關(guān)于GRE、GPA、TOEFL的重要性

I have a Chinese friend who is going to apply in USA, he is studying CS.

As an experienced professor, would you help me to help him answer several questions about how you pick your gradate students from China?

For example, when you pick your students, which capability is most important thing that your are favor of? Do you think GPA is important? Do you think GRE and TOEFL scores are important? Do you think the rank of the Chinese University where th student is from is important? Do you think you'd rather pick students who has several years of working experiences or fresh granduated students?

Thanks you so much for your help! I really appreciate your time!

Thanks for asking my opinion.

GPA (as long as is in upper rank), GRE and TOEFL are not important. They just measured the abilities of students but not necessary their capabilities.

BTW, Since GRE score are so high for most Chinese apllicants, most of us just ignore them.

For our applied research programs like ours, work experience is definitelt an advantage. Critical thinking and independence are traits for successful PhD students.

For CS, I believe critical thinking is very important since many of their reseach work is abstract.

Best Regards.

UT-Austin中國教授關(guān)于PhD申請(qǐng)的訪談

Q：根據(jù)您在 UT-Austin 和 UK 的經(jīng)歷，PhD 申請(qǐng)中最重要的因素有哪些？如何成功申P(guān)hD？ 美國 PhD 委員會(huì)處理申請(qǐng)過程是什么？

A：最主要的是需要讓教授知道你，并且知道你的優(yōu)點(diǎn)，對(duì)他/她有價(jià)值。成功的因素主要有三個(gè)，任意一個(gè)即可：

• 高GPA
• 不錯(cuò)的研究成果
• 教授的熟人強(qiáng)烈推薦

處理申請(qǐng)主要有兩種方法，在非常頂尖的學(xué)校，比如CMU和UT，所有的教授都會(huì)參與對(duì)申請(qǐng)者的打分，只有大家都認(rèn)為不錯(cuò)的人才能被錄取。另一種是完全由教授決定，但前提是教授手里要有資金。但是不過哪種方法，申請(qǐng)之前去套磁都是必要的。

Q：導(dǎo)師的選擇有沒有注意點(diǎn)，教授或者助理教授、亞裔或者白人？

A：教授的水平和人品跟他的種族以及級(jí)別基本沒有關(guān)系。很多國內(nèi)的說法流行說白人的資深教授好，但是我見過無數(shù)坑爹的白人資深教授。反而是不少中國的年輕教授比較能對(duì)中國學(xué)生因材施教，也了解他們的想法。比如仲盛老師和劉向陽老師，他們?cè)诿绹臅r(shí)候都是華人ap，但是培養(yǎng)學(xué)生都非常的出色，前程也很好。多了解一個(gè)教授是很重要的，可以通過各種途徑，比如找學(xué)長(zhǎng)問。

Q：出國做科研需要英語很好、GRE托福很高？

A：錯(cuò)，英語可能是國外的教授最不看重的一個(gè)能力。我在香港的導(dǎo)師倪明選教授（現(xiàn)澳門大學(xué)副校長(zhǎng)）說過：“英語可以提高，人品沒法提高。”我的博士導(dǎo)師林善成教授（美國工程院院士）也說他從來不看GRE的成績(jī)。

Q：出國做科研只有在頂尖學(xué)校才會(huì)做好？

A：錯(cuò)，現(xiàn)在CS領(lǐng)域大把大把的一流科研出現(xiàn)在美國50名之后甚至100名之后的學(xué)校。只要努力，誰都能做出一流的科研。

對(duì)于喜愛科研的同學(xué)來說，出國留學(xué)遠(yuǎn)比考研容易且性價(jià)比高。我的第一個(gè)博士生便是考研落榜生，花了一年準(zhǔn)備考研沒上，又花了一個(gè)月便考完托福GRE來了我這里。在我組里還沒到兩年就發(fā)表了兩篇一流會(huì)議的文章，另外有兩篇在投，名聲在外被IBM研究院和愛立信研究院邀請(qǐng)去暑假實(shí)習(xí)。我的另一個(gè)博士生，在原來學(xué)校讀研的時(shí)候不順利自己退學(xué)，到了我組里兩年不到也發(fā)表了兩篇一流會(huì)議文章。出去開會(huì)的時(shí)候，很多同行羨慕我找到這么出色的學(xué)生，我其實(shí)想說如果他們國內(nèi)考研考上了或者讀研順利，那我就沒戲了。不知道沒考上是幸運(yùn)還是不幸。

Q：在 UT-Austin 讀 CS PhD 是一種怎樣的體驗(yàn)？

A：沒啥特殊體驗(yàn)，就是每天來辦公室做幾個(gè)小時(shí)。大家不要把phd和做研究想得太神秘，

我前兩年隔壁辦公室的圖靈獎(jiǎng)得主Emerson教授就是一個(gè)普通老爺爺，每天都是迷迷糊糊的，拖著布鞋，拿著可樂。

Q：成功取得 PhD 學(xué)位的關(guān)鍵因素是什么 （譬如能力、熱愛、堅(jiān)持）？

A：基本的能力其實(shí)大家都具備。就是堅(jiān)持最重要了。

Q：計(jì)算機(jī)網(wǎng)絡(luò)和分布式系統(tǒng)等研究方向的 PhD畢業(yè)以后在學(xué)術(shù)界和工業(yè)界就業(yè)去向如何

？

A：網(wǎng)絡(luò)在學(xué)術(shù)界的就業(yè)不是最熱門，但也還可以。在工業(yè)界一直是就業(yè)熱門，網(wǎng)絡(luò)的phd從來都是所向披靡，offer拿到手軟。分布式系統(tǒng)在學(xué)術(shù)界和工業(yè)界都是熱點(diǎn)，尤其結(jié)合了大數(shù)據(jù)以后。

Q：PhD 相比于 MS 畢業(yè)工作三四年來說，有哪些優(yōu)勢(shì)？

A：對(duì)于工業(yè)界的工作來說，phd最大的優(yōu)勢(shì)是有完備的發(fā)現(xiàn)問題-解決問題-報(bào)告成果這一系列的訓(xùn)練，如果單純看工資，可能phd和ms畢業(yè)三四年差不多，但是phd可以擔(dān)當(dāng)公司的研究員，而ms大部分是程序員，能接觸核心技術(shù)的非phd，你在公司需要作出的努力也早就夠你拿個(gè)phd了。另外像Google這樣phd建立的公司，非常喜歡phd。

賓大小秘關(guān)于GT權(quán)重的回復(fù)

詢問信：

Dear Sir,

I have ever heard that the secretary of graduate school will first shift some application materials according to applicants\' GRE, TOEFL scores and GPA. Then the graduate committee make decisions according to the shifted materials. Is that true for seas@uUpenn? My GRE score is not competitive, V: 150, Q: 170, AW:3. But I have some publications publicated on important and famous international journals.

I have taken TOELF exam, and I am not sure whether my TOEFL is above 100 or not.

But seas@Upenn is my dream school. How I wish I could persue the doctoral program in seas@Upenn. Would you please tell me if my TOEFL is not above 100, do I still get chance to be addmitted by seas@Upenn?

Sincerely.

小秘回復(fù)：

Your description of the admissions process is roughly accurate, althoughit better describes the MSE process.

It is _possible_ to gain admission even with a low TOEFL, GRE-Verbal, etc (the GRE-Verbal isn't terribly important) - none of these are immediate summary disqualifiers. The most important things are academic/research ability, trancripts, the schools you've attended – of course papers published in international journals are very significant.

同校學(xué)生比較

上次和一個(gè)Professor聊天，正好聊到申請(qǐng)了。當(dāng)時(shí)他正笑呵呵的警告我以后不許和他的那個(gè)清華的學(xué)生講中文，然后就聊到GRE了。

他說中國學(xué)生的GRE太高了，他反正不管。我馬上問他Care什么，他說一般他都先讓RICHARD（我們系的男小秘）把幾個(gè)學(xué)校的學(xué)生檔案拿過來，我馬上說大概能看出區(qū)別嗎？他說，能呀，很多都是一個(gè)系的，直接比較分?jǐn)?shù)就可以了。

UMich教授：GPA低，教授都不好意思提名

我現(xiàn)在的導(dǎo)師正好剛在Umich做過一年多的訪問學(xué)者，而且就是在那里ME系里的Graduate Chair手下做的，又都是中國人所以和他很熟。所以我申請(qǐng)的時(shí)候就把Umich做為了我的一個(gè)重點(diǎn)學(xué)校準(zhǔn)備：申請(qǐng)材料精心填寫、交申請(qǐng)費(fèi)、給TG正式成績(jī)。

材料寄出后又請(qǐng)老板親自給對(duì)方寫了一封Email拜托對(duì)方多多關(guān)照。Email發(fā)出后一直沒有回音，好心的老板又寫了一封也一直沒有回音。讓我很是沮喪，老板也不好意思再問了。就這樣在日日夜夜的企盼中等來了Umich的據(jù)信，頓時(shí)很受打擊，一看信的落款就是那個(gè)家伙。心里想這個(gè)人真是不夠意思，太望舊了。事情就這么過去了。五月份我老板又要去美國訪問半個(gè)多月，而且要途徑Umich，于是我跟老板說了一下，要他碰見那個(gè)人的時(shí)候順便幫我問一下我被據(jù)的原因。收到了這么多的據(jù)信，沒有一封信寫得比較實(shí)在，都是一些套話，根本就沒有告訴你被據(jù)的具體原因。老板也欣然答應(yīng)了。

今天老板從美國回來了，我碰見他的第一件事情就是問他幫我詢問的結(jié)果。老板說：我在Umich見了他兩面，也談到你的問題。我給他寫的信他都看到了于是特別留意了一下你的材料，可是發(fā)現(xiàn)你的GPA特別低，而且Umich的競(jìng)爭(zhēng)也特別激烈，不太好意思給committee提名。如果你的GPA再高一點(diǎn)的話，提名然后給找個(gè)獎(jiǎng)什么的并不太難。所以一直沒好意思給我回信。

就這么我的一個(gè)最大的希望就破滅了?，F(xiàn)在回想起來人家說得話也是實(shí)話，大三、大四的時(shí)候忙著考G，一直沒怎么好好上過課，作業(yè)都是copy的，考試的時(shí)候隨便混過去的，都是系里的老師也沒有為難，給的分?jǐn)?shù)都及格了。所以算GPA的時(shí)候低得讓我吃驚，算完后告訴了當(dāng)時(shí)飛躍的版主Levis尋求一些安慰，又被他給打擊了一下，說確實(shí)低了一些。估計(jì)其他的學(xué)校據(jù)的最主要的原因也是因?yàn)檫@個(gè)問題。當(dāng)時(shí)沒有認(rèn)識(shí)到GPA的重要性，現(xiàn)在后悔也晚了。

研究潛力最為重要

I think the most important thing during your application process is to show your research POTENTIAL as much as possible. Your T&G scores needn't to be very high. Neither is your GPA. In general, the scores are not so important for the first class schools than for the less famous schools. However, you do need to get a minimum score, eg. T≥100 G≥300 and the GPA is not too low.

You should contact with your interested professor as early as possible to get your application results as soon as possible. As for the application fees, you can neglect it completely. If a department really want to recruit you, it will pay for you in some cases.

耶魯教授看美國研究生招生

角色的不同造成相異的視野與角度，所謂“橫看成嶺側(cè)成峰，遠(yuǎn)近高低各不同”，同樣的峰　山石，竟有不同的面貌。我想從一個(gè)年輕教師的角度，報(bào)告一下教授對(duì)研究生如何評(píng)鑒，有怎樣期許。

我任教的耶魯大學(xué)地質(zhì)與地球物理學(xué)系每年收到一百到一百八十份入學(xué)申請(qǐng)，從中選取十至十五人發(fā)出入學(xué)許可。我們審查委員會(huì)首先根據(jù)學(xué)生的學(xué)業(yè)成績(jī)、修習(xí)科目、推薦信、個(gè)人陳述、GRE、TOEFL 分?jǐn)?shù)、研究成果（學(xué)、碩士論文）等將學(xué)生分成上、中、下三等，提報(bào)教授會(huì)議，教授們就個(gè)人的專業(yè)范疇對(duì)個(gè)別的學(xué)生可以更改評(píng)鑒的級(jí)別或從中等“解救”到上等，或從上等降級(jí)；然後再把上等生根據(jù)其研究興趣分到本系的四個(gè)學(xué)術(shù)分科里去，再逐一討論，評(píng)定名次，各組的前四人列入“夠格”的候選名單，這是第一階段。

這第一階段的選拔根據(jù)一些比較硬性的標(biāo)準(zhǔn)，譬如說外籍學(xué)生 TOEFL 在100分以下不予考慮、基本學(xué)科訓(xùn)練不足者剔除、中國學(xué)生GRE數(shù)學(xué)在95%以下，閱讀145分以下出局。

第二階段則考慮到一些實(shí)際性的因素，譬如說甲教授指導(dǎo)學(xué)生人數(shù)已經(jīng)太多，不堪重負(fù)，即使有很好的學(xué)生也只好割愛；又如TA的搭配，能說流利英語的學(xué)生受到重點(diǎn)照顧，例如，乙教授有重大研究計(jì)劃，經(jīng)費(fèi)設(shè)備一應(yīng)俱全，只欠學(xué)生，則有這方面研究經(jīng)驗(yàn)或興趣的學(xué)生可得到賞識(shí)。因此第二階段中，“人治”的因素大增，個(gè)別教授的舉薦與偏愛往往能使排名在后的優(yōu)等生破格中選。

總結(jié)來說，基本的成績(jī)首先要夠一定水準(zhǔn)，然後個(gè)別的研究興趣與經(jīng)驗(yàn)，及教授的研究經(jīng)費(fèi)及學(xué)生結(jié)構(gòu)等綜合的考慮，則成為最后決定因素。

一般而言，有研究經(jīng)驗(yàn)及文章發(fā)表的學(xué)生往往能在最後一關(guān)脫穎而出。我們收的中國籍（大陸、臺(tái)灣）學(xué)生通常在國內(nèi)多半已有碩士學(xué)位或已是研究生，多數(shù)并有文章發(fā)表，若是本科畢業(yè)生則必然是全A學(xué)生，TOEFL和GRE成績(jī)領(lǐng)袖群倫者。中國學(xué)生的推薦信往往千篇一律，好話連篇，而寫信的教授素不相識(shí)，因此推薦信的推薦份量往往不如美國學(xué)生。此時(shí)，學(xué)生本人的個(gè)人陳述變得非常重要，成為我們認(rèn)識(shí)學(xué)生的智慧、文筆、情操、興趣、動(dòng)機(jī)的主要根據(jù)。

我們認(rèn)為，事在人為。聰明能干、自動(dòng)自發(fā)的學(xué)生是我們的最愛。一方面，這樣的學(xué)生能自已開創(chuàng)，創(chuàng)造形勢(shì)，化不可能為可能，成為有原創(chuàng)力的研究伙伴，自行整合校園乃至校外的研究資源，開發(fā)研究經(jīng)費(fèi)及課題；另方面，這樣的學(xué)生能青出于藍(lán)而勝于藍(lán)；在學(xué)期間老師有教學(xué)相長(zhǎng)的收獲；畢業(yè)以后，光照學(xué)界，老師亦與有榮焉。

伯克利教授談GT、Paper及套磁的作用

Dear Micheal,

I read your statement again during the weekend. It is clearly that there is a cultural difference here. In the US, the criteria to admit a student primarily rests on merits (connection may count in some private elitist schools where that the parents being an alumni will or a big contributor add points). Especially in public university, the admission process is very transparent and conducted in a committee. When I was at Berkeley, I ran for one of the two student representatives to the doctoral admission committee.I read admission files with my other 5 faculty members. During the three rounds of discussion, I had one vote just like each of my professors. We admitted the students purely based on merits. There was a separately financial support committee composing with faculty and students who do not need financial aids (for fear of conflict of interests).

We developed a table by GRE, TOFEL, SAT and GPA. It is easy to screen off about 70% of the applicants who were obviously unqualified -- low scores, lack of supporting documents, incomplete forms, major errors, missing deadlines, etc. Then we divided the remaining 30% into three groups -- must accept (students we love to get when they may go to other better schools), should accept (those who are good enough), and the marginal case (usually we put them on waiting list). Doctoral admission also relies on another important factor -- whether there is a professor in the department who has common research interest to take care of the applicant for the next 5 years. This is not very important for a master applicant because most master programs are fairly standardized.

So, with a reasonably good TOFEL and GRE score, how can you market yourself better? Your academic training, your academic interest, your future career or scholastic plan, and your dedication are the stuff the admission committee will examine. Letters of reference are very important but you don't have that advantage because unless you are applying to a school where your professor at *** graduated (or your professor is known in the field worldwide); otherwise, these letters from China (sometimes with translation) will not help too much. Of course, you must make sure none of these letters hurt you. Then what makes you different from other applicants and deserving a fellowship?

Some work sample, prior studies, project reports will give the committee a chance to gauge your level of competence. So only send stuff that is good -- it is tough because of the translation requirement.

Life is full of constraints and opportunities. So you try your best and hope of the best. When opportunities present to you, seize them. Your score is good enough to get to most schools (don't worry about the verbal part of the GRE -- which you are very good as a non-English speaker and your TOFEL is OK because it is above 100. You will be accepted by some schools but then what?

You have several options:

• Get to the cheapest and most affordable place that accepts you -- it may be in a remote, isolated place. But if you are lucky, you study a subject you like in a program that is solid.
• Get to a good university but financially you are bit vulnerable -- you may be in the field you like or not.

Here good university means the fame of the whole university -- not the standardof the department. You should do some research about the reputation by department in the fields you applied for. The fame of the whole university is more relevant for undergraduates. For graduate

school, the department's reputation counts more because of the network of professors in that field. For example, computer science at Penn is lousy but its MBA is the best in the world (better than Harvard and Yale). Also my previous university FIU is lousy but its hopitality/sea cruise program is one of the best two in the country (the other one in Las Vegas). It is about the niche.

All this related back to your goals. It seems to me that you just want to study in the US regardless interest. If you show such desperation in your application, it may raise a red flag. Committees want to accept good students in terms of academic potential. So try your best to hide that motivation. Since it is a practical concern to you, you may like to go for option a. as defined above.

Use the 2 years as a time to get adjusted, obtain a very high GPA, complete one or two excellent term papers. You should know the niche of graduate schools in your field, so you may consider transfer in the second year, or get another master degrees or a doctoral. Also develop your connection and identify your true academic interest in those two years. Make sure that when you applied for Ph.D. or for a job, you have 2 to 3 American professors who are fully supportive.

Given your ability, you would like the US because this is a place that appreciates handwork and individual merit. It is a country full of opportunities. Guanxi helps but not a factor that limits you. However, you must be very independent and always lonely in mastering your future and taking your responsibilities.

Another matter is about the long-term career future. Computer, IT & biotech will be the future of the world just like electricity, telephone, telegram around the turn of the 20th century. Despite the setback of the high-tech industries (dot.com, etc.) in the past 10 months, these fields have lot of future.

However what I saw in the past decade was that most Asian students came and stayed in engineering and then computer; then make a good living. Few of them were entrepreneurial enough to have their own companies. Most of them lived in an isolated social and cultural setting that they did not fully integrate into the US mainstream. The main reason is that they lack the language skill, so they could not get to marketing or key positions which \ require bargaining, negotiation, and management skill -- all in English.

It is clear that when they were studying here, they did not use English. They found a place to excel in their technical and mathematical expressions. This specialization hurts them in the long-term, so some of them returned to Taiwan and Korea. Some try to work in import and export trade between Asian and the US. Some worked for Dell, I BM, Microsoft but worked in China. I just let you know you can choose if you prepare yourselves early enough. It is up to you to pick the best path for you.You will be on your own most time but I am happy to discuss with you if I have less time pressure.

Take care and good luck!

【注：作者據(jù)稱系一華人教授，有人說是伯克利的，但也有將該信標(biāo)為賓夕法尼亞大學(xué)的，實(shí)情不詳，有偽造嫌疑，但其觀點(diǎn)依然有一定的參考意義】

UMN物理教授答學(xué)生問

明尼蘇達(dá)大學(xué)雙城分校物理系教授Yushi面試中答學(xué)生問：

每個(gè)教授的評(píng)審方法不同，有的非?？粗谿PA，有的看重Paper，有的看中GRE，不過Yushi的標(biāo)準(zhǔn)是Paper不重要，因?yàn)樗惶湃芜@邊的Paper含金量，不過有總歸是好的，GRE也沒用。因?yàn)樗f這個(gè)不能反映什么東西，是最后考慮的問題。所以他很無奈的選擇GPA作為標(biāo)準(zhǔn)，至于什么推薦信之類的也就根本沒什么作用了，除非有什么特殊的關(guān)系之類。

我問他你面了這么多學(xué)生，那個(gè)學(xué)校的學(xué)生你最感興趣，他說他在北京最感興趣的是一個(gè)南開的學(xué)生，然后說科大每年多有海量的申請(qǐng)者，而且會(huì)錄取很多，不過他最近幾年陸續(xù)招了很多復(fù)旦南大的學(xué)生感覺質(zhì)量很不錯(cuò)，所以他們?cè)诳紤]給復(fù)旦南大更多的機(jī)會(huì)，今后幾年錄取這兩個(gè)學(xué)校的學(xué)生會(huì)增加的。

透露出的信息點(diǎn)：1）GRE不重要；2）國內(nèi)學(xué)校很重要

斯坦福教授：Paper未必很重要

其實(shí)也算不上面試，正好套的教授來中國，他在committee，就順便見面聊了半個(gè)小時(shí)，我申的就是他的組（斯坦福的）。他告訴我，斯坦福的committee把每個(gè)申請(qǐng)者的GPA、GT和推薦信按照某種加權(quán)的算法rank所有的申請(qǐng)者，（他們沒有cutoff，即使你的GT沒有達(dá)到minimum)

對(duì)中國學(xué)生，一般只錄取前5%美國學(xué)生大概50%+。這里面的一個(gè)問題是如果你的research非常出色，有很牛的paper，如果不在推薦信中強(qiáng)烈體現(xiàn)的話，基本上作用不大。

而大家網(wǎng)申填的publications和寄過去的CV，基本上committee不看。一般Research所起的作用是：在兩個(gè)申請(qǐng)者的ranking不相上下的時(shí)候，如果你在Research上很有建樹，那么就是你了！

而且由于assistantship基本上是系里的經(jīng)費(fèi)，所以委員會(huì)決定的事情。教授想更改非常困難，即使你們之間非常鐵，他也很難幫上忙。而且由于rank是量化的，很難出現(xiàn)某個(gè)很有權(quán)威的教授，說想要誰，委員會(huì)就會(huì)妥協(xié)的情形。

聽完真的想當(dāng)郁悶,大家平時(shí)都喊著牛paper是多么的重要，research多么重要，對(duì)于這種committee的牛校，如果GPA和GT不highlight,基本仍然等于沒戲。即使你又再多的paper，再牛的paper（當(dāng)然牛到Science/Nature量級(jí)的也許會(huì)有質(zhì)變）

最后Professor只能和我說“I can help you, but the help needs a reasonable ranking.”

當(dāng)時(shí)犧牲了那么多的時(shí)間，搞科研，出paper，甚至犧牲了一些課程的Grade，現(xiàn)在看來也許未必是對(duì)的選擇啊?！驹搶W(xué)生系清華大學(xué)學(xué)生】

MIT & CMU CS教授談?wù)猩?/strong>

去年曾經(jīng)聽過MIT的CSAIL的co-director Prof. Zue還有CMU的計(jì)算機(jī)系主任Prof. Wing的講座。他們都講了他們是如何選擇申請(qǐng)者的。當(dāng)時(shí)我各寫了一篇blog，現(xiàn)在我把里面關(guān)于申請(qǐng)的這一段貼出來吧，看看最牛的學(xué)校是怎么選擇申請(qǐng)者的。

Prof. Zue:

我們也很難挑選這些申請(qǐng)者——想來也是，從1600人里挑40人，敢來申請(qǐng)的有多少不是牛人的。我們主要看的是三項(xiàng)：

1. 成績(jī)，但是這是你在申請(qǐng)時(shí)已經(jīng)不能改變的；

2. 推薦信，我們要求寫推薦信的人是了解你的老師，而且不是寫“該同學(xué)在該科表現(xiàn)出色，獲得95分”之類，這些都是在成績(jī)單上可以看出來的，我們要知道的是你有什么與眾不同的表現(xiàn)；

3. 也是最重要的，Statement of Objective（也即一般所講的Personal Statement），這一段老教授講得非常有趣，把能想起來的寫一下：

1) If you say, I was dreaming of going to MIT when I was very young...扔到垃圾桶里! Everyone is dreaming of going to MIT!

2) And if you say, MIT is the best university of the world, it has the 1st ranking CS department...扔到垃圾桶里，這些我們都知道...

3) But if you say, 我對(duì)你們的某個(gè)專業(yè)或者某個(gè)方向感興趣...um? 可以看一下…

如果是說，我對(duì)某教授研究的課題比較有興趣，希望能參與它的研究，it's better!

4) And if you say, I'm interested in ..., but I have a better idea. That's great!

老教授要求我們寫PS時(shí)重點(diǎn)考慮三個(gè)問題：

1. Why do you come here?

2. Who do you want to study with?

3. What do you want to do?

從始至終，他都在強(qiáng)調(diào)一句話: Make yourself different. 這才是申請(qǐng)的最關(guān)鍵之處。

Prof. Wing:

The depardment of CS of CMU is very selective, We only adimit 177 students this year from over 1000 applications... But we are looking fro very bright students from China, because you received very good training in math...And our requirement is:

1. Good grade

2. Good recommendation

3. the most important is your research potential.

That means you must have strong ability to do research. So you'd better get involved to project, have creativity and communacation skill, which means skill to speak and write. And as the most popular language in CS field is English, so your English must be good enough....Please work hard and apply for us and have good luck!

為什么要面試？

昨天問了一個(gè)米國老師為什么要面試，他說I think they want to know how good your English is.  They may also ask you questions on your career goal, and see how much you know about XXX school.

伯克利CS招生官談CS四大名校錄取

前幾個(gè)星期，我去參加了系給本科生舉辦的一個(gè)Grad school talk，邀請(qǐng)了本校的CS Grad admission 的負(fù)責(zé)人和斯坦福的EE grad admission的負(fù)責(zé)人。其中，伯克利的負(fù)責(zé)人展示了一個(gè)圖表，里面顯示了去年斯坦福、CMU、MIT、伯克利 4所大學(xué)的計(jì)算機(jī)科學(xué)PhD的錄取情況。很可惜，我記憶力不太好，所以具體的數(shù)字記不太清楚了，不過以下是幾點(diǎn)我記得的東西：

1. 4所大學(xué)加在一起招收了260多個(gè)人

2. 伯克利招收了80-90多個(gè)PhD學(xué)生

3. 有10名左右的學(xué)生，同時(shí)被3或4所大學(xué)錄取

4. 有30名左右的學(xué)生，同時(shí)被斯坦福和伯克利錄取

5. MIT特別喜歡搞后門（直升）

6. 伯克利一共受到了大約1.1-1.3k份application

7. 斯坦福的PhD錄取不是按照Field來分的，也沒有任何Field的Quota

8. 伯克利的PhD是按照field來分的，因此申請(qǐng)AI的同學(xué)將和別的想搞AI的同學(xué)競(jìng)爭(zhēng)幾個(gè)位置

9. GRE太高沒啥用

10. 牛推大于一切

11. 伯克利的負(fù)責(zé)人建議：大二開始做研究

12. 大于3.7-3.8的gpa是基本的

13. 從1,2,3,4點(diǎn)可以看出，每個(gè)學(xué)校選擇的時(shí)候口味是有一定的區(qū)別的，因此多報(bào)的話，很有利于提高機(jī)會(huì)

14. 任課老師的推薦信基本不會(huì)有任何幫助

15. 你去業(yè)界作研究后，讓你負(fù)責(zé)人寫的推薦信很大程度上取決于你負(fù)責(zé)人的知名度和學(xué)歷

16. 如果你拿到Fellowship的話，一定要通知學(xué)校，會(huì)對(duì)你的錄取有一定的幫助

17. 在CS PhD出路方面，貌似很大部分最后是去業(yè)界帶研究team

18. 在被錄取后的選校方面，他們推薦以你想要研究的Field的導(dǎo)師來選擇學(xué)校，而不是以名聲

19. Admission的結(jié)果很大程度上取決于你所什么方面的教授對(duì)你的看法，因此....

補(bǔ)充幾點(diǎn)在陸奇講座上聽到的東西

陸奇是代表微軟來作每個(gè)學(xué)期的 microsoft executive talk，他當(dāng)時(shí)一件汗衫，看起來感覺和學(xué)生差不多，完全不像李開復(fù)那種一看就是高管的樣子。

這些是陸奇對(duì)在一名（也許是伯克利的）CS PhD留學(xué)生的提問之后作出的回答。他的主題是：在當(dāng)今社會(huì)，中國留美CS PhD越來越多的情況下，怎么才能夠脫穎而出，做到像他那樣的成功。

1. 機(jī)會(huì)就像公車，錯(cuò)過一班永遠(yuǎn)有下一班，但不要以為一班車就是把你從你現(xiàn)在的位置帶到你自己所期望的位置。

2. Make yourself uncomfortable，如果你的日子很舒服輕松了，那他覺得move on的時(shí)間到了，如果你的日子不舒服，那他覺得是一件很好的事情，因?yàn)槟阍谶M(jìn)步在成長(zhǎng)。

3. If you are smart and you work hard, you can achieve anything. (他反復(fù)強(qiáng)調(diào)的這點(diǎn))

課外活動(dòng)、高中成績(jī)及各種計(jì)算機(jī)證書

International applicants seeking assistantship/admission: please do not send me emails blindly without reading this.

Every day I receive emails from international candidates for information, for advice, and especially requesting financial assistance. Unfortunately, I delete most of them without even reading them carefully. I get so many irrelevant emails that cannot reply to each. Extremely rarely has one caught my attention. Please do not send me emails without reading this first. I am only interested in corresponding to candidates who are exceptional, and have a strong interest and experience in my immediate area of specialization, and I highly prefer ones with a Master's degree. 

For your benefit, I have included in this page, some writing tips, when corresponding to me, and perhaps also in writing to other faculty members in US Universities. When corresponding, do not mention your age/date of birth, your gender, nationality etc. These are superfluous information. 

I am NOT interested in: Extra-curricular activities, performance in high school or before, TOEFL scores, hobbies, software skills, UNIX/ORACLE/NOVELL certification and other items unrelated to my research area*. 

I am SOMEWHAT interested in: Your awards/academic honors, academic records, GPA, GRE score (2200+), experience. 

I am HIGHLY interested in: Your publications, research interests (have supporting evidence for your claim, don't just claim you are interested, unless you actually have done some work in the area). 

 (Incidentally, I do not know what NOVELL or Microsoft certification are. Research is more concerned with things like advanced mathematical skills, creativity etc. Too much software skill actually may even count negative in my group). 

藤校對(duì)中國大學(xué)的偏見

Boss剛好認(rèn)識(shí)某藤校 CS Master Program的系主任，幫我強(qiáng)推了。系主任說他們有bias towards Universities in China，然后說會(huì)look out for his application.

GRE Sub只是浮云

Although this is specified in the main documents, we do overlook this requirement if we find a good student. So please do apply, and then lets keep our fingers crossed. Let me know what you have decided.

GT在初篩中很重要

Letters of recommendation all tend to say that so and so is the best student, etc. So, not much differentiation exists there.

GRE and TOEFL are heavily counted at the beginning of the screening process. Once you pass that phase, the scores are not emphasized all that much.

對(duì)清北的學(xué)生而言，不要太在意GT

Well, since you want an honest evaluation, I'll tell you what I can. What we've found in the past is that students from China generally put so much effort into their GRE and TOEFL, that the scores mean nothing.

We've admitted students who got perfect GRE scores and high TOEFLs who didn't turn out to be all that bright or to speak English all that well. Also, what we've seen are letters of recommendation that were clearly written by the students and only signed by the professors. Since we're just reading the student's words, we tend to focus on the statement of purpose (where you describe your skills and what you want to do), and on the grade transcripts.

I've read many statements of purpose that are two pages long and say nothing. If somebody can actually describe their skills and the work they want to do, that stands out. Likewise, we'll look at grades from all the students applying from a particular school.

Usually, you can find one or two classes that have professors who grade more harshly. If somebody gets a good grade from a hard professor, that stands out as well.

It's obviously easier to consider students from schools like Tsinghua because we get enough applications from there that you can really compare the students to each other. Coming from a less famous school certainly hurts your chances, but it doesn't ruin them.

That's what I look for when I consider students from China. So, I hope you understand that nothing that you've sent me allows me to tell you how well you might do if you applied here. 

I would need to see your whole application.

Thanks,

興趣最重要

羅格斯大學(xué)的教授，他就和我說了我非常受用的3句話——套辭是必須的，沒有什么所謂的核心技術(shù)，你對(duì)這個(gè)研究的興趣超越一切背景。

注：上述觀點(diǎn)，系某背景極弱學(xué)生在計(jì)劃申美國PhD，近乎絕望的情況下，咨詢自己羅格斯大學(xué)的一位教授（該同學(xué)親戚）時(shí)，該教授對(duì)該同學(xué)說的一句話。

該同學(xué)背景：本科gpa2.3，碩士gpa3.7（但是學(xué)分很少，課程也很少），gre311，toefl100，無文章。推薦信一個(gè)普通教授（本科畢設(shè)導(dǎo)師），經(jīng)理（畢業(yè)實(shí)習(xí)經(jīng)理），副教授（碩士專業(yè)課和我交往不錯(cuò)的一個(gè)老師）。2012年畢業(yè)，2013Fall找啟德遭遇全拒，2014Fall  DIY依靠套磁申到美國TOP50學(xué)校全獎(jiǎng)PhD，生物技術(shù)專業(yè)。

該同學(xué)自述：“在2013年11月，在我遞交完第一波申請(qǐng)材料時(shí)候，我心情是很絕望的。連我自己都想不出一個(gè)理由讓別人要我。我的ps里面充滿了套話，空話，我如果是教授，除非這個(gè)人的背景很棒，否則我不會(huì)要他的。”

奧巴馬的推薦信足夠讓你上哈佛嗎？

Quora上的一個(gè)提問，一位招生官的回答：

我認(rèn)為可能性不大。我在一所大學(xué)的商學(xué)院招生委員會(huì)工作，雖然不是哈佛，但也是非常頂尖的商學(xué)院之一，錄取標(biāo)準(zhǔn)是相當(dāng)嚴(yán)苛。當(dāng)閱讀申請(qǐng)者的推薦信時(shí)，如果申請(qǐng)人的工作經(jīng)歷和申請(qǐng)文書并不是很好的時(shí)候，他的推薦信將成為我們決定是否錄取他的關(guān)鍵性一票。如果有位申請(qǐng)者的GMAT分?jǐn)?shù)相對(duì)較低，但是他的文書寫的很好，工作經(jīng)驗(yàn)很完美，在本科階段也表現(xiàn)的很好，我會(huì)閱讀他的推薦信。如果他的推薦信都是正面的評(píng)價(jià)，我會(huì)投票錄取他。然而，如果成績(jī)和考試分?jǐn)?shù)僅僅達(dá)到了平均水平的話，即使該申請(qǐng)者有一份很有份量的推薦信，我也不是很愿意錄取他，因?yàn)槲抑浪麑⒑茈y跟得上他的同學(xué)，特別是大部分同學(xué)都有很高的成績(jī)和考試分?jǐn)?shù)。

UIUC教授看重的點(diǎn)

背景：教授為UIUC某資深教授，推測(cè)應(yīng)該是常年在招生委員會(huì)里的。我暑期在這里和他做項(xiàng)目，這周就要回家了，走前和他聊了聊申請(qǐng)問題。
1. GRE重要，因?yàn)榭傂枰豁?xiàng)指標(biāo)來給所有學(xué)生排序，而這項(xiàng)的成績(jī)是對(duì)所有人統(tǒng)一的。 
2. GPA是要和本科學(xué)校結(jié)合起來看的，他認(rèn)為中國top校的中上等學(xué)生要好于第十名學(xué)校的第一名學(xué)生（泛指）。
3. 我問他最看重phd applicants的什么，他脫口而出的是leadership，然后解釋說要有independence，能夠獨(dú)立地解決問題，有成為領(lǐng)域領(lǐng)軍者的潛力。
4. 教授們喜歡能夠閱讀文獻(xiàn)、能真正理解、能把自己放在作者的角度上思考下一步要做什么的學(xué)生。
5. Be yourself，每個(gè)學(xué)生都是非常不同的，不必關(guān)心別人什么樣。一些特殊的經(jīng)歷，比如雙學(xué)位等，可以引起教授的興趣。
6. 我怨念的大一GPA低問題：他說可以向教授解釋
7. 不要上太多課，做research不關(guān)注你能上多少課，得多少高分，因?yàn)檎n程往往廣而不深。
8. 套磁要有針對(duì)性，他作為一個(gè)做通信網(wǎng)絡(luò)的人，收到一封聲稱對(duì)他方向感興趣，但個(gè)人簡(jiǎn)歷全是電磁場(chǎng)研究的郵件，唯一的想法就是這個(gè)人把同一封郵件發(fā)給了一百個(gè)人。
9. 感覺他們很重視國際學(xué)生的reading, writing and communication skills，他說要給我寫推薦信時(shí)多次提到我的這方面能力。