15 Pieces of Advice I Wish My PhD Advisor Had Given Me

8/7/2019 15 Pieces of Advice I Wish My PhD Advisor Had Given Me

1/12

10 pieces of advice I wish myPhD advisor had given me

Jim KuroseDepartment of Computer Science

University of MassachusettsAmherst, MA USA

http://www.cs.umass.edu/~kurose

1112

1415

13

With great affection and apologiesto my advisors

Mischa Schwartz, EE Department,

Columbia University

Yechiam Yemini

CS DepartmentColumbia University

and a disclaimer..


2/12

I wish Idtaken more

math courses!

#1

Take math courses!

everymath courseIve taken hasbeen valuable

wont have timelater

research fieldsdraw increasinglyon math as they

mature theory is timeless!

study

broadly

I wish Idtaken more

math courses!

Important coursesoutside CS

signal processing control theory information theory nonlinear optimization stochastic processes game theory domains: systems

biology, economics,

#1

studybroadly


3/12

I wish Idtaken more

math courses!

you will never

again have somuch relaxedtime to study,learn, think

Sorry for that pieceof bad news but

its true

#1study

broadly

Choosing, defining a researchproblempick your problems carefully!

whats the fundamental issueyoure solving?

will the problem be of interestfive, ten years from now?

focus on fundamentals in aworld with an increasingly shortattention span

A fool can ask more

questions in a minute

than a wise man/woman(or a Yoda) can answer in a lifetime

QoSmulticast

congestion control

P2Psensor networks

energy

#2


4/12

Choosing, defining a research

problem There are lots of smartpeople out there!

avoid crowded areasunless you have aunique talent, viewpoint

low-hanging fruit hasbeen picked

researchers working onnext big thing are not inthe crowd

take risks (its research)Wisdom of crowds?

Choosing, defining a researchproblem complexity,

sophistication arethemselves not ofinterest

simple is sometimesharder!

time

solutions in use

understanding ofproblem area

solutions proposed

early middle latesol

utioncomplexity

[adapted from Hluchyj 2001]

!!

!!

maximum impact / mindshare


5/12

Choosing, defining a research

problemavoid point solutions

insights that cut acrosssolution space vs pointsolution

what broaderconclusions can bedrawn from your work?

solution space

.

You are here (but shouldnt be)

Publishing publish where you will get

mindshare, impact

there islife beyond sigcomm,infocom

quality over quantity:

avoid LPUs

1 widely-read/referencedpaper >> K mediocre papers

PhD thesis != magnum opus dont be driven by conference

deadlines

dont submit just to get reviews

#3


6/12

Time: your most precious resource

give yourself time tothink

manage your timecarefully: consciouslychoose how youspend your time

learn to multiplex

Only how to book

recommended by BillClinton

Ive given away 50+

#4

Time

Learn how to write reallywell

can notoverstressimportance of good writing

the most important course?

unfair advantage inpaper, proposal review

outstanding investment ofyour time

study role models

http://www-net.cs.umass.edu/kurose/writing/

"No tale is so good that it can't be

spoiled in the tellingProverb

#5


7/12

Top-10 tips for writing a paper

1: Every paper tells astory

2. Write top down

3. Introduction: crucial,formulaic

4. Master basics oforganized writing

5. Put yourself in placeof reader

6. Write precisely (bespecific, dontembellish)

7. No one (not even yourmother) is as interestedin this topic as you

8. State results carefully

9. Study the art of writing

10. Good writing takestime

Recommended reading:

Writing for Computer Science

by Justin ZobelThe Elements of Style

by William Strunk E. B. White(50 years old and still a classic!)


8/12

Learn how to speak reallywell

cant overstress importance ofgood speaking

important course toteach/take?

unfair advantage in mindshare

convey exciting story/message

thoughtful

engaging

clear, concise

practice, practice, practice videotape, critique yourself

study role models

#6

Learn the processof doing research

our field is a guild

grad student = apprentice

professor = master artisan

its about more than the resultsin your thesis

youll be generating resultsfor a lifetime

knowing process is whatsmost important

why your advisor cant (andshouldnt) solve (or evendefine) the problem for you

apprenticeship

#7


9/12

Think about what you want to do afterwards

academia: teaching schools

research-1 schools

big v small; public vprivate

country?

do you love (or at leastlike) to teach? students?

industry:

many different types ofindustry settings

startup

big industry

research labs

research institutes

#8

A community of scholars

meet people, listen,collaborate

good students, colleagues,

friends

approach, talk withpeople

interactions with peerstudents

research discussions

paper presentations

practice talks

.

#9


10/12

Identify role models

who does somethingyou care about reallywell?

how do they do it?

many role models:

no one does everything

find your balance

get a mentor be a mentor

#10

The last word

Have fun enjoywhat you are doing

Best piece of advice Iever received (1984):Pick a place, job where youll

have fun, enjoy living, enjoyyour colleagues. Without that,no level of success will makeyou happy.Jack K. Wolf

#11


11/12

What have others added (1)?

learn how to deal with rejection

itll happen now and then, for the rest of yourprofessional life (hopefully not with your partner)

learn from rejection: Whywas paper/proposalrejected? What did/didnt reviewers see/like?

know your secret weapon

what unfair advantage do you have over everyoneelse?

learn how to change topics boring to do same thing for 30 years!

What have others added (2)?

learn how to deal with stress

life balance, life changes, too much work

learn how to multiplex

youll be doing it the rest of your life

learn how to read/review/write fast, but well

and follow the 90/10 rule


12/12

Questions!

What did I miss?

What advice would you give to 1st year student?

15 Pieces of Advice I Wish My PhD Advisor Had Given Me

Documents