註冊 登入
 



打印

跪求Phd 建議

跪求Phd 建議 E-mail 此主題給朋友

[隱藏]
 
等小弟講講background, 30就黎加1,家中有一小孩,undergrad係美國普通public u 讀cs,不過gpa得3.1,好彩既係,我求過一個professor比我參與research project, 叫做有co-author 到2份papers,雖然唔係直接cs research,但算係叫做有data analysis 經驗
畢業之後,一路努力做野下,入到可以話行內數1數2既大公司做software engineer, 近幾個月前,再轉做data scientist後,對於一路都有興趣做research同教書既我,都好想升讀Phd

自知undergrad gpa低分,所以上個學期開始master of cs希望主攻 data science(ML同computer vision),2020年fall應該讀完,用master既gpa補救下,就算升唔到phd,有master對工作上都係有幫助

工作上希望可以出1-2份white paper, 唔知有冇用

終極目標係想做professor,雖然我有美國身份,美國做professor都得,不過都想睇香港有冇可能,因為父母係香港,所以跪求建議

入US既phd唔知機會有幾大,base on而加有既野,想入比較好既grad school可能有d難度,唔知有咩意見比小弟?us讀完係us搵assitant professor位都唔係簡單既事,香港會唔會易d?[worst case入一般既學校,我讀完返去做research scientist]

UK phd時間比較短,仲有機會part time 讀,咁我頭1,2年可能唔洗放棄份工(london有office, 可以調過去) 但想反黎美國做professor係咪變到機會好微?如果用uk phd反香港做professor又可唔可行?


新加坡都有office同樣可以考慮part time phd


最後就係試香港既phd, 但我覺得可以係美國做professor既機會係零,香港讀完香港到做tenure 既機會又如果?

先謝過各位!



精選樓盤
先唔講機會,首先,值唔值先,呢個除咗自己,無人可以代答。值的話,就去試,個人認為首選美國,但無次選。


引用:
原帖由 pharaoh 於 2019-6-1 02:20 PM 發表

等小弟講講background, 30就黎加1,家中有一小孩,undergrad係美國普通public u 讀cs,不過gpa得3.1,好彩既係,我求過一個professor比我參與research project, 叫做有co-author 到2份papers,雖然唔係直接cs research,但算係叫做有data analysis 經驗
畢業之後,一路努力做野下,入到可以話行內數1數2既大 ...
對於美國PhD玩法小弟唔清楚,呢度有位猛人應該可以比到你想要既答案

不過就香港想入行就prof....,你就諗定點樣入top U PhD啦,畢竟香港既競爭好大,而且位極少(同時人工都唔差..,過十萬都有)。
以City/Poly泥計,大部份新入泥既prof都係名校出生,三大會更甚..,所以入之前,先睇下間U夠唔夠猛/submit既paper夠唔夠強(我記得之前識既assistant prof都係UofT+submit左篇野去nature度先入到泥做Prof)

因此,如果你話對比香港做prof同美國做prof既難度的話..,容我直說,香港一定係更難,畢竟即使香港土炮PhD都時有聽聞做到美國Prof(不過冇乜聽過係一線,可能UST會多D吧),但土炮PhD做就香港Prof倒係微乎其微(當然啦,如果你既interest area夠hot既,咁應該ok既)

[ 本帖最後由 uzumymw888 於 2019-6-1 05:38 PM 編輯 ]



買個馬來西亞phd 全世界都認同


回覆 引用 TOP

[隱藏]
 
Hong Kong U


回覆 引用 TOP

八大機會接近零 除非係CS 頂會有幾篇先有機會

單係人工智能今年頂會 大陸教授發表數都有幾百篇 全部都有可能去八大爭位。ust 岩岩先邀請到騰訊首席AI科學家張潼 出任要職 此人來頭甚猛。可見 ust 要求很高!



回覆 引用 TOP

引用:
原帖由 uzumymw888 於 2019-6-1 05:10 PM 發表



對於美國PhD玩法小弟唔清楚,呢度有位猛人應該可以比到你想要既答案

不過就香港想入行就prof....,你就諗定點樣入top U PhD啦,畢竟香港既競爭好大,而且位極少(同時人工都唔差..,過十萬都有)。
以City/Poly泥計,大部份新入泥既prof都係名校出生,三大會更甚..,所以入之前,先睇下間U夠唔夠猛/submit既paper夠唔夠強(我記得之前識既assistant ...
睇返csranking 排名, ust 頂會人均產值已經有美國一流大學水平, 所以科大請
得返黎既一定要係精兵.



回覆 引用 TOP

Well .. i guess I will chime in. First, read the article "hD & Beyond" up top, which provides some information (though not specific to CS).


"入US既phd唔知機會有幾大" .. well 100% because there are always bad schools who will accept anyone. You goal obviously is a top tier department (not school .. you should know very well that it is much more important to work with the right faculty than going to the brand name school, although brand name school usually has good faculty), with a full ride. For that, if you have some research experience, it is possible. The key is not grades or even GRE scores (you have to have good numbers for both to be on the list), but whether you can convince the admission committee that a) you are motivated and passionate about research, b) have the skills and can learn fast, and c) smart enough to ask the right question. We judge applicants more by how they think, then what they know.


Here is one very standard trick. Find faculty you want to work with, read their research and cold-email them asking for a chance to discuss research. We *love* people who want to talk about our work. Make sure you do not just recite stuff and kiss ass ("oh .. you seminar paper is so clever .. blah blah blah"), but offer your own ideas and extensions.


"us讀完係us搵assitant professor位都唔係簡單既事". Now we have to talk about the differences in fields. Since you are interested in data science, you need to consider business school information system departments, and not just CS. (Just to give you some context, i am in a b-school, but have worked with colleagues in CS and engineering, including to be on their search committees, so i know a little about both.) It is much easier to start with a tenure-track faculty job in b-school fields than pure science or engineering (where CS would reside), which may require years of post-docs. Post-doc is relatively less common in b-school disciplines. I have 3 PhD students graduating this year, and all 3 found tenure-track jobs (i know, small sample anecdotal but i have no formal stat now). In addition, the data analytics area is hot. We hired two new tenure-track faculty this year, and it took us like 8!! campus visit to find them. For CS, in the US, you number 1 goal as a faculty member is to get grant. For b-school (including IS), your only goal is to publish. So the game is very different. Personally, i think the publication "game" is a lot more fair, and merit-based, than the grant "game". You need to consider the longer term career path (i.e. tenure requirement) before jumping into a field. Doing a PhD is just a start. Funding-wise, b-school PhD program (like the one i ran), guarantee 4 years and usually support a 5th. CS programs (at least the ones i know of) only guarantee 2 years from the school, and you need to work for a prof who will find funding for the rest.


Let me also clear up the misconception that CS is more technical than b-school. That is certainly true at the undergrad level. If we don't pull the punches, all the undergrads will fail. However, at the research level, it is often as technical. The major different is that b-school research is more applied, meaning relevant to a real world setting (you can still do applied theory under that), than only methodological. The publication game is also different. In CS, you can publish in good conferences (like ACM-EC, which i actually have done before when i was in industry .. that is also why i know both sides a bit) and the turn-around time is fast. In b-school, we only care about big papers in big journals which can take a long time.


"UK phd時間比較短,仲有機會part time 讀" .. your goal is NOT to finish it fast, but to have enough time to create a good research track record to get into a good tenure-track job. I was visiting Durham (which is NOT a bad school) last year and had long conversations about issues of PhD programs with friends over there. The UK system in my opinion is very inadequate. They will let non-research active faculty supervise PhD students. The shorter time means that you have LESS time to explore research with different faculty (now we do have more course work in the first 2 years, but i push students to start research asap and we require a second year paper for that purpose). I would not recommend going to the UK for a PhD. Even if you get into Cambridge (incidentally i have a coauthor there, so i have nothing against them) or Oxford, you still need to consider any US options very carefully.


I understand you have family issues. But purely from a career development point of view, I would not go back to HK. The action is here in the US. Now, China is developing very rapidly in some of these areas but the Chinese academia has lots of problems and I would not recommend going there. 


Lastly, it will be useful if you discuss your research interests more. Just to give you a sense of where I come from, while my training is not CS (actually i am a behavioral economist though i have very broad interests), i have started AI (deep learning) research recently. I have one project (also collaboration with a bio-engineer colleague) using AI/ML to analyze neuro-imaging data (yes, brain scan!) with subjects engaging in business decision-making task. I have another using deep learning (both as a classifier, but also a "simulator" generating "human-like" data, you know about that kind of work, right .. like the deep fake stuff?) studying human behaviors (yes, i have lots of experimental data) in supply chain decisions. I also do consulting silicon valley companies on data science/AI issues (amongst other). BTW, those are pretty lucrative .. so you are in the "right" discipline



引用:
原帖由 IWC-Rolex 於 2019-6-1 04:41 PM 發表

先唔講機會,首先,值唔值先,呢個除咗自己,無人可以代答。值的話,就去試,個人認為首選美國,但無次選。
sorry, 可能我講野1999,其實我心入邊一早決定想去讀,對我黎絕對值,所以想上黎問多d
睇黎大家既建議都係美國讀



回覆 引用 TOP

[隱藏]
 
引用:
原帖由 uzumymw888 於 2019-6-1 05:10 PM 發表



對於美國PhD玩法小弟唔清楚,呢度有位猛人應該可以比到你想要既答案

不過就香港想入行就prof....,你就諗定點樣入top U PhD啦,畢竟香港既競爭好大,而且位極少(同時人工都唔差..,過十萬都有)。
以City/Poly泥計,大部份新入泥既prof都係名校出生,三大會更甚..,所以入之前,先睇下間U夠唔夠猛/submit既paper夠唔夠強(我記得之前識既assistant ...
多謝回覆,絕對答左其中一個我想知既問題,咁樣我可以心左條心留係美國讀,
當然prof narius 講左個好重要既論點,係睇我之後緊邊個老細多過邊間學校,
但以你所見,香港對top U 既定義係咩?我而加master果間cs ranking us頭25(雖然我覺得係有濫收生問題, master大部份都係咁), gpa暫時A-,我自己估/想,phd應該可以去到20-50內,頭20我覺得有時真係好睇命水。所以如果us top20係咪都唔夠玩?

我而加公司保守估計一年幾萬份application一定有,再年年炒走唔少人,對於要同其他業界精英同phd去鬥過你死我活既話,我有些少體驗,我知要上tenure係好難, 我打算抱住遇神殺神遇佛殺佛既決心



引用:
原帖由 narius 於 2019-6-1 09:55 PM 發表

Well .. i guess I will chime in. First, read the article "hD & Beyond" up top, which provides some information (though not specific to CS).


"入US既phd唔知機會有幾大" .. well 100% because there are always ...
引用:
原帖由 narius 於 2019-6-1 09:55 PM 發表

Well .. i guess I will chime in. First, read the article "hD & Beyond" up top, which provides some information (though not specific to CS).


"入US既phd唔知機會有幾大" .. well 100% because there are always ...
感謝professor,回覆非常有用,

其實gpa同gre大概要幾多分先可以on the list? 除左搵相關既professor去問research機會(打算問完自己學校再問其他學校既professor),你見過仲有冇咩方法可以以唔太高gpa又脫穎而出?

我之前undergard既research係關於一d醫學院既研究,主要係對比唔同medical physic equation對tumor病人要用既劑量得出既結果(有mri,ct但唔係用ai做research同埋我只係data analysis多,果幾廿條equations我係唔明), 做左幾年software engineer 而加data science工作research利用簡單既ML model去做legal同business相關既野,例如計算用戶係恐怖份子機率,洗黑錢,sanction普通NLP name matching等,要用到既tech知識其實唔係真係好難,反而係feature engineering難。可以話同prod一樣好多research topic我都okay, 好雜食下, business, healthcare, legal, 去到pure science 都okay
[using AI/ML to analyze neuro-imaging data (yes, brain scan!)]<=簡直夢想research topic之一

所以prof narius建議去試b-school既phd,我絕對有興趣,可以apply唔同model既research(applied machine learning)係好有趣,不過唔知b-school phd application應該點入手,同埋你會點define好定唔好既department for phd? 除左超級名校,有冇咩學校你會推介我留意下?

你有consult開bay area公司實在係太好,會完全知我情況,介唔介意話私下話我知教授你教邊間U
(I have 3 PhD students graduating this year, and all 3 found tenure-track jobs)<= 求收埋我做學生



引用:
原帖由 pharaoh 於 2019-6-2 05:33 AM 發表

感謝professor,回覆非常有用,

其實gpa同gre大概要幾多分先可以on the list? 除左搵相關既professor去問research機會(打算問完自己學校再問其他學校既professor),你見過仲有冇咩方法可以以唔太高gpa又脫穎而出?

我之前undergard既research係關於一d醫學院既研究,主要係對比唔同medical physic equat ...
I am more than happy to give you a bit of advise (I have been doing that enough for students who just emailed and dropped into my office, i may as well do it online!).

However, to give you proper advise, it is much easier to have some personal details. PM me with your email and we can chat a bit via email.



引用:
原帖由 narius 於 2019-6-2 09:38 AM 發表



I am more than happy to give you a bit of advise (I have been doing that enough for students who just emailed and dropped into my office, i may as well do it online!).

However, to give you prop ...
教授,我新手send唔到pm



引用:
原帖由 pharaoh 於 2019-6-2 11:22 AM 發表



教授,我新手send唔到pm
No problem .. i will pm you.



[隱藏]
 
引用:
原帖由 narius 於 2019-6-2 12:43 PM 發表



No problem .. i will pm you.
narius so nice so nice...hope i had a professor like u



 42 123
 提示:支持鍵盤翻頁 ←左 右→
 
 




 

重要聲明:本討論區是以即時上載留言的方式運作,香港討論區對所有留言的真實性、完整性及立場等,不負任何法律責任。而一切留言之言論只代表留言者個人意 見,並非本網站之立場,讀者及用戶不應信賴內容,並應自行判斷內容之真實性。於有關情形下,讀者及用戶應尋求專業意見(如涉及醫療、法律或投資等問題)。 由於本討論區受到「即時上載留言」運作方式所規限,故不能完全監察所有留言,若讀者及用戶發現有留言出現問題,請聯絡我們。香港討論區有權刪除任何留言及拒絕任何人士上載留言 (刪除前或不會作事先警告及通知 ), 同時亦有不刪除留言的權利,如有任何爭議,管理員擁有最終的詮釋權 。用戶切勿撰寫粗言穢語、誹謗、渲染色情暴力或人身攻擊的言論,敬請自律。本網站保留一切法律權利。


Copyright©2003- Discuss.com.hk Limited. All Rights Reserved.
版權所有,不得轉載。