SlideShare una empresa de Scribd logo
1 de 31
Descargar para leer sin conexión
Experimenting in Equilibrium
Stefan Wager
Stanford University
SAMSI Causal Inference
Duke, NC, 9 December 2019
joint work with Kuang Xu
Modern computational infrastructure enables us to routinely and
quickly run large-scale data analyses, and has led to a resurgence
of interest in experimental design.
Many companies, ranging from pharmaceuticals to “traditional”
tech, invest heavily in running multiple randomized trials to
optimize their products.
In recent years, we’ve seen the rise of platforms that support
miniature economies. Experimentation in this setting is harder.
Motivating Example
The following is a toy version of a problem that comes up with
sharing economy platforms:
A platform wants to satisfy demand using freelance workers.
Each day, the platform commits to a payment pi delivered to
worker i for each unit of demand served.
On seeing the offered pi , each worker decides to become
“active” or not.
Demand is randomly allocated among workers who are active
and are not already busy.
The platform and workers have divergent 1-st order preferences:
Workers would prefer high payment and few active workers.
Platform would prefer low payments and many active workers.
Question: How can we set the payments to optimize utility?
Motivating Example
Question: How can we set the payments to optimize utility?
Idea 1: Run a case-control randomized trial, give different
workers different payments.
This won’t work because of interference. Workers who are
paid more are more likely to become active, and cannibalize
demand from others.
Motivating Example
Question: How can we set the payments to optimize utility?
Idea 1: Run a case-control randomized trial, give different
workers different payments.
This won’t work because of interference. Workers who are
paid more are more likely to become active, and cannibalize
demand from others.
Idea 2: Run a randomized trial on non-interfering workers.
But all workers interfere with each other. In principle, you
could randomize across cities, at the cost of loss of power.
Motivating Example
Question: How can we set the payments to optimize utility?
Idea 1: Run a case-control randomized trial, give different
workers different payments.
This won’t work because of interference. Workers who are
paid more are more likely to become active, and cannibalize
demand from others.
Idea 2: Run a randomized trial on non-interfering workers.
But all workers interfere with each other. In principle, you
could randomize across cities, at the cost of loss of power.
Idea 3: Model and correct for interference?
In a large sample mean-field limit, we may be able to
understand quite well how interference works.
Interference
When experimenting in a marketplace, interference is ubiquitous.
In statistics, the classical approach to interference starts from
cutting up the exposure graph (Aronow and Samii, 2017; Athey,
Eckles and Imbens, 2018; Basse, Feller and Toulis, 2019; Hudgens
and Halloran, 2008; Leung, 2019; Manski, 2012; Sobel, 2006).
Main question: Can we design more powerful experiments that are
robust to interference using a little bit of modeling instead.
Key Assumption: Workers respond to expected revenue
In order to correct for interference, our core assumption is that all
interference is mediated by driver response to expected revenue.
Strong assumption, but aligned with empirical evidence in the
ride sharing context (Hall, Horton and Knoepfle, 2019).
As with the sufficient statistics approach in economics (Chetty,
2009), we don’t specify a full model and instead just rely on some
simple relationships.
Key Assumption: Workers respond to expected revenue
In order to correct for interference, our core assumption is that all
interference is mediated by driver response to expected revenue.
Strong assumption, but aligned with empirical evidence in the
ride sharing context (Hall, Horton and Knoepfle, 2019).
As with the sufficient statistics approach in economics (Chetty,
2009), we don’t specify a full model and instead just rely on some
simple relationships.
=⇒ All interference is due to demand cannibalization, and
mediated by total supply.
A simple model
In order to correct for interference, we assume the following model:
The platform chooses a distribution π, and promises a
payment Pi
iid
∼ π to each worker.
If a fraction µ of workers are active, the expected amount of
demand served by any worker if they become active is q(µ).
Workers have random outside options Bi such that, given
the distribution π, the i-th worker is active with probability
fBi
(pi q(µ(π))) = 1/ (1 + exp [−β (pi q(µ(π)) − Bi )]) .
Note: the expected revenue of the i-th worker is pi q(µ(π)).
The system is in equilibrium, i.e., the fraction of active
workers is µ(π) = E [fBi
(pi q(µ(π)))].
Key Idea: A local experiment
We start by running an experiment where we independently
perturb each works payment by a small random amount:
pi = p + ζεi , εi
iid
∼ {±1} .
Under reasonable assumptions, local experimentation does not
alter total supply, and so does not lead to any interference.
Key Idea: A local experiment
We start by running an experiment where we independently
perturb each works payment by a small random amount:
pi = p + ζεi , εi
iid
∼ {±1} .
Under reasonable assumptions, local experimentation does not
alter total supply, and so does not lead to any interference.
Write Zi for whether the i-th worker gets active, and estimate
∆ ←
1
ζ
OLS (Zi ∼ εi )
for the marginal response ∆ of workers to changes in p.
The marginal response function is not of direct policy interest in
itself, because it ignores cannibalization effects.
But given our key assumption, knowing ∆ gets us a long way
towards answering policy-relevant questions.
Simulation study
0 10 20 30 40 50 60
0.00.20.40.60.81.0
payment
fraction
demand served
fraction of suppliers active
demand per active supplier
10 15 20 25 30
19.520.521.522.5
payment
meanutility
optimal
local exp.
global exp.
Consider the following simple simulation study. A platform wants
to choose a payment p that maximizes a utility function U(p).
The experiment is run over a horizon of T = 200 days.
There is no interference across days.
There are large demand fluctuations across days (e.g., due
to weather or special events).
Simulation study
The platform considers the following experimental strategies:
Global experimentation: Each day up to T deploy a shared
random price pt and observe the realized utility Ut. At time
T, fit a spline Ut ∼ pt and deploy the max thereafter.
Local experimentation: Estimate ∆ via price perturbations
pit = pt + ζεit. Obtain an estimate of dU(p)/dp that
accounts for interference. Update pt+1 via gradient descent.
Simulation study
0.0 0.2 0.4 0.6 0.8
0.00.20.40.60.8
in−sample mean regret
futureexpectedregret
q
q
q
q
q
q
q
q q
q
local experimentation
global experimentation
q
q
q
q
q
q
q
q
q
qq
q
q
q
q
qqq
qq
q
q
q
q
qq
qq
qqqqqqqq
q
qqqqqqq
q
qqqqqq
qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq
qqqqqqqqqqqqqqqqqqqqqqqqqq
qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq
0 50 100 150 200
18202224262830
time period
payment
The left panel compares the regret of local vs. global exp.
The right panel illustrates convergence of the pt via local exp.
Mean-field analysis
We adopt an asymptotic setting with n → ∞ workers who could
potentially become active.
Assumption 1: Workers observe a daily state variable A that
allows them to anticipate demand,
lim
n→∞
E |D/n − dA| A = a = 0.
I’ll implicitly condition on a everywhere, and use an a-subscript to
remind us of this.
Assumption 2: The “marketplace dynamics” are scale-invariant:
If there are D units of demand and T = n
i=1 Zi active workers, Ω
units of demand get served, where Ω/T ≈ ω(D/T) for large n,
and ω(·) is a known regular allocation function (taken to be
smooth, concave, non-decreasing, etc.)
A simple model
In order to correct for interference, we assume the following model:
The platform chooses a distribution π, and promises a
payment Pi
iid
∼ π to each worker.
If a fraction µ of workers are active and conditionally on daily
state A = a, the expected amount of demand served by
any worker if they become active is qa(µ).
Workers have random outside options Bi such that, given
the distribution π, the i-th worker is active with probability
fBi
(pi q(µ(π))) = 1/ (1 + exp [−β (pi q(µ(π)) − Bi )]) .
Note: the expected revenue of the i-th worker is pi q(µ(π)).
The system is in equilibrium, i.e., the fraction of active
workers is µa(π) = E fBi
(pi qA(µ(π))) A = a .
NB: The distribution of outside options Bi may depend on state A.
Mean-field analysis
We adopt an asymptotic setting with n → ∞ workers who could
potentially become active.
Fact 1: Given the choice of payment distribution π, an
equilibrium with µa(π) = E fBi
(pi q(µA(π))) A = a exists and is
unique. The number of active workers has a binomial(µa(π), n)
distribution.
Fact 2: As n → ∞, the equilibrium (and relevant derivatives)
converge to a mean-field limit.
Mean-field analysis
Fact 3: Recall our local experiment where we independently
perturb each worker’s payment by a small random amount,
pi = p + ζnεi , εi
iid
∼ {±1} .
Write Zi for whether the i-th worker gets active, and estimate
∆ ←
1
ζn
OLS (Zi ∼ εi ) .
Then, if ζn → 0 and ζn
√
n → ∞,
∆ →p ∆a(p) = q(µa(p))E fBi
(pq(µA(p))) A = a ,
and we refer to ∆a(p) as the marginal response function.
Mean-field analysis
Fact 4: Under out assumptions, the marginal response function
∆ and the supply response dµ(p)/dp are linked via the system
dµa(p)
dp
= ∆a(p) − p∆a(p)
da
µ2
a(p)
ω (da/µa(p))
ω(da/µa(p))
dµa(p)
dp
.
Apart from ∆(p), all other quantities in this equation, da and
µa(p), can be readily observed.
Theorem. The local experimentation strategy outlined above
consistently recovers dµa(p)/dp as n → ∞.
Learning via Local Experimentation
The ultimate goal of the platform is to maximize its utility U, for
our purposes taken as total cost minus total revenue.
Write γ for the platform’s revenue per unit of demand served. In
the mean-field limit, the utility then converges to
n−1
Ua(p) = (γ − p) ω(da/µa(p)) µa(p), U(p) = E [UA(p)] .
Once we know dµa(p)/dp, working out the utility derivative
dUa(p)/dp amounts to calculus.
We consider a platform that uses these estimates to optimize U(p)
by gradient descent (or rather ascent).
A First-Order Algorithm
We now proceed to optimize payments via a variant of mirror
descent Specify a step size η, an interval I = [c−, c+], and an
initial payment p1. Then, at time period t = 1, 2, ...:
1. Deploy randomized payment perturbations εit around pt.
2. Estimate ∆ by regressing market participation on εit.
3. Translate this into an estimate Γt of dUAt (p)/dp via the
transformation implied by the mean-field limit.
4. Perform a gradient update, where θt = t
s=1 sΓs:
pt+1 = argminp
1
2η
t
s=1
s(p − ps)2
− θtp : p ∈ I
If the Ua(p) functions are strongly concave, this attains a 1/t rate
of convergence in large markets, both in regret and squared error.
A First-Order Algorithm
If the Ua(p) functions are strongly concave, this attains a 1/t rate
of convergence in large markets, both in regret and squared error.
Theorem. If the Ua(p) functions are σ-strongly concave,
|ua(p)| ≤ M, and we use a step size η > σ−1 then
lim
n→∞
P
1
T
T
t=1
t (UAt (p) − UAt (pt)) ≤
ηM2
2
= 1,
for any fixed payment p ∈ [c−, c+].
Corollary. If in addition the day-specific states At are IID, then
lim sup
n→∞
P (p∗
− ¯pT )2
≤
ηM2
σT
16 log δ−1
+ 4 ≥ 1 − δ,
p∗ = argmax {E [UA(p)] : p ∈ I} and ¯pT = 2
T(T+1)
T
t=1 t pt.
Comparison with global experimentation
Conceptually, our problem is closely related to the literature on
continuous-armed bandits, motivated by the following setting:
In each time period, the analyst deploys pt, and observes a
reward Ut = U(pt) + noise.
We want to control regret T−1 T
t=1 (U(p∗) − U(pt)).
Some references include Bubeck et al. (2017), Flaxman et al.
(2005), Kleinberg (2005) and Shamir (2013).
The optimal regret in this problem scales as 1/
√
T, even if we
know U(p) is quadratic (Shamir, 2013).
Comparison with global experimentation
Here, instead, the gradients we get via our approach enable a 1/T
rate of convergence.
In other words, if local experimentation is applicable it
fundamentally changes the difficulty of the problem relative to
the continuous-armed bandits setting.
The gain from local experimentation is comparable to the gain we
could get from running two function evaluations with the same
noise (Duchi et al., 2015).
Extensions via generalized earning functions
The core assumption that enables our approach is that workers
care only about expected revenue, and thus respond to payments
pi and market-level congestion q(µa(π)) via their product.
Then, we showed that the mean-field limit is characterized by the
following balance condition.
µa(π) = E fBi
(pi q(µA(π))) A = a .
The form of this balance condition is crucial: If fB can have a
generic dependence on pi and q, we may run into intractable
difficulties.
Extensions via generalized earning functions
One way to generalize this setting is to let workers respond to pi
and q via a (known) generalized earning function (GEF) θ,
µa(π) = E fBi
(θ(pi , q(µA(π)))) A = a .
Example: Risk aversion. Workers respond to the expectation of
a concave function of revenue. In the binary case where each
worker serves 0 or 1 units of demand, we get θ(p, q) = β(p)q for
some concave β(·).
Example: Supply-side surge pricing. The platform commits to
paying the i-th worker s(D/T)pi for some increasing surge
multiplier s(·). Surge is automatic and anticipated by workers.
With surge, the mean-field limit of expected revenue of the i-th
worker is θ(p, q) = pqs(ω−1(q)).
Extensions via generalized earning functions
One way to generalize this setting is to let workers respond to pi
and q via a (known) generalized earning function (GEF) θ,
µa(π) = E fBi
(θ(pi , q(µA(π)))) A = a .
With GEF, the balance condition implies that a marginal
response function can be estimated via local perturbations:
∆a(p) = pθ(p, q(µa(p)))E fBi
(θ(p, q(µA(p)))) A = a .
Then, dµa(p)/dp can be linked to ∆a(p) via a linear system that
depends on system dynamics, thus enabling local experimentation:
dµa(p)
dp
= p + q (µa(p))
dµa(p)
dp
q θ(p, q(µa(p)))
E fBi
(θ(p, q(µA(p)))) A = a .
NB: The above are conjectures; no formal results yet with GEF.
Simulation study: surge pricing
0.0 0.2 0.4 0.6 0.8
0.00.20.40.60.8
in−sample mean regret
futureexpectedregret
q
q
q
q
q
q
q
q q
q
local experimentation
global experimentation
0.0 0.2 0.4 0.6 0.8
0.00.20.40.60.8
in−sample mean regret
futureexpectedregret
q
q
q
q
q
q
q q
q
q
local experimentation
global experimentation
The left panel is the simulation experiment from the beginning.
The right panel shows results with an extension of our method
that allows for surge pricing.
Most work on experimental design assumes no interference, but
this assumption often fails in a marketplace setting.
We showed, however, that in some cases we can correct for
interference with better power using some light-weight modeling.
exposure graph mechanism
graph cutting sparse and known arbitrary
model based complete mean-field game
There are more open questions than closed ones.
Thanks!

Más contenido relacionado

Similar a Causal Inference Opening Workshop - Experimenting in Equilibrium - Stefan Wager, December 9, 2019

Read the article Competition or Complement Six Sigma and TOC that.docx
Read the article Competition or Complement Six Sigma and TOC that.docxRead the article Competition or Complement Six Sigma and TOC that.docx
Read the article Competition or Complement Six Sigma and TOC that.docxmakdul
 
Lecture 1: NBERMetrics
Lecture 1: NBERMetricsLecture 1: NBERMetrics
Lecture 1: NBERMetricsNBER
 
SURE Model_Panel data.pptx
SURE Model_Panel data.pptxSURE Model_Panel data.pptx
SURE Model_Panel data.pptxGeetaShreeprabha
 
Introduction to Econometrics for under gruadute class.pptx
Introduction to Econometrics for under gruadute class.pptxIntroduction to Econometrics for under gruadute class.pptx
Introduction to Econometrics for under gruadute class.pptxtadegebreyesus
 
InstructionDue Date 6 pm on October 28 (Wed)Part IProbability a.docx
InstructionDue Date 6 pm on October 28 (Wed)Part IProbability a.docxInstructionDue Date 6 pm on October 28 (Wed)Part IProbability a.docx
InstructionDue Date 6 pm on October 28 (Wed)Part IProbability a.docxdirkrplav
 
Operations Management VTU BE Mechanical 2015 Solved paper
Operations Management VTU BE Mechanical 2015 Solved paperOperations Management VTU BE Mechanical 2015 Solved paper
Operations Management VTU BE Mechanical 2015 Solved paperSomashekar S.M
 
Nber Lecture Final
Nber Lecture FinalNber Lecture Final
Nber Lecture FinalNBER
 
Degree presentation: Indirect Inference Applied to Financial Econometrics
Degree presentation: Indirect Inference Applied to Financial EconometricsDegree presentation: Indirect Inference Applied to Financial Econometrics
Degree presentation: Indirect Inference Applied to Financial EconometricsJean Duchesne
 
Amortized Analysis of Algorithms
Amortized Analysis of Algorithms Amortized Analysis of Algorithms
Amortized Analysis of Algorithms sathish sak
 
Creative destruction & job mobility: flexibility in the land of Schumpeter
Creative destruction & job mobility: flexibility in the land of SchumpeterCreative destruction & job mobility: flexibility in the land of Schumpeter
Creative destruction & job mobility: flexibility in the land of SchumpeterADEMU_Project
 
Classification methods and assessment
Classification methods and assessmentClassification methods and assessment
Classification methods and assessmentLeonardo Auslender
 
Hypothesis Testing techniques in social research.ppt
Hypothesis Testing techniques in social research.pptHypothesis Testing techniques in social research.ppt
Hypothesis Testing techniques in social research.pptSolomonkiplimo
 

Similar a Causal Inference Opening Workshop - Experimenting in Equilibrium - Stefan Wager, December 9, 2019 (18)

PREDICT 422 - Module 1.pptx
PREDICT 422 - Module 1.pptxPREDICT 422 - Module 1.pptx
PREDICT 422 - Module 1.pptx
 
Read the article Competition or Complement Six Sigma and TOC that.docx
Read the article Competition or Complement Six Sigma and TOC that.docxRead the article Competition or Complement Six Sigma and TOC that.docx
Read the article Competition or Complement Six Sigma and TOC that.docx
 
Lecture 1: NBERMetrics
Lecture 1: NBERMetricsLecture 1: NBERMetrics
Lecture 1: NBERMetrics
 
Dynamic optimal taxation,
Dynamic  optimal  taxation,Dynamic  optimal  taxation,
Dynamic optimal taxation,
 
Ch.6.pdf
Ch.6.pdfCh.6.pdf
Ch.6.pdf
 
SURE Model_Panel data.pptx
SURE Model_Panel data.pptxSURE Model_Panel data.pptx
SURE Model_Panel data.pptx
 
Introduction to Econometrics for under gruadute class.pptx
Introduction to Econometrics for under gruadute class.pptxIntroduction to Econometrics for under gruadute class.pptx
Introduction to Econometrics for under gruadute class.pptx
 
InstructionDue Date 6 pm on October 28 (Wed)Part IProbability a.docx
InstructionDue Date 6 pm on October 28 (Wed)Part IProbability a.docxInstructionDue Date 6 pm on October 28 (Wed)Part IProbability a.docx
InstructionDue Date 6 pm on October 28 (Wed)Part IProbability a.docx
 
Operations Management VTU BE Mechanical 2015 Solved paper
Operations Management VTU BE Mechanical 2015 Solved paperOperations Management VTU BE Mechanical 2015 Solved paper
Operations Management VTU BE Mechanical 2015 Solved paper
 
Causal Inference in Marketing
Causal Inference in MarketingCausal Inference in Marketing
Causal Inference in Marketing
 
Nber Lecture Final
Nber Lecture FinalNber Lecture Final
Nber Lecture Final
 
Degree presentation: Indirect Inference Applied to Financial Econometrics
Degree presentation: Indirect Inference Applied to Financial EconometricsDegree presentation: Indirect Inference Applied to Financial Econometrics
Degree presentation: Indirect Inference Applied to Financial Econometrics
 
Amortized Analysis of Algorithms
Amortized Analysis of Algorithms Amortized Analysis of Algorithms
Amortized Analysis of Algorithms
 
Creative destruction & job mobility: flexibility in the land of Schumpeter
Creative destruction & job mobility: flexibility in the land of SchumpeterCreative destruction & job mobility: flexibility in the land of Schumpeter
Creative destruction & job mobility: flexibility in the land of Schumpeter
 
macro_mini_thesis
macro_mini_thesismacro_mini_thesis
macro_mini_thesis
 
Classification methods and assessment
Classification methods and assessmentClassification methods and assessment
Classification methods and assessment
 
Hypothesis Testing techniques in social research.ppt
Hypothesis Testing techniques in social research.pptHypothesis Testing techniques in social research.ppt
Hypothesis Testing techniques in social research.ppt
 
Macro Development.pptx
Macro Development.pptxMacro Development.pptx
Macro Development.pptx
 

Más de The Statistical and Applied Mathematical Sciences Institute

Más de The Statistical and Applied Mathematical Sciences Institute (20)

Causal Inference Opening Workshop - Latent Variable Models, Causal Inference,...
Causal Inference Opening Workshop - Latent Variable Models, Causal Inference,...Causal Inference Opening Workshop - Latent Variable Models, Causal Inference,...
Causal Inference Opening Workshop - Latent Variable Models, Causal Inference,...
 
2019 Fall Series: Special Guest Lecture - 0-1 Phase Transitions in High Dimen...
2019 Fall Series: Special Guest Lecture - 0-1 Phase Transitions in High Dimen...2019 Fall Series: Special Guest Lecture - 0-1 Phase Transitions in High Dimen...
2019 Fall Series: Special Guest Lecture - 0-1 Phase Transitions in High Dimen...
 
Causal Inference Opening Workshop - Causal Discovery in Neuroimaging Data - F...
Causal Inference Opening Workshop - Causal Discovery in Neuroimaging Data - F...Causal Inference Opening Workshop - Causal Discovery in Neuroimaging Data - F...
Causal Inference Opening Workshop - Causal Discovery in Neuroimaging Data - F...
 
Causal Inference Opening Workshop - Smooth Extensions to BART for Heterogeneo...
Causal Inference Opening Workshop - Smooth Extensions to BART for Heterogeneo...Causal Inference Opening Workshop - Smooth Extensions to BART for Heterogeneo...
Causal Inference Opening Workshop - Smooth Extensions to BART for Heterogeneo...
 
Causal Inference Opening Workshop - A Bracketing Relationship between Differe...
Causal Inference Opening Workshop - A Bracketing Relationship between Differe...Causal Inference Opening Workshop - A Bracketing Relationship between Differe...
Causal Inference Opening Workshop - A Bracketing Relationship between Differe...
 
Causal Inference Opening Workshop - Testing Weak Nulls in Matched Observation...
Causal Inference Opening Workshop - Testing Weak Nulls in Matched Observation...Causal Inference Opening Workshop - Testing Weak Nulls in Matched Observation...
Causal Inference Opening Workshop - Testing Weak Nulls in Matched Observation...
 
Causal Inference Opening Workshop - Difference-in-differences: more than meet...
Causal Inference Opening Workshop - Difference-in-differences: more than meet...Causal Inference Opening Workshop - Difference-in-differences: more than meet...
Causal Inference Opening Workshop - Difference-in-differences: more than meet...
 
Causal Inference Opening Workshop - New Statistical Learning Methods for Esti...
Causal Inference Opening Workshop - New Statistical Learning Methods for Esti...Causal Inference Opening Workshop - New Statistical Learning Methods for Esti...
Causal Inference Opening Workshop - New Statistical Learning Methods for Esti...
 
Causal Inference Opening Workshop - Bipartite Causal Inference with Interfere...
Causal Inference Opening Workshop - Bipartite Causal Inference with Interfere...Causal Inference Opening Workshop - Bipartite Causal Inference with Interfere...
Causal Inference Opening Workshop - Bipartite Causal Inference with Interfere...
 
Causal Inference Opening Workshop - Bridging the Gap Between Causal Literatur...
Causal Inference Opening Workshop - Bridging the Gap Between Causal Literatur...Causal Inference Opening Workshop - Bridging the Gap Between Causal Literatur...
Causal Inference Opening Workshop - Bridging the Gap Between Causal Literatur...
 
Causal Inference Opening Workshop - Some Applications of Reinforcement Learni...
Causal Inference Opening Workshop - Some Applications of Reinforcement Learni...Causal Inference Opening Workshop - Some Applications of Reinforcement Learni...
Causal Inference Opening Workshop - Some Applications of Reinforcement Learni...
 
Causal Inference Opening Workshop - Bracketing Bounds for Differences-in-Diff...
Causal Inference Opening Workshop - Bracketing Bounds for Differences-in-Diff...Causal Inference Opening Workshop - Bracketing Bounds for Differences-in-Diff...
Causal Inference Opening Workshop - Bracketing Bounds for Differences-in-Diff...
 
Causal Inference Opening Workshop - Assisting the Impact of State Polcies: Br...
Causal Inference Opening Workshop - Assisting the Impact of State Polcies: Br...Causal Inference Opening Workshop - Assisting the Impact of State Polcies: Br...
Causal Inference Opening Workshop - Assisting the Impact of State Polcies: Br...
 
Causal Inference Opening Workshop - Targeted Learning for Causal Inference Ba...
Causal Inference Opening Workshop - Targeted Learning for Causal Inference Ba...Causal Inference Opening Workshop - Targeted Learning for Causal Inference Ba...
Causal Inference Opening Workshop - Targeted Learning for Causal Inference Ba...
 
Causal Inference Opening Workshop - Bayesian Nonparametric Models for Treatme...
Causal Inference Opening Workshop - Bayesian Nonparametric Models for Treatme...Causal Inference Opening Workshop - Bayesian Nonparametric Models for Treatme...
Causal Inference Opening Workshop - Bayesian Nonparametric Models for Treatme...
 
2019 Fall Series: Special Guest Lecture - Adversarial Risk Analysis of the Ge...
2019 Fall Series: Special Guest Lecture - Adversarial Risk Analysis of the Ge...2019 Fall Series: Special Guest Lecture - Adversarial Risk Analysis of the Ge...
2019 Fall Series: Special Guest Lecture - Adversarial Risk Analysis of the Ge...
 
2019 Fall Series: Professional Development, Writing Academic Papers…What Work...
2019 Fall Series: Professional Development, Writing Academic Papers…What Work...2019 Fall Series: Professional Development, Writing Academic Papers…What Work...
2019 Fall Series: Professional Development, Writing Academic Papers…What Work...
 
2019 GDRR: Blockchain Data Analytics - Machine Learning in/for Blockchain: Fu...
2019 GDRR: Blockchain Data Analytics - Machine Learning in/for Blockchain: Fu...2019 GDRR: Blockchain Data Analytics - Machine Learning in/for Blockchain: Fu...
2019 GDRR: Blockchain Data Analytics - Machine Learning in/for Blockchain: Fu...
 
2019 GDRR: Blockchain Data Analytics - QuTrack: Model Life Cycle Management f...
2019 GDRR: Blockchain Data Analytics - QuTrack: Model Life Cycle Management f...2019 GDRR: Blockchain Data Analytics - QuTrack: Model Life Cycle Management f...
2019 GDRR: Blockchain Data Analytics - QuTrack: Model Life Cycle Management f...
 
2019 GDRR: Blockchain Data Analytics - Modeling Cryptocurrency Markets with T...
2019 GDRR: Blockchain Data Analytics - Modeling Cryptocurrency Markets with T...2019 GDRR: Blockchain Data Analytics - Modeling Cryptocurrency Markets with T...
2019 GDRR: Blockchain Data Analytics - Modeling Cryptocurrency Markets with T...
 

Último

ISYU TUNGKOL SA SEKSWLADIDA (ISSUE ABOUT SEXUALITY
ISYU TUNGKOL SA SEKSWLADIDA (ISSUE ABOUT SEXUALITYISYU TUNGKOL SA SEKSWLADIDA (ISSUE ABOUT SEXUALITY
ISYU TUNGKOL SA SEKSWLADIDA (ISSUE ABOUT SEXUALITYKayeClaireEstoconing
 
Roles & Responsibilities in Pharmacovigilance
Roles & Responsibilities in PharmacovigilanceRoles & Responsibilities in Pharmacovigilance
Roles & Responsibilities in PharmacovigilanceSamikshaHamane
 
Karra SKD Conference Presentation Revised.pptx
Karra SKD Conference Presentation Revised.pptxKarra SKD Conference Presentation Revised.pptx
Karra SKD Conference Presentation Revised.pptxAshokKarra1
 
DATA STRUCTURE AND ALGORITHM for beginners
DATA STRUCTURE AND ALGORITHM for beginnersDATA STRUCTURE AND ALGORITHM for beginners
DATA STRUCTURE AND ALGORITHM for beginnersSabitha Banu
 
How to Add Barcode on PDF Report in Odoo 17
How to Add Barcode on PDF Report in Odoo 17How to Add Barcode on PDF Report in Odoo 17
How to Add Barcode on PDF Report in Odoo 17Celine George
 
Grade 9 Q4-MELC1-Active and Passive Voice.pptx
Grade 9 Q4-MELC1-Active and Passive Voice.pptxGrade 9 Q4-MELC1-Active and Passive Voice.pptx
Grade 9 Q4-MELC1-Active and Passive Voice.pptxChelloAnnAsuncion2
 
Q4 English4 Week3 PPT Melcnmg-based.pptx
Q4 English4 Week3 PPT Melcnmg-based.pptxQ4 English4 Week3 PPT Melcnmg-based.pptx
Q4 English4 Week3 PPT Melcnmg-based.pptxnelietumpap1
 
ANG SEKTOR NG agrikultura.pptx QUARTER 4
ANG SEKTOR NG agrikultura.pptx QUARTER 4ANG SEKTOR NG agrikultura.pptx QUARTER 4
ANG SEKTOR NG agrikultura.pptx QUARTER 4MiaBumagat1
 
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17Celine George
 
Earth Day Presentation wow hello nice great
Earth Day Presentation wow hello nice greatEarth Day Presentation wow hello nice great
Earth Day Presentation wow hello nice greatYousafMalik24
 
MULTIDISCIPLINRY NATURE OF THE ENVIRONMENTAL STUDIES.pptx
MULTIDISCIPLINRY NATURE OF THE ENVIRONMENTAL STUDIES.pptxMULTIDISCIPLINRY NATURE OF THE ENVIRONMENTAL STUDIES.pptx
MULTIDISCIPLINRY NATURE OF THE ENVIRONMENTAL STUDIES.pptxAnupkumar Sharma
 
ACC 2024 Chronicles. Cardiology. Exam.pdf
ACC 2024 Chronicles. Cardiology. Exam.pdfACC 2024 Chronicles. Cardiology. Exam.pdf
ACC 2024 Chronicles. Cardiology. Exam.pdfSpandanaRallapalli
 
What is Model Inheritance in Odoo 17 ERP
What is Model Inheritance in Odoo 17 ERPWhat is Model Inheritance in Odoo 17 ERP
What is Model Inheritance in Odoo 17 ERPCeline George
 
Computed Fields and api Depends in the Odoo 17
Computed Fields and api Depends in the Odoo 17Computed Fields and api Depends in the Odoo 17
Computed Fields and api Depends in the Odoo 17Celine George
 
Difference Between Search & Browse Methods in Odoo 17
Difference Between Search & Browse Methods in Odoo 17Difference Between Search & Browse Methods in Odoo 17
Difference Between Search & Browse Methods in Odoo 17Celine George
 
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...JhezDiaz1
 

Último (20)

ISYU TUNGKOL SA SEKSWLADIDA (ISSUE ABOUT SEXUALITY
ISYU TUNGKOL SA SEKSWLADIDA (ISSUE ABOUT SEXUALITYISYU TUNGKOL SA SEKSWLADIDA (ISSUE ABOUT SEXUALITY
ISYU TUNGKOL SA SEKSWLADIDA (ISSUE ABOUT SEXUALITY
 
Roles & Responsibilities in Pharmacovigilance
Roles & Responsibilities in PharmacovigilanceRoles & Responsibilities in Pharmacovigilance
Roles & Responsibilities in Pharmacovigilance
 
Karra SKD Conference Presentation Revised.pptx
Karra SKD Conference Presentation Revised.pptxKarra SKD Conference Presentation Revised.pptx
Karra SKD Conference Presentation Revised.pptx
 
DATA STRUCTURE AND ALGORITHM for beginners
DATA STRUCTURE AND ALGORITHM for beginnersDATA STRUCTURE AND ALGORITHM for beginners
DATA STRUCTURE AND ALGORITHM for beginners
 
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptxFINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
FINALS_OF_LEFT_ON_C'N_EL_DORADO_2024.pptx
 
How to Add Barcode on PDF Report in Odoo 17
How to Add Barcode on PDF Report in Odoo 17How to Add Barcode on PDF Report in Odoo 17
How to Add Barcode on PDF Report in Odoo 17
 
Grade 9 Q4-MELC1-Active and Passive Voice.pptx
Grade 9 Q4-MELC1-Active and Passive Voice.pptxGrade 9 Q4-MELC1-Active and Passive Voice.pptx
Grade 9 Q4-MELC1-Active and Passive Voice.pptx
 
OS-operating systems- ch04 (Threads) ...
OS-operating systems- ch04 (Threads) ...OS-operating systems- ch04 (Threads) ...
OS-operating systems- ch04 (Threads) ...
 
Q4 English4 Week3 PPT Melcnmg-based.pptx
Q4 English4 Week3 PPT Melcnmg-based.pptxQ4 English4 Week3 PPT Melcnmg-based.pptx
Q4 English4 Week3 PPT Melcnmg-based.pptx
 
ANG SEKTOR NG agrikultura.pptx QUARTER 4
ANG SEKTOR NG agrikultura.pptx QUARTER 4ANG SEKTOR NG agrikultura.pptx QUARTER 4
ANG SEKTOR NG agrikultura.pptx QUARTER 4
 
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
Incoming and Outgoing Shipments in 3 STEPS Using Odoo 17
 
Earth Day Presentation wow hello nice great
Earth Day Presentation wow hello nice greatEarth Day Presentation wow hello nice great
Earth Day Presentation wow hello nice great
 
LEFT_ON_C'N_ PRELIMS_EL_DORADO_2024.pptx
LEFT_ON_C'N_ PRELIMS_EL_DORADO_2024.pptxLEFT_ON_C'N_ PRELIMS_EL_DORADO_2024.pptx
LEFT_ON_C'N_ PRELIMS_EL_DORADO_2024.pptx
 
Raw materials used in Herbal Cosmetics.pptx
Raw materials used in Herbal Cosmetics.pptxRaw materials used in Herbal Cosmetics.pptx
Raw materials used in Herbal Cosmetics.pptx
 
MULTIDISCIPLINRY NATURE OF THE ENVIRONMENTAL STUDIES.pptx
MULTIDISCIPLINRY NATURE OF THE ENVIRONMENTAL STUDIES.pptxMULTIDISCIPLINRY NATURE OF THE ENVIRONMENTAL STUDIES.pptx
MULTIDISCIPLINRY NATURE OF THE ENVIRONMENTAL STUDIES.pptx
 
ACC 2024 Chronicles. Cardiology. Exam.pdf
ACC 2024 Chronicles. Cardiology. Exam.pdfACC 2024 Chronicles. Cardiology. Exam.pdf
ACC 2024 Chronicles. Cardiology. Exam.pdf
 
What is Model Inheritance in Odoo 17 ERP
What is Model Inheritance in Odoo 17 ERPWhat is Model Inheritance in Odoo 17 ERP
What is Model Inheritance in Odoo 17 ERP
 
Computed Fields and api Depends in the Odoo 17
Computed Fields and api Depends in the Odoo 17Computed Fields and api Depends in the Odoo 17
Computed Fields and api Depends in the Odoo 17
 
Difference Between Search & Browse Methods in Odoo 17
Difference Between Search & Browse Methods in Odoo 17Difference Between Search & Browse Methods in Odoo 17
Difference Between Search & Browse Methods in Odoo 17
 
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
ENGLISH 7_Q4_LESSON 2_ Employing a Variety of Strategies for Effective Interp...
 

Causal Inference Opening Workshop - Experimenting in Equilibrium - Stefan Wager, December 9, 2019

  • 1. Experimenting in Equilibrium Stefan Wager Stanford University SAMSI Causal Inference Duke, NC, 9 December 2019 joint work with Kuang Xu
  • 2. Modern computational infrastructure enables us to routinely and quickly run large-scale data analyses, and has led to a resurgence of interest in experimental design. Many companies, ranging from pharmaceuticals to “traditional” tech, invest heavily in running multiple randomized trials to optimize their products. In recent years, we’ve seen the rise of platforms that support miniature economies. Experimentation in this setting is harder.
  • 3. Motivating Example The following is a toy version of a problem that comes up with sharing economy platforms: A platform wants to satisfy demand using freelance workers. Each day, the platform commits to a payment pi delivered to worker i for each unit of demand served. On seeing the offered pi , each worker decides to become “active” or not. Demand is randomly allocated among workers who are active and are not already busy. The platform and workers have divergent 1-st order preferences: Workers would prefer high payment and few active workers. Platform would prefer low payments and many active workers. Question: How can we set the payments to optimize utility?
  • 4. Motivating Example Question: How can we set the payments to optimize utility? Idea 1: Run a case-control randomized trial, give different workers different payments. This won’t work because of interference. Workers who are paid more are more likely to become active, and cannibalize demand from others.
  • 5. Motivating Example Question: How can we set the payments to optimize utility? Idea 1: Run a case-control randomized trial, give different workers different payments. This won’t work because of interference. Workers who are paid more are more likely to become active, and cannibalize demand from others. Idea 2: Run a randomized trial on non-interfering workers. But all workers interfere with each other. In principle, you could randomize across cities, at the cost of loss of power.
  • 6. Motivating Example Question: How can we set the payments to optimize utility? Idea 1: Run a case-control randomized trial, give different workers different payments. This won’t work because of interference. Workers who are paid more are more likely to become active, and cannibalize demand from others. Idea 2: Run a randomized trial on non-interfering workers. But all workers interfere with each other. In principle, you could randomize across cities, at the cost of loss of power. Idea 3: Model and correct for interference? In a large sample mean-field limit, we may be able to understand quite well how interference works.
  • 7. Interference When experimenting in a marketplace, interference is ubiquitous. In statistics, the classical approach to interference starts from cutting up the exposure graph (Aronow and Samii, 2017; Athey, Eckles and Imbens, 2018; Basse, Feller and Toulis, 2019; Hudgens and Halloran, 2008; Leung, 2019; Manski, 2012; Sobel, 2006). Main question: Can we design more powerful experiments that are robust to interference using a little bit of modeling instead.
  • 8. Key Assumption: Workers respond to expected revenue In order to correct for interference, our core assumption is that all interference is mediated by driver response to expected revenue. Strong assumption, but aligned with empirical evidence in the ride sharing context (Hall, Horton and Knoepfle, 2019). As with the sufficient statistics approach in economics (Chetty, 2009), we don’t specify a full model and instead just rely on some simple relationships.
  • 9. Key Assumption: Workers respond to expected revenue In order to correct for interference, our core assumption is that all interference is mediated by driver response to expected revenue. Strong assumption, but aligned with empirical evidence in the ride sharing context (Hall, Horton and Knoepfle, 2019). As with the sufficient statistics approach in economics (Chetty, 2009), we don’t specify a full model and instead just rely on some simple relationships. =⇒ All interference is due to demand cannibalization, and mediated by total supply.
  • 10. A simple model In order to correct for interference, we assume the following model: The platform chooses a distribution π, and promises a payment Pi iid ∼ π to each worker. If a fraction µ of workers are active, the expected amount of demand served by any worker if they become active is q(µ). Workers have random outside options Bi such that, given the distribution π, the i-th worker is active with probability fBi (pi q(µ(π))) = 1/ (1 + exp [−β (pi q(µ(π)) − Bi )]) . Note: the expected revenue of the i-th worker is pi q(µ(π)). The system is in equilibrium, i.e., the fraction of active workers is µ(π) = E [fBi (pi q(µ(π)))].
  • 11. Key Idea: A local experiment We start by running an experiment where we independently perturb each works payment by a small random amount: pi = p + ζεi , εi iid ∼ {±1} . Under reasonable assumptions, local experimentation does not alter total supply, and so does not lead to any interference.
  • 12. Key Idea: A local experiment We start by running an experiment where we independently perturb each works payment by a small random amount: pi = p + ζεi , εi iid ∼ {±1} . Under reasonable assumptions, local experimentation does not alter total supply, and so does not lead to any interference. Write Zi for whether the i-th worker gets active, and estimate ∆ ← 1 ζ OLS (Zi ∼ εi ) for the marginal response ∆ of workers to changes in p. The marginal response function is not of direct policy interest in itself, because it ignores cannibalization effects. But given our key assumption, knowing ∆ gets us a long way towards answering policy-relevant questions.
  • 13. Simulation study 0 10 20 30 40 50 60 0.00.20.40.60.81.0 payment fraction demand served fraction of suppliers active demand per active supplier 10 15 20 25 30 19.520.521.522.5 payment meanutility optimal local exp. global exp. Consider the following simple simulation study. A platform wants to choose a payment p that maximizes a utility function U(p). The experiment is run over a horizon of T = 200 days. There is no interference across days. There are large demand fluctuations across days (e.g., due to weather or special events).
  • 14. Simulation study The platform considers the following experimental strategies: Global experimentation: Each day up to T deploy a shared random price pt and observe the realized utility Ut. At time T, fit a spline Ut ∼ pt and deploy the max thereafter. Local experimentation: Estimate ∆ via price perturbations pit = pt + ζεit. Obtain an estimate of dU(p)/dp that accounts for interference. Update pt+1 via gradient descent.
  • 15. Simulation study 0.0 0.2 0.4 0.6 0.8 0.00.20.40.60.8 in−sample mean regret futureexpectedregret q q q q q q q q q q local experimentation global experimentation q q q q q q q q q qq q q q q qqq qq q q q q qq qq qqqqqqqq q qqqqqqq q qqqqqq qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq qqqqqqqqqqqqqqqqqqqqqqqqqq qqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqqq 0 50 100 150 200 18202224262830 time period payment The left panel compares the regret of local vs. global exp. The right panel illustrates convergence of the pt via local exp.
  • 16. Mean-field analysis We adopt an asymptotic setting with n → ∞ workers who could potentially become active. Assumption 1: Workers observe a daily state variable A that allows them to anticipate demand, lim n→∞ E |D/n − dA| A = a = 0. I’ll implicitly condition on a everywhere, and use an a-subscript to remind us of this. Assumption 2: The “marketplace dynamics” are scale-invariant: If there are D units of demand and T = n i=1 Zi active workers, Ω units of demand get served, where Ω/T ≈ ω(D/T) for large n, and ω(·) is a known regular allocation function (taken to be smooth, concave, non-decreasing, etc.)
  • 17. A simple model In order to correct for interference, we assume the following model: The platform chooses a distribution π, and promises a payment Pi iid ∼ π to each worker. If a fraction µ of workers are active and conditionally on daily state A = a, the expected amount of demand served by any worker if they become active is qa(µ). Workers have random outside options Bi such that, given the distribution π, the i-th worker is active with probability fBi (pi q(µ(π))) = 1/ (1 + exp [−β (pi q(µ(π)) − Bi )]) . Note: the expected revenue of the i-th worker is pi q(µ(π)). The system is in equilibrium, i.e., the fraction of active workers is µa(π) = E fBi (pi qA(µ(π))) A = a . NB: The distribution of outside options Bi may depend on state A.
  • 18. Mean-field analysis We adopt an asymptotic setting with n → ∞ workers who could potentially become active. Fact 1: Given the choice of payment distribution π, an equilibrium with µa(π) = E fBi (pi q(µA(π))) A = a exists and is unique. The number of active workers has a binomial(µa(π), n) distribution. Fact 2: As n → ∞, the equilibrium (and relevant derivatives) converge to a mean-field limit.
  • 19. Mean-field analysis Fact 3: Recall our local experiment where we independently perturb each worker’s payment by a small random amount, pi = p + ζnεi , εi iid ∼ {±1} . Write Zi for whether the i-th worker gets active, and estimate ∆ ← 1 ζn OLS (Zi ∼ εi ) . Then, if ζn → 0 and ζn √ n → ∞, ∆ →p ∆a(p) = q(µa(p))E fBi (pq(µA(p))) A = a , and we refer to ∆a(p) as the marginal response function.
  • 20. Mean-field analysis Fact 4: Under out assumptions, the marginal response function ∆ and the supply response dµ(p)/dp are linked via the system dµa(p) dp = ∆a(p) − p∆a(p) da µ2 a(p) ω (da/µa(p)) ω(da/µa(p)) dµa(p) dp . Apart from ∆(p), all other quantities in this equation, da and µa(p), can be readily observed. Theorem. The local experimentation strategy outlined above consistently recovers dµa(p)/dp as n → ∞.
  • 21. Learning via Local Experimentation The ultimate goal of the platform is to maximize its utility U, for our purposes taken as total cost minus total revenue. Write γ for the platform’s revenue per unit of demand served. In the mean-field limit, the utility then converges to n−1 Ua(p) = (γ − p) ω(da/µa(p)) µa(p), U(p) = E [UA(p)] . Once we know dµa(p)/dp, working out the utility derivative dUa(p)/dp amounts to calculus. We consider a platform that uses these estimates to optimize U(p) by gradient descent (or rather ascent).
  • 22. A First-Order Algorithm We now proceed to optimize payments via a variant of mirror descent Specify a step size η, an interval I = [c−, c+], and an initial payment p1. Then, at time period t = 1, 2, ...: 1. Deploy randomized payment perturbations εit around pt. 2. Estimate ∆ by regressing market participation on εit. 3. Translate this into an estimate Γt of dUAt (p)/dp via the transformation implied by the mean-field limit. 4. Perform a gradient update, where θt = t s=1 sΓs: pt+1 = argminp 1 2η t s=1 s(p − ps)2 − θtp : p ∈ I If the Ua(p) functions are strongly concave, this attains a 1/t rate of convergence in large markets, both in regret and squared error.
  • 23. A First-Order Algorithm If the Ua(p) functions are strongly concave, this attains a 1/t rate of convergence in large markets, both in regret and squared error. Theorem. If the Ua(p) functions are σ-strongly concave, |ua(p)| ≤ M, and we use a step size η > σ−1 then lim n→∞ P 1 T T t=1 t (UAt (p) − UAt (pt)) ≤ ηM2 2 = 1, for any fixed payment p ∈ [c−, c+]. Corollary. If in addition the day-specific states At are IID, then lim sup n→∞ P (p∗ − ¯pT )2 ≤ ηM2 σT 16 log δ−1 + 4 ≥ 1 − δ, p∗ = argmax {E [UA(p)] : p ∈ I} and ¯pT = 2 T(T+1) T t=1 t pt.
  • 24. Comparison with global experimentation Conceptually, our problem is closely related to the literature on continuous-armed bandits, motivated by the following setting: In each time period, the analyst deploys pt, and observes a reward Ut = U(pt) + noise. We want to control regret T−1 T t=1 (U(p∗) − U(pt)). Some references include Bubeck et al. (2017), Flaxman et al. (2005), Kleinberg (2005) and Shamir (2013). The optimal regret in this problem scales as 1/ √ T, even if we know U(p) is quadratic (Shamir, 2013).
  • 25. Comparison with global experimentation Here, instead, the gradients we get via our approach enable a 1/T rate of convergence. In other words, if local experimentation is applicable it fundamentally changes the difficulty of the problem relative to the continuous-armed bandits setting. The gain from local experimentation is comparable to the gain we could get from running two function evaluations with the same noise (Duchi et al., 2015).
  • 26. Extensions via generalized earning functions The core assumption that enables our approach is that workers care only about expected revenue, and thus respond to payments pi and market-level congestion q(µa(π)) via their product. Then, we showed that the mean-field limit is characterized by the following balance condition. µa(π) = E fBi (pi q(µA(π))) A = a . The form of this balance condition is crucial: If fB can have a generic dependence on pi and q, we may run into intractable difficulties.
  • 27. Extensions via generalized earning functions One way to generalize this setting is to let workers respond to pi and q via a (known) generalized earning function (GEF) θ, µa(π) = E fBi (θ(pi , q(µA(π)))) A = a . Example: Risk aversion. Workers respond to the expectation of a concave function of revenue. In the binary case where each worker serves 0 or 1 units of demand, we get θ(p, q) = β(p)q for some concave β(·). Example: Supply-side surge pricing. The platform commits to paying the i-th worker s(D/T)pi for some increasing surge multiplier s(·). Surge is automatic and anticipated by workers. With surge, the mean-field limit of expected revenue of the i-th worker is θ(p, q) = pqs(ω−1(q)).
  • 28. Extensions via generalized earning functions One way to generalize this setting is to let workers respond to pi and q via a (known) generalized earning function (GEF) θ, µa(π) = E fBi (θ(pi , q(µA(π)))) A = a . With GEF, the balance condition implies that a marginal response function can be estimated via local perturbations: ∆a(p) = pθ(p, q(µa(p)))E fBi (θ(p, q(µA(p)))) A = a . Then, dµa(p)/dp can be linked to ∆a(p) via a linear system that depends on system dynamics, thus enabling local experimentation: dµa(p) dp = p + q (µa(p)) dµa(p) dp q θ(p, q(µa(p))) E fBi (θ(p, q(µA(p)))) A = a . NB: The above are conjectures; no formal results yet with GEF.
  • 29. Simulation study: surge pricing 0.0 0.2 0.4 0.6 0.8 0.00.20.40.60.8 in−sample mean regret futureexpectedregret q q q q q q q q q q local experimentation global experimentation 0.0 0.2 0.4 0.6 0.8 0.00.20.40.60.8 in−sample mean regret futureexpectedregret q q q q q q q q q q local experimentation global experimentation The left panel is the simulation experiment from the beginning. The right panel shows results with an extension of our method that allows for surge pricing.
  • 30. Most work on experimental design assumes no interference, but this assumption often fails in a marketplace setting. We showed, however, that in some cases we can correct for interference with better power using some light-weight modeling. exposure graph mechanism graph cutting sparse and known arbitrary model based complete mean-field game There are more open questions than closed ones.