only valuable experts can be valued

33
1 Only Valuable Experts Can be Valued Moshe Babaioff, Microsoft Research, Silicon Valley Liad Blumrosen, Hebrew U, Dept. of Economics Nicolas Lambert, Stanford GSB Omer Reingold, Microsoft Research, Silicon Valley and Weizmann.

Upload: ilana

Post on 23-Mar-2016

19 views

Category:

Documents


0 download

DESCRIPTION

Only Valuable Experts Can be Valued. Moshe Babaioff , Microsoft Research, Silicon Valley Liad Blumrosen , Hebrew U, Dept. of Economics Nicolas Lambert , Stanford GSB Omer Reingold , Microsoft Research, Silicon Valley and Weizmann. Probabilities of Events. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Only Valuable Experts Can be Valued

1

Only Valuable Experts Can be

ValuedMoshe Babaioff, Microsoft Research, Silicon Valley

Liad Blumrosen, Hebrew U, Dept. of Economics

Nicolas Lambert, Stanford GSB

Omer Reingold, Microsoft Research, Silicon Valley and Weizmann.

Page 2: Only Valuable Experts Can be Valued

2

Probabilities of Events• Often, estimating probabilities of

future events is important.• Examples:

– Weather: probability of rain tomorrow

– Online advertising: what is the click probability of the next visitor on our web-site?

– What % of Toyota cars is defective?

– Many applications in financial markets.

Page 3: Only Valuable Experts Can be Valued

3

Contracts and Screening

Uncertain about the probability of a future event.

Claims he knows this probability.

Averse to uncertainty, is willing to pay $$$ to reduce it.

May be uninformedand pretend to be informed to get $$$…

A decision maker (Alice)

An expert (Bob)

Page 4: Only Valuable Experts Can be Valued

4

Contracts and Screening

• Goal: screen experts.• That is, design contracts such

that:

– Informed experts will:1. accept the contract2. reveal the true probability.

– Uninformed experts will reject the contract.

• Contracts can be based on outcomes only: True probabilities are never revealed.

Page 5: Only Valuable Experts Can be Valued

5

This work• We characterize settings where Alice can

separate good experts from bad experts.

• We discuss what is a “valuable” expert, and its relation to screening of experts.

Page 6: Only Valuable Experts Can be Valued

6

Outline• Model

• With prior:– An easy impossibility result– A positive result

• No priors

• Extensions

Page 7: Only Valuable Experts Can be Valued

7

Model (1/4)• Ω - Finite set of outcomes.• p - A (true) distribution over Ω.

– Unknown to Alice• Φ - The set of possible distributions.

– Φ may be restricted, examples to come…

• Bayesian assumptions:Prior f on Φ.

• f is known to Alice, Bob.

No Prior on Φ.

In the beginning of this talk Later….

Page 8: Only Valuable Experts Can be Valued

8

Model (2/4)• Contract: π(q,ω)

payment to Bob when reporting q when outcome is ω.– Bob is risk neutral.

• Bob’s expected payment depends on what he knows:

– Informed: π(q,p) = Eω~p [ π(q,ω) ]– Uninformed: Ep~f Eω~p [ π(q,ω) ] =

π(q,E[p])

Reported probability

Realized outcome

Notation: E[payment] upon reporting q when the true probability is p.

Page 9: Only Valuable Experts Can be Valued

9

Model (3/4)• Bob δ-accepts the contract if he has

a report q with payment > δ– Otherwise, we say that Bob δ-rejects.

• For avoiding handling ties, we aim that for δa> δr :– an informed Bob will δa -accept – an uninformed Bob will δr –reject

δa

δr

0

Page 10: Only Valuable Experts Can be Valued

10

Model (4/4)

• We actually study a more general model: – experts are ε-informed.– Mixed strategies are allowed.

• This talk: perfectly informed, pure strategies.

Page 11: Only Valuable Experts Can be Valued

11

Example• A binary event: Ω = { , }

Pr( ) = αp Φ = { (α , 1- α) | α [0,1] }

• Alice does not know p.– Knows, however, that α ~ U[0,1] .Sees an a-priori probability of ½.

• Bob claims he knows the realization of p.

• Contract: Bob reports α. Is paid according to or .

Page 12: Only Valuable Experts Can be Valued

12

Main message

• The ability to screen experts closely relates to the structure of Φ– Roughly, on whether Φ is convex or not.

• If all possible experts are “valuable” to Alice, then screening is possible.

Page 13: Only Valuable Experts Can be Valued

13

An Easy Impossibility

• Reason: when the true probability is E[p] a true expert accepts the contract.

Proof:o Informed experts always accept the contract π.That is, for all p, we have π(p,p) > δa.

o Then, an uninformed agent can get > δa by reporting E[p]:Ep~f Eω~p [π(E[p],ω)] =π(E[p],E[p]) > δa

Proposition: screening is impossible.

Page 14: Only Valuable Experts Can be Valued

14

Valuable Experts• So screening is impossible when E[p] Φ.• But experts knowing that the true

distribution is E[p] are not really valuable to Alice.

• In the binary-outcome example:−Alice’s prior is U[0,1], so she believes

that Pr( ) = ½.−An expert knowing that p=½ is not that

helpful…What if all experts are “valuable”?

Page 15: Only Valuable Experts Can be Valued

15

Possibility result• Theorem: when E[p] is not in Φ (and Φ is

closed), screening is possible.– When all experts are valuable, we can

screen...

• Immediate questions:−Is non-convex Φ natural?−Can we expect Φ not to contain E[p]?

E[p]

Φ

Page 16: Only Valuable Experts Can be Valued

16

Non-convex Φ: examples

• Example 1: a coin which is either fair (p=1/2) or biased (p=3/4)– For any (non-trivial) prior, E[p] not in Φ.

• Example 2: many standard distributions are not closed under mixing.– E.g., uniform, normal, etc.

1/2 3/4

Page 17: Only Valuable Experts Can be Valued

17

Non-convex Φ: examples

• Example 3: The binary outcome example. But now, Alice observes two samples.– For example, we wish to know the failure rate

in cars, and we thoroughly check 2 random Toyotas.

Page 18: Only Valuable Experts Can be Valued

18

Non-convex Φ: examples

Ω = { ( , ), ( , ), ( , ) , ( , ) }Φ = { ( α2, α(1- α), (1- α)α, (1-α)2 ) } α [0,1]

• Example 3: The binary outcome example.

But now, Alice observes two samples.

For every prior, E[p] is not in Φ, and thus screening is possible.

Φ is not convex!− Moreover, for every p,p’,

the convex combination of the above vectors is not in Φ.

Page 19: Only Valuable Experts Can be Valued

19

Possibility result• Theorem: when E[p] is not in Φ (and Φ is

closed), screening is possible.Φ

• One definition before proceeding to the proof…

Page 20: Only Valuable Experts Can be Valued

20

Scoring rules• Scoring rules:

– Contracts that elicit distributions from experts.• S(q,ω) = payment for an expert reporting q

when the realized outcome is ω.– A scoring rule is strictly proper if the expert is

always strictly better off by reporting the true distribution.

– Strictly proper scoring rules are known to exist [Brier ‘50, , Good ’52, Savage ‘71,…]

• We want, in addition, to screen good experts from bad.

Page 21: Only Valuable Experts Can be Valued

21

Possibility result• Theorem: when E[p] is not in Φ (and Φ is

closed), screening is possible.Φ• Proof:

• Let s be some strictly proper scoring rule.− On the full probability space

• The following contracts screens experts:

rq

ra qpEsqqspEsqsq

)'],[()','(inf

)],[(),(2),('

We need to show:1. An informed expert δa-accepts.2. An uninformed expert δr-rejects.

Page 22: Only Valuable Experts Can be Valued

22

Possibility result• Theorem: when E[p] is not in Φ (and Φ is

closed), screening is possible.Φ• Proof:

• Let s be some strictly proper scoring rule.• On the full probability space

• The following contracts screens experts:

rq

ra qpEsqqspEsqsq

)'],[()','(inf

)],[(),(2),('

An informed expert reporting the truth p gains: rrrapp 2),(

≥1

Page 23: Only Valuable Experts Can be Valued

23

Possibility result• Theorem: when E[p] is not in Φ (and Φ is

closed), screening is possible.Φ• Proof:

• Let s be some strictly proper scoring rule.• The following contracts screens experts:

Since s is strictly proper, for every q:

An uninformed expert will gain: rrrapEq 02])[,(

0])[],[(])[,( pEpEspEqs

rq

ra qpEsqqspEsqsq

)'],[()','(inf

)],[(),(2),('

Page 24: Only Valuable Experts Can be Valued

24

Outline• Model

• With prior:– An easy impossibility result– A positive result

• No priors

• Extensions

Page 25: Only Valuable Experts Can be Valued

25

Related work• [Olszewski & Sandroni 2007] studied a

similar model:– A binary event with unknown probability p.– No priors:

• An uninformed expert accepts a contract if it is good in the worst case.

• Theorem [O&S]:

– Use min-max theorems.– The probability space is convex.

All informed experts accept a contract

An uninformed expert also accepts it

Page 26: Only Valuable Experts Can be Valued

26

Valuable Experts: No Prior

• We claim: invaluable agents are also behind this impossibility.– But what is a valuable agent without priors?

• What are Alice’s utility function and actions?– A(p) : Alice’s action when she knows p.– U( A(p),p ): utility maximizing actions.

Theorem:Φ is convex (and closed)

There exists p such thatU(A(p),p)=U(A(“reject”),p)no prior on Φ

• Interpretation: if Φ is convex then some informed expert is not valuable.

Page 27: Only Valuable Experts Can be Valued

27

No prior: positive result• We have an analogues positive result for

the no-prior case:non-convex Φ screening is

possible.

• (we use the with-prior positive result in the proof)

Page 28: Only Valuable Experts Can be Valued

28

Outline• Model

• With prior:– An easy impossibility result– A positive result

• No priors

• Extensions

Page 29: Only Valuable Experts Can be Valued

29

Extension: forecasting• A related line of research is forecasting:

– An unbounded sequence of events.– An expert provides a forecast before

each event occurs.– Goal: test the expert.

Page 30: Only Valuable Experts Can be Valued

30

Forecasting: related work

• Negative results are known:– Informed experts pass the test uninformed

experts can do it too. [e.g., Foster & Vohra ‘97, Fudenberg & Levine ‘99]

– When forecasting is possible, decisions can be delayed arbitrarily. [Olszewski & Sandroni ‘09]

• Some works around this impossibility:– [Olszewski & Sandroni ‘09] show a counter-example

by constructing non-convex set of distributions. – [Al-Najjar & Sandroni & Smorodinsky & Weinstein ‘10]

Describe a class of distribution such that decisions can be made in time. The relevant class of distributions also admits non-convexities.

Page 31: Only Valuable Experts Can be Valued

31

Extension: forecasting• We extend our approach to forecasting

settings.– In the works.

• We characterize conditions on the set of distributions that allow expert testing.

• Analysis is more involved, but the ideas are similar.– Results relate to the convexity of Φ.– For example: two samples at each period enable

testing.

Page 32: Only Valuable Experts Can be Valued

32

Summary• A decision maker want to hire an expert.

– For learning the probability of some future event.– The expert may be a charlatan.

• Can the decision maker separate good experts from bad ones?

• We characterize the settings where such screening is possible.– With or without priors on Φ.

• We design screening contracts.

Page 33: Only Valuable Experts Can be Valued

33

Thanks!