chance to start it over and take another path in further study. However, PhD is final. To be a scientist, I will not be possible to move far away from what I will choose as my research direction during PhD study. Although there is still a chance, I can not afford the danger of depending my future solely on this tiny probability. As I have said, I have been thinking and discovering for more than a month since my arrival. However, 5 weeks with full of shocks are definitely not enough for me to construct a fullview, even just briefly, of the worldwide contemporary situation in Computer Science studies.
 
This will probably be my most important decision through my entire life. It needs to be analyzid further, and it needs to be decided soon. I never expect a global optimal solution being found, but with luck, I hope I could find a path for me that is optimal by now, and will not turn to be the worst in the future.
 
(Today's photo: London Paddington Station)
DSC02424
2007/11/5

关于维护汉语使用纯洁性的声明

    自即日起,本人所写一切文章,或全部使用中文,或全部使用英文,除少数特殊情况外(见下文)绝不将中文与其它语言混杂使用。今特此声明,请各位监督。
    以下是被允许在中文中混杂使用其它语言(包含但不限于英文)的特殊情况:
    1. 地名、人名,如London、Greenwich、Issac Newton等。
    2. 没有通用翻译或翻译后可能引起歧义的专有名词,如Prolog、Visual C++、Matlab等。
    3. 由于本人能力所限,对其中文翻译的准确性没有把握的词汇或短语,如Horn Clause,Refutation by Resolution等。
    4. 原文引用。
 
(今日照片:午门前,上世纪初的摄影作品,抱歉我忘记作者的名字了)
午门前
2007/11/3

Another weekend

Ah, five weeks have passed, time can be really fast when you got full of things to do. Tired but also rewarding, indeed, rewarding. Even if I have not leant anything academic from the lectures, at least I realized that rigorous theoretical research is also a vivid and fruitful stream in Computer Sience studies.  More importantly, it is really a great pleasure to know those computer scientists who really LOVE what they are doing, with STRONG PASSION that one who have not tasted a bit can hardly imagine. This, is indeed the real essence of academic excellency, for it does not only have so many distinctive researchers, but the aroma of dedication.

By the way, it is also important to note that a half of this semester has gone away. Being a first year student and a last year student in the same time does not only means that I have to receive double amount of trash mail comparing to others. It also implies that I have to start my PhD application process by now, right after I began to get some shallow ideas about what I really want to pursuit in the future. Good luck to myself.

(Plus: Still, I can not understand those business school guys, yet they are also Chinese, who share quite simliar experience with us before coming to here. However, from my perspective, they seem to be in another world, which is entirely different from ours at Imperial... Alas)

DSC02262

 

2007/11/2

About Theroem Prover and Others

1. Why Prolog is not enough:
 
Because Prolog only deals with horn clauses(?!). The definition of horn clause is that it only have at most one positive literal. Remember how we define predicate in Prolog? The only acceptable form in Prolog is that such like 'A :- B,C,D.'. It means "A if B and C and D", or written in logic form as B & C & D -> A. Therefore, if we rearrange it into clause form (A clause means a first order logic formula that is constructed as a disjunction of literals, say, atoms or negated atoms), it would become !B | !C | !D | A. See, only the head of the clause (A) remains to be positive, all others are negative now. The goal input by user is negated to form another clause (also horn clausem, note when negated, existential quantifier in the goal became univeral, as it shold be in a clause). Then, Prolog resolves on this set of horn clauses, using linear searching strategy, namely, always resolve with the last resolved clause (certainly it would start from the negeted goal), to derive the empty clause and hence prove that the goal can be deduted from the evidence. Otherwise, Prolog returns 'no' to say that it can not find a proof. The soundness and completeness of this mechanism can be prooved. However, in general, to be a really complete theorem prover in the sense of Automated Reasoning, it should have the ability of dealing with all kinds of clauses. Prolog failed to achieve this goal. Thus, other methods are needed, which actually form the most part of the content of this course.

2. Other issues about the nature of theorem prover will be added in separate articles, cause I am still working on it, so I decide to further discuss it later, when I could have clearer ideas in mind.
 
3. Sorry for disappointing DuanDuan, I forgot what did I wanted to say yesterday... :(
 
(Today's photo, 19th century factory steam engine shown in Science Museum)
DSC02309
2007/11/1

Automated Reasonong

To understand what is 'automated reasonong', we have to first understand the exact meaning of the term 'reasoning'.
 
What is 'reasoning'? You could consult a dictionary to find out the standard meaning of the English word 'reasoning'. However, as many other words in the terminology of Computer Science, the concept held by 'reasoning' is much narrower in our context than that in daily life usage. Suppose we have a set of propositions (evidence), which are known to be true. Then, the process of 'reasoning' means to derive conclusions (aslo propositions) from the given evidence, based on the law of Predicate Logic (also known as First Order Logic). The following sentences serve as an example. We know: (1)all universities in London do not have large campus, (2)Imperial College is located in London. Then, we can conclude: (1)Imperical College is small in size. This example can also be written in logical form: ((In_London(X) --> Small_In_Size(X)) AND (In_London(Imperial_College)) |= Small_In_Size(Imperial_College).
 
However, it can be easily imagined that when we have a huge set of evidence, the number of protential conclusions would also become very large. Thus, at most cases, we are more concerned about the question that whether a certain conclusion could be fairly made from a certain set of true propositions, rather than what we could conclude in general from the given evidence, cause there are often too many of them. This is also the main theme through this course. In other words, in this course, 'reasoning' means to find out whether a certain proposition would stand when a set of other propositions hold.
 
After above description, it would be easy to see what does 'automated reasoning' really means: rigid and algorithmic methods of solving the problem regarding to logically proving a proposition from given evidence, both in Predicate Logic form. In other words, it is all about the mechanism of computer based theorem prover.
 
cas-prolog