Download - Machine Learning & Mechanism Design
![Page 1: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/1.jpg)
Machine Learning & Mechanism Design:Dynamic and Discriminatory Pricing in Auctions
Jason D. HartlineMicrosoft Research – Silicon Valley
(Joint with with Maria-Florina Balcan,
Avrim Blum, and Yishay Mansour)
August 5, 2005
![Page 2: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/2.jpg)
The Problem
Sellers can extract more of surplus with discriminatory pricing.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20051
![Page 3: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/3.jpg)
The Problem
Sellers can extract more of surplus with discriminatory pricing.
Two approaches:
1. Distinguish between products.(E.g., software versioning, airline tickets, etc.)
2. Price discriminate with observable customer features.(E.g., college tuition, DVDs, car insurance, shipping)
DISCRIMINATORY AUCTIONS – AUGUST 5, 20051
![Page 4: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/4.jpg)
The Problem
Sellers can extract more of surplus with discriminatory pricing.
Two approaches:
1. Distinguish between products.(E.g., software versioning, airline tickets, etc.)
2. Price discriminate with observable customer features.(E.g., college tuition, DVDs, car insurance, shipping)
Goal: design mechanism to optimally price discriminate.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20051
![Page 5: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/5.jpg)
Optimal Mechanism Design
Typical Economic approach to optimal mechanism design:
• Assume valuations are from known distribution.
• Design optimal auction for distribution.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20052
![Page 6: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/6.jpg)
Optimal Mechanism Design
Typical Economic approach to optimal mechanism design:
• Assume valuations are from known distribution.
• Design optimal auction for distribution.
Notes on optimal mechanism design problem:
• Solved by Myerson (for single-parameter case).
• non-identical distributions =⇒ discriminatory pricing.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20052
![Page 7: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/7.jpg)
Optimal Mechanism Design
Typical Economic approach to optimal mechanism design:
• Assume valuations are from known distribution.
• Design optimal auction for distribution.
Notes on optimal mechanism design problem:
• Solved by Myerson (for single-parameter case).
• non-identical distributions =⇒ discriminatory pricing.
• Assumed known distribution ignores:
– incentives (of acquiring distribution)
– performance (from inaccurate distribution)
DISCRIMINATORY AUCTIONS – AUGUST 5, 20052
![Page 8: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/8.jpg)
Optimal Mechanism Design
Typical Economic approach to optimal mechanism design:
• Assume valuations are from known distribution.
• Design optimal auction for distribution.
Notes on optimal mechanism design problem:
• Solved by Myerson (for single-parameter case).
• non-identical distributions =⇒ discriminatory pricing.
• Assumed known distribution ignores:
– incentives (of acquiring distribution)
– performance (from inaccurate distribution)
Goal: understand how quality and incentives of learning distributionaffect profit.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20052
![Page 9: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/9.jpg)
Setting
1. Unlimited supply of stuff to sell.
2. bidders with private valuations for stuff.
3. make each bidder an offer.
4. revenue is incentive compatible function of offer and valuation.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 10: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/10.jpg)
Setting
1. Unlimited supply of stuff to sell.(Example 1: MS Office Professional (PV) & Student Version (SV))
2. bidders with private valuations for stuff.
3. make each bidder an offer.
4. revenue is incentive compatible function of offer and valuation.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 11: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/11.jpg)
Setting
1. Unlimited supply of stuff to sell.(Example 1: MS Office Professional (PV) & Student Version (SV))
2. bidders with private valuations for stuff.(Example 1: Bidder: “PV worth $400, SV worth $300”)
3. make each bidder an offer.
4. revenue is incentive compatible function of offer and valuation.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 12: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/12.jpg)
Setting
1. Unlimited supply of stuff to sell.(Example 1: MS Office Professional (PV) & Student Version (SV))
2. bidders with private valuations for stuff.(Example 1: Bidder: “PV worth $400, SV worth $300”)
3. make each bidder an offer.(Example 1: Seller: “PV costs $369.88, SV costs $124.99”)
4. revenue is incentive compatible function of offer and valuation.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 13: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/13.jpg)
Setting
1. Unlimited supply of stuff to sell.(Example 1: MS Office Professional (PV) & Student Version (SV))
2. bidders with private valuations for stuff.(Example 1: Bidder: “PV worth $400, SV worth $300”)
3. make each bidder an offer.(Example 1: Seller: “PV costs $369.88, SV costs $124.99”)
4. revenue is incentive compatible function of offer and valuation.(Example 1: Sold: SV for $124.99!)
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 14: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/14.jpg)
Setting
1. Unlimited supply of stuff to sell.(Example 1: MS Office Professional (PV) & Student Version (SV))(Example 2: Tuition for in state (IS) and out of state (OS) students)
2. bidders with private valuations for stuff.(Example 1: Bidder: “PV worth $400, SV worth $300”)
3. make each bidder an offer.(Example 1: Seller: “PV costs $369.88, SV costs $124.99”)
4. revenue is incentive compatible function of offer and valuation.(Example 1: Sold: SV for $124.99!)
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 15: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/15.jpg)
Setting
1. Unlimited supply of stuff to sell.(Example 1: MS Office Professional (PV) & Student Version (SV))(Example 2: Tuition for in state (IS) and out of state (OS) students)
2. bidders with private valuations for stuff.(Example 1: Bidder: “PV worth $400, SV worth $300”)(Example 2: Bidder (OS): “Tuition worth $15,000”)
3. make each bidder an offer.(Example 1: Seller: “PV costs $369.88, SV costs $124.99”)
4. revenue is incentive compatible function of offer and valuation.(Example 1: Sold: SV for $124.99!)
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 16: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/16.jpg)
Setting
1. Unlimited supply of stuff to sell.(Example 1: MS Office Professional (PV) & Student Version (SV))(Example 2: Tuition for in state (IS) and out of state (OS) students)
2. bidders with private valuations for stuff.(Example 1: Bidder: “PV worth $400, SV worth $300”)(Example 2: Bidder (OS): “Tuition worth $15,000”)
3. make each bidder an offer.(Example 1: Seller: “PV costs $369.88, SV costs $124.99”)(Example 2: Seller: “IS costs $9,256.80, OS costs $16,855.30,”)
4. revenue is incentive compatible function of offer and valuation.(Example 1: Sold: SV for $124.99!)
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 17: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/17.jpg)
Setting
1. Unlimited supply of stuff to sell.(Example 1: MS Office Professional (PV) & Student Version (SV))(Example 2: Tuition for in state (IS) and out of state (OS) students)
2. bidders with private valuations for stuff.(Example 1: Bidder: “PV worth $400, SV worth $300”)(Example 2: Bidder (OS): “Tuition worth $15,000”)
3. make each bidder an offer.(Example 1: Seller: “PV costs $369.88, SV costs $124.99”)(Example 2: Seller: “IS costs $9,256.80, OS costs $16,855.30,”)
4. revenue is incentive compatible function of offer and valuation.(Example 1: Sold: SV for $124.99!)(Example 2: No Sale!)
DISCRIMINATORY AUCTIONS – AUGUST 5, 20053
![Page 18: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/18.jpg)
Overview
1.=⇒ Auction Problem
(a) Random Sampling Solution
(b) Retrospective bounds.
(c) Software Versioning Example.
2. Online Auction Problem
(a) Expert Learning based Auction.
(b) Expert Learning with non-uniform bounds.
3. Conclusions
DISCRIMINATORY AUCTIONS – AUGUST 5, 20054
![Page 19: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/19.jpg)
Auction Problem
The Unlimited Supply Auction Problem:
Given:
• unlimited supply of stuff.
• Set S of n bidders with valuations for stuff.
• class G of reasonable offers.
Design: Auction with profit near that of optimal single offer.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20055
![Page 20: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/20.jpg)
Auction Problem
The Unlimited Supply Auction Problem:
Given:
• unlimited supply of stuff.
• Set S of n bidders with valuations for stuff.
• class G of reasonable offers.
Design: Auction with profit near that of optimal single offer.
Notation:
• g(i) = payoff from bidder i when offered g.
• g(S) =∑
i∈S g(i).
• optG(S) = argmaxg∈G g(S).
• OPTG(S) = maxg∈G g(S).
DISCRIMINATORY AUCTIONS – AUGUST 5, 20055
![Page 21: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/21.jpg)
Random Sampling Auction
Random Sampling Optimal Offer Auction, RSOOG
1. Randomly partition bidders into two sets: S1 and S2.
2. compute g1 (resp. g2), optimal offer for S1 (resp. S2)
3. Offer g1 to S2 and g2 to S1.
S
(Random Sampling Auction from [Goldberg, Hartline, Wright 2001])
DISCRIMINATORY AUCTIONS – AUGUST 5, 20056
![Page 22: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/22.jpg)
Random Sampling Auction
Random Sampling Optimal Offer Auction, RSOOG
1. Randomly partition bidders into two sets: S1 and S2.
2. compute g1 (resp. g2), optimal offer for S1 (resp. S2)
3. Offer g1 to S2 and g2 to S1.
SS1
S2
(Random Sampling Auction from [Goldberg, Hartline, Wright 2001])
DISCRIMINATORY AUCTIONS – AUGUST 5, 20056
![Page 23: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/23.jpg)
Random Sampling Auction
Random Sampling Optimal Offer Auction, RSOOG
1. Randomly partition bidders into two sets: S1 and S2.
2. compute g1 (resp. g2), optimal offer for S1 (resp. S2)
3. Offer g1 to S2 and g2 to S1.
SS1
S2
g1 = opt(S1)
g2 = opt(S2)
(Random Sampling Auction from [Goldberg, Hartline, Wright 2001])
DISCRIMINATORY AUCTIONS – AUGUST 5, 20056
![Page 24: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/24.jpg)
Random Sampling Auction
Random Sampling Optimal Offer Auction, RSOOG
1. Randomly partition bidders into two sets: S1 and S2.
2. compute g1 (resp. g2), optimal offer for S1 (resp. S2)
3. Offer g1 to S2 and g2 to S1.
SS1
S2
g1 = opt(S1)
g2 = opt(S2)
g1 = opt(S1)
g2 = opt(S2)
(Random Sampling Auction from [Goldberg, Hartline, Wright 2001])
DISCRIMINATORY AUCTIONS – AUGUST 5, 20056
![Page 25: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/25.jpg)
Random Sampling Auction
Random Sampling Optimal Offer Auction, RSOOG
1. Randomly partition bidders into two sets: S1 and S2.
2. compute g1 (resp. g2), optimal offer for S1 (resp. S2)
3. Offer g1 to S2 and g2 to S1.
SS1
S2
g1 = opt(S1)
g2 = opt(S2)
g1 = opt(S1)
g2 = opt(S2)
Question: when is RSOOG good?
(Random Sampling Auction from [Goldberg, Hartline, Wright 2001])
DISCRIMINATORY AUCTIONS – AUGUST 5, 20056
![Page 26: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/26.jpg)
Performance Analysis
Lemma: For g and random partitions S1 and S2:
Pr[|g(S1) − g(S2)| > εmax(p, g(S))] ≤ 2e−ε2p/2h.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20057
![Page 27: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/27.jpg)
Performance Analysis
Lemma: For g and random partitions S1 and S2:
Pr[|g(S1) − g(S2)| > εmax(p, g(S))] ≤ 2e−ε2p/2h.
Consider:
• Use p = OPTG .
• If |G| e−ε2 OPTG /2h ≤ δ,
• union bound probability any g ∈ G is bad by δ.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20057
![Page 28: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/28.jpg)
Performance Analysis
Lemma: For g and random partitions S1 and S2:
Pr[|g(S1) − g(S2)| > εmax(p, g(S))] ≤ 2e−ε2p/2h.
Consider:
• Use p = OPTG .
• If |G| e−ε2 OPTG /2h ≤ δ,
• union bound probability any g ∈ G is bad by δ.
Theorem: With probability 1 − δ profit from RSOOG is at least
(1 − ε)OPTG −O( hε2 log |G|
δ )
DISCRIMINATORY AUCTIONS – AUGUST 5, 20057
![Page 29: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/29.jpg)
Performance Analysis
Lemma: For g and random partitions S1 and S2:
Pr[|g(S1) − g(S2)| > εmax(p, g(S))] ≤ 2e−ε2p/2h.
Consider:
• Use p = OPTG .
• If |G| e−ε2 OPTG /2h ≤ δ,
• union bound probability any g ∈ G is bad by δ.
Theorem: With probability 1 − δ profit from RSOOG is at least
(1 − ε)OPTG −O( hε2 log |G|
δ )
Interpretation: O(h log |G|) is convergence time.
DISCRIMINATORY AUCTIONS – AUGUST 5, 20057
![Page 30: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/30.jpg)
Example
Example: Selling tee shirts. (discretized prices)
• Bidders with valuations in [1, h] for a tee shirt.
• Reasonable offers: G = {price 2i for i ∈ {1, . . . , log h}}.
• Convergence Time: O(h log |G|) = O(h log log h)
DISCRIMINATORY AUCTIONS – AUGUST 5, 20058
![Page 31: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/31.jpg)
Overview
1. Auction Problem
(a) Random Sampling Solution
(b)=⇒ Retrospective bounds.
(c) Software Versioning Example.
2. Online Auction Problem
(a) Expert Learning based Auction.
(b) Expert Learning with non-uniform bounds.
3. Conclusions
DISCRIMINATORY AUCTIONS – AUGUST 5, 20059
![Page 32: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/32.jpg)
|G| = ∞?
What if |G| = ∞?
DISCRIMINATORY AUCTIONS – AUGUST 5, 200510
![Page 33: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/33.jpg)
|G| = ∞?
What if |G| = ∞?
Example: Selling tee shirts. (non-discretized prices)
• Bidders with valuations v1, . . . , vn in [1, h] for a tee shirt.
• Reasonable offers: G = {price p ∈ [1, h]}.
• Convergence Time:
DISCRIMINATORY AUCTIONS – AUGUST 5, 200510
![Page 34: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/34.jpg)
|G| = ∞?
What if |G| = ∞?
Example: Selling tee shirts. (non-discretized prices)
• Bidders with valuations v1, . . . , vn in [1, h] for a tee shirt.
• Reasonable offers: G = {price p ∈ [1, h]}.
• Convergence Time:
Observation:
• Suppose RSOOG on S only offers g ∈ GS ⊂ G.
• Then RSOOGS(S) is same as RSOOG(S).
• Retrospectively perform analysis on GS instead of G.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200510
![Page 35: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/35.jpg)
|G| = ∞?
What if |G| = ∞?
Example: Selling tee shirts. (non-discretized prices)
• Bidders with valuations v1, . . . , vn in [1, h] for a tee shirt.
• Reasonable offers: G = {price p ∈ [1, h]}.
• Convergence Time:
Observation:
• Suppose RSOOG on S only offers g ∈ GS ⊂ G.
• Then RSOOGS(S) is same as RSOOG(S).
• Retrospectively perform analysis on GS instead of G.
Consider: GS = {“offer vi” : i ∈ S}.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200510
![Page 36: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/36.jpg)
|G| = ∞?
What if |G| = ∞?
Example: Selling tee shirts. (non-discretized prices)
• Bidders with valuations v1, . . . , vn in [1, h] for a tee shirt.
• Reasonable offers: G = {price p ∈ [1, h]}.
• Convergence Time: O(h |GS |) = O(h log n)
Observation:
• Suppose RSOOG on S only offers g ∈ GS ⊂ G.
• Then RSOOGS(S) is same as RSOOG(S).
• Retrospectively perform analysis on GS instead of G.
Consider: GS = {“offer vi” : i ∈ S}.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200510
![Page 37: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/37.jpg)
Example: Software Versioning
Example: Software versioning.
• Microsoft has M possible versions of MS Office,
• but can only package and sell m � M distinct versions.(E.g., Student and Professional).
DISCRIMINATORY AUCTIONS – AUGUST 5, 200511
![Page 38: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/38.jpg)
Example: Software Versioning
Example: Software versioning.
• Microsoft has M possible versions of MS Office,
• but can only package and sell m � M distinct versions.(E.g., Student and Professional).
• Reasonable offers: G = {p ∈ (R ∪ {∞})M with pj = ∞ for allbut m items}.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200511
![Page 39: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/39.jpg)
Example: Software Versioning
Example: Software versioning.
• Microsoft has M possible versions of MS Office,
• but can only package and sell m � M distinct versions.(E.g., Student and Professional).
• Reasonable offers: G = {p ∈ (R ∪ {∞})M with pj = ∞ for allbut m items}.
• What is |GS |?
DISCRIMINATORY AUCTIONS – AUGUST 5, 200511
![Page 40: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/40.jpg)
Example: Software Versioning
Example: Software versioning.
• Microsoft has M possible versions of MS Office,
• but can only package and sell m � M distinct versions.(E.g., Student and Professional).
• Reasonable offers: G = {p ∈ (R ∪ {∞})M with pj = ∞ for allbut m items}.
• What is |GS |?
Lemma: |GS | = O((nm2M)m).
DISCRIMINATORY AUCTIONS – AUGUST 5, 200511
![Page 41: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/41.jpg)
Example: Software Versioning
Example: Software versioning.
• Microsoft has M possible versions of MS Office,
• but can only package and sell m � M distinct versions.(E.g., Student and Professional).
• Reasonable offers: G = {p ∈ (R ∪ {∞})M with pj = ∞ for allbut m items}.
• What is |GS |?
Lemma: |GS | = O((nm2M)m).
Convergence time = O(hm log(nm2M)) .
DISCRIMINATORY AUCTIONS – AUGUST 5, 200511
![Page 42: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/42.jpg)
Example: Software Versioning
Example: Software versioning.
• Microsoft has M possible versions of MS Office,
• but can only package and sell m � M distinct versions.(E.g., Student and Professional).
• Reasonable offers: G = {p ∈ (R ∪ {∞})M with pj = ∞ for allbut m items}.
• What is |GS |?
Lemma: |GS | = O((nm2M)m).
Convergence time = O(hm log(nm2M)) ≈ O(hm log n).
DISCRIMINATORY AUCTIONS – AUGUST 5, 200511
![Page 43: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/43.jpg)
Proof of Lemma
Lemma: |GS | = O((nm2M)m).
DISCRIMINATORY AUCTIONS – AUGUST 5, 200512
![Page 44: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/44.jpg)
Proof of Lemma
Lemma: |GS | = O((nm2M)m).
Proof:
1. Fix set of m items to sell.
2. Bidder i’s valuation divides price space into m + 1 convex regions.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200512
![Page 45: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/45.jpg)
Proof of Lemma
Lemma: |GS | = O((nm2M)m).
Proof:
1. Fix set of m items to sell.
2. Bidder i’s valuation divides price space into m + 1 convex regions.
3. Regions are joined by (m + 1)2 hyperplanes.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200512
![Page 46: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/46.jpg)
Proof of Lemma
Lemma: |GS | = O((nm2M)m).
Proof:
1. Fix set of m items to sell.
2. Bidder i’s valuation divides price space into m + 1 convex regions.
3. Regions are joined by (m + 1)2 hyperplanes.
4. n bidders total for n(m + 1)2 hyperplanes.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200512
![Page 47: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/47.jpg)
Proof of Lemma
Lemma: |GS | = O((nm2M)m).
Proof:
1. Fix set of m items to sell.
2. Bidder i’s valuation divides price space into m + 1 convex regions.
3. Regions are joined by (m + 1)2 hyperplanes.
4. n bidders total for n(m + 1)2 hyperplanes.
5. RSOOG offer price must be at intersection of hyperplanes.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200512
![Page 48: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/48.jpg)
Proof of Lemma
Lemma: |GS | = O((nm2M)m).
Proof:
1. Fix set of m items to sell.
2. Bidder i’s valuation divides price space into m + 1 convex regions.
3. Regions are joined by (m + 1)2 hyperplanes.
4. n bidders total for n(m + 1)2 hyperplanes.
5. RSOOG offer price must be at intersection of hyperplanes.
6. K = n(m + 1)2 hyperplanes in m dimensions intersect in Km.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200512
![Page 49: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/49.jpg)
Proof of Lemma
Lemma: |GS | = O((nm2M)m).
Proof:
1. Fix set of m items to sell.
2. Bidder i’s valuation divides price space into m + 1 convex regions.
3. Regions are joined by (m + 1)2 hyperplanes.
4. n bidders total for n(m + 1)2 hyperplanes.
5. RSOOG offer price must be at intersection of hyperplanes.
6. K = n(m + 1)2 hyperplanes in m dimensions intersect in Km.
7. Sum over Mm possible m-item sets.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200512
![Page 50: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/50.jpg)
Other Results
See paper for details on:
• Bounds for RSOOG for item-pricing in combinatorial auctions.
• Bounds for RSOOG on bidders with observable features.
• Better bounds with ε-covers of G.
• Better random sampling auction with structural risk minimization.
• Using approximation algorithms in RSOOG .
DISCRIMINATORY AUCTIONS – AUGUST 5, 200513
![Page 51: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/51.jpg)
Overview
1. Auction Problem
(a) Random Sampling Solution
(b) Retrospective bounds.
(c) Software Versioning Example.
2.=⇒ Online Auction Problem
(a) Expert Learning based Auction.
(b) Expert Learning with non-uniform bounds.
3. Conclusions
DISCRIMINATORY AUCTIONS – AUGUST 5, 200514
![Page 52: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/52.jpg)
Online Auction Problem
Online Auction Problem:
• unlimited supply of stuff.
• class G of reasonable offers.
• Bidders arrive one at a time and place bids, b1, b2, . . .
• Auctioneer makes offer g from G before next bidder arrives.
• Goal: Auction with profit close to optimal single offer.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200515
![Page 53: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/53.jpg)
Online Auction Problem
Online Auction Problem:
• unlimited supply of stuff.
• class G of reasonable offers.
• Bidders arrive one at a time and place bids, b1, b2, . . .
• Auctioneer makes offer g from G before next bidder arrives.
• Goal: Auction with profit close to optimal single offer.
Two Difficulties:
1. Incentive Compatibility requirement:
2. Online Requirement (do not know future):
DISCRIMINATORY AUCTIONS – AUGUST 5, 200515
![Page 54: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/54.jpg)
Online Auction Problem
Online Auction Problem:
• unlimited supply of stuff.
• class G of reasonable offers.
• Bidders arrive one at a time and place bids, b1, b2, . . .
• Auctioneer makes offer g from G before next bidder arrives.
• Goal: Auction with profit close to optimal single offer.
Two Difficulties:
1. Incentive Compatibility requirement:
offer to bidder i not function of bi.
2. Online Requirement (do not know future):
DISCRIMINATORY AUCTIONS – AUGUST 5, 200515
![Page 55: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/55.jpg)
Online Auction Problem
Online Auction Problem:
• unlimited supply of stuff.
• class G of reasonable offers.
• Bidders arrive one at a time and place bids, b1, b2, . . .
• Auctioneer makes offer g from G before next bidder arrives.
• Goal: Auction with profit close to optimal single offer.
Two Difficulties:
1. Incentive Compatibility requirement:
offer to bidder i not function of bi.
2. Online Requirement (do not know future):
price offered bidder i not function of future bids.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200515
![Page 56: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/56.jpg)
Online Auction Problem
Online Auction Problem:
• unlimited supply of stuff.
• class G of reasonable offers.
• Bidders arrive one at a time and place bids, b1, b2, . . .
• Auctioneer makes offer g from G before next bidder arrives.
• Goal: Auction with profit close to optimal single offer.
Two Difficulties:
1. Incentive Compatibility requirement:
offer to bidder i not function of bi.
2. Online Requirement (do not know future):
price offered bidder i not function of future bids.
Conclusion: offer for bidder i based only on prior bids: b1, . . . , bi−1.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200515
![Page 57: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/57.jpg)
Assumptions
1. We learn each bidders full valuation.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200516
![Page 58: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/58.jpg)
Assumptions
1. We learn each bidders full valuation.
for partial information case see multi-armed bandit solutions:[Blum, Kumar, Rudra, Wu ’03][Kleinberg, Leighton ’03][Blum, Hartline 05]
DISCRIMINATORY AUCTIONS – AUGUST 5, 200516
![Page 59: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/59.jpg)
Assumptions
1. We learn each bidders full valuation.
for partial information case see multi-armed bandit solutions:[Blum, Kumar, Rudra, Wu ’03][Kleinberg, Leighton ’03][Blum, Hartline 05]
2. Bidders cannot come back.
3. Bidders cannot lie about their arrival time.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200516
![Page 60: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/60.jpg)
Assumptions
1. We learn each bidders full valuation.
for partial information case see multi-armed bandit solutions:[Blum, Kumar, Rudra, Wu ’03][Kleinberg, Leighton ’03][Blum, Hartline 05]
2. Bidders cannot come back.
3. Bidders cannot lie about their arrival time.
for temporal strategyproofness see: [Hajiaghayi, Kleinberg, Parkes ’04]
DISCRIMINATORY AUCTIONS – AUGUST 5, 200516
![Page 61: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/61.jpg)
Assumptions
1. We learn each bidders full valuation.
for partial information case see multi-armed bandit solutions:[Blum, Kumar, Rudra, Wu ’03][Kleinberg, Leighton ’03][Blum, Hartline 05]
2. Bidders cannot come back.
3. Bidders cannot lie about their arrival time.
for temporal strategyproofness see: [Hajiaghayi, Kleinberg, Parkes ’04]
4. items in unlimited supply.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200516
![Page 62: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/62.jpg)
Assumptions
1. We learn each bidders full valuation.
for partial information case see multi-armed bandit solutions:[Blum, Kumar, Rudra, Wu ’03][Kleinberg, Leighton ’03][Blum, Hartline 05]
2. Bidders cannot come back.
3. Bidders cannot lie about their arrival time.
for temporal strategyproofness see: [Hajiaghayi, Kleinberg, Parkes ’04]
4. items in unlimited supply.
for limited supply see:[Hajiaghayi, Kleinberg, Parkes ’04][Kleinberg ’05]
DISCRIMINATORY AUCTIONS – AUGUST 5, 200516
![Page 63: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/63.jpg)
Online Learning
Expert Online Learning Problem:
In round i:
1. Each of k experts propose a strategy.
2. We choose an expert’s strategy.
3. Find out how each strategy performed (payoff)
Goal: Obtain payoff close to single best expert overall (in hindsight).
DISCRIMINATORY AUCTIONS – AUGUST 5, 200517
![Page 64: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/64.jpg)
Online Learning
Expert Online Learning Problem:
In round i:
1. Each of k experts propose a strategy.
2. We choose an expert’s strategy.
3. Find out how each strategy performed (payoff)
Goal: Obtain payoff close to single best expert overall (in hindsight).
Weighted Majority Algorithm: (for round i)
Let h be maximum payoff. For expert j, let sj be total payoff thus far.
Choose expert j ’s strategy with probability proportional to (1+2ε)sj/h.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200517
![Page 65: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/65.jpg)
Online Learning
Expert Online Learning Problem:
In round i:
1. Each of k experts propose a strategy.
2. We choose an expert’s strategy.
3. Find out how each strategy performed (payoff)
Goal: Obtain payoff close to single best expert overall (in hindsight).
Weighted Majority Algorithm: (for round i)
Let h be maximum payoff. For expert j, let sj be total payoff thus far.
Choose expert j ’s strategy with probability proportional to (1+2ε)sj/h.
Result: E[payoff] = (1 − ε)OPT− h2ε log k.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200517
![Page 66: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/66.jpg)
Application to Online Auctions
Application: (to online auctions) [Blum Kumar Rudra Wu 2003]
DISCRIMINATORY AUCTIONS – AUGUST 5, 200518
![Page 67: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/67.jpg)
Application to Online Auctions
Application: (to online auctions) [Blum Kumar Rudra Wu 2003]
1. Expert for each g ∈ G
DISCRIMINATORY AUCTIONS – AUGUST 5, 200518
![Page 68: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/68.jpg)
Application to Online Auctions
Application: (to online auctions) [Blum Kumar Rudra Wu 2003]
1. Expert for each g ∈ G
2. Best expert ⇒ best offer.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200518
![Page 69: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/69.jpg)
Application to Online Auctions
Application: (to online auctions) [Blum Kumar Rudra Wu 2003]
1. Expert for each g ∈ G
2. Best expert ⇒ best offer.
Result: E[profit] = (1 − ε)OPTG −hε log |G|.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200518
![Page 70: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/70.jpg)
Application to Online Auctions
Application: (to online auctions) [Blum Kumar Rudra Wu 2003]
1. Expert for each g ∈ G
2. Best expert ⇒ best offer.
Result: E[profit] = (1 − ε)OPTG −hε log |G|.
Note: Same convergence time as for RSOOG .
DISCRIMINATORY AUCTIONS – AUGUST 5, 200518
![Page 71: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/71.jpg)
Example
Example: Selling tee shirts. (discretized prices)
• Bidders with valuations in [1, h] for a tee shirt.
• Reasonable offers: G = {price 2i for i ∈ {1, . . . , log h}}.
• Convergence Time: O(h log |G|)
DISCRIMINATORY AUCTIONS – AUGUST 5, 200519
![Page 72: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/72.jpg)
Example
Example: Selling tee shirts. (discretized prices)
• Bidders with valuations in [1, h] for a tee shirt.
• Reasonable offers: G = {price 2i for i ∈ {1, . . . , log h}}.
• Convergence Time: O(h log |G|)= O(h log log h).
DISCRIMINATORY AUCTIONS – AUGUST 5, 200519
![Page 73: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/73.jpg)
Better Bounds?
Can we get better bounds?
Retrospective technique like using GS does not work.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200520
![Page 74: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/74.jpg)
Overview
1. Auction Problem
(a) Random Sampling Solution
(b) Retrospective bounds.
(c) Software Versioning Example.
2. Online Auction Problem
(a) Expert Learning based Auction.
(b)=⇒ Expert Learning with non-uniform bounds.
3. Conclusions
DISCRIMINATORY AUCTIONS – AUGUST 5, 200521
![Page 75: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/75.jpg)
Non-uniform Bounds on Payoff
Expert Online Learning Problem: In round i:
1. Each of k experts propose a strategy.
2. We choose an expert’s strategy.
3. Find out how each strategy performed (payoff)
4. Expert i’s payoff is always less than hi.
Goal: Obtain payoff close to single best expert overall (in hindsight).
DISCRIMINATORY AUCTIONS – AUGUST 5, 200522
![Page 76: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/76.jpg)
Non-uniform Bounds on Payoff
Expert Online Learning Problem: In round i:
1. Each of k experts propose a strategy.
2. We choose an expert’s strategy.
3. Find out how each strategy performed (payoff)
4. Expert i’s payoff is always less than hi.
Goal: Obtain payoff close to single best expert overall (in hindsight).
Non-uniform Experts Algorithm: [Kalai ’01][Blum, Hartline ’05]
1. (initialization) For each expert, j, add initial score, sj , as:
hi × number of heads flipped in a row.
2. Run deterministic “go with best expert” algorithm.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200522
![Page 77: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/77.jpg)
Non-uniform Bounds on Payoff
Expert Online Learning Problem: In round i:
1. Each of k experts propose a strategy.
2. We choose an expert’s strategy.
3. Find out how each strategy performed (payoff)
4. Expert i’s payoff is always less than hi.
Goal: Obtain payoff close to single best expert overall (in hindsight).
Non-uniform Experts Algorithm: [Kalai ’01][Blum, Hartline ’05]
1. (initialization) For each expert, j, add initial score, sj , as:
hi × number of heads flipped in a row.
2. Run deterministic “go with best expert” algorithm.
Result: E[profit] ≥ OPT /2 −∑
i hi.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200522
![Page 78: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/78.jpg)
Application to Online Auctions
Application: (to online auctions)
DISCRIMINATORY AUCTIONS – AUGUST 5, 200523
![Page 79: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/79.jpg)
Application to Online Auctions
Application: (to online auctions)
1. Bound hg for each g ∈ G.
2. Expert for each g ∈ G
3. Best expert ⇒ best offer.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200523
![Page 80: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/80.jpg)
Application to Online Auctions
Application: (to online auctions)
1. Bound hg for each g ∈ G.
2. Expert for each g ∈ G
3. Best expert ⇒ best offer.
Result: E[profit] = OPTG /2 −∑
g∈G hg .
DISCRIMINATORY AUCTIONS – AUGUST 5, 200523
![Page 81: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/81.jpg)
Application to Online Auctions
Application: (to online auctions)
1. Bound hg for each g ∈ G.
2. Expert for each g ∈ G
3. Best expert ⇒ best offer.
Result: E[profit] = OPTG /2 −∑
g∈G hg .
Note: Convergence time =∑
g∈G hg
DISCRIMINATORY AUCTIONS – AUGUST 5, 200523
![Page 82: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/82.jpg)
Example
Example: Selling tee shirts. (discretized prices)
• Bidders with valuations in [1, h] for a tee shirt.
• Reasonable offers: G = {price 2i for i ∈ {1, . . . , log h}}.
• Convergence Time:∑
g∈G hg .
DISCRIMINATORY AUCTIONS – AUGUST 5, 200524
![Page 83: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/83.jpg)
Example
Example: Selling tee shirts. (discretized prices)
• Bidders with valuations in [1, h] for a tee shirt.
• Reasonable offers: G = {price 2i for i ∈ {1, . . . , log h}}.
• Convergence Time:∑
g∈G hg =∑log h
i 2i .
DISCRIMINATORY AUCTIONS – AUGUST 5, 200524
![Page 84: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/84.jpg)
Example
Example: Selling tee shirts. (discretized prices)
• Bidders with valuations in [1, h] for a tee shirt.
• Reasonable offers: G = {price 2i for i ∈ {1, . . . , log h}}.
• Convergence Time:∑
g∈G hg =∑log h
i 2i ≤ 2h.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200524
![Page 85: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/85.jpg)
Example
Example: Selling tee shirts. (discretized prices)
• Bidders with valuations in [1, h] for a tee shirt.
• Reasonable offers: G = {price 2i for i ∈ {1, . . . , log h}}.
• Convergence Time:∑
g∈G hg =∑log h
i 2i ≤ 2h.
Note: this is optimal up to constant factors.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200524
![Page 86: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/86.jpg)
Conclusions
1. Used machine learning techniques for auction design/analysis.
2. Prior-free discriminatory optimal mechanism design.
(a) distinguishing between products (and selecting products to sell).
(b) price discriminate based on observable customer features.
3. Similar bounds for offline and online auctions.
4. Retrospective analysis for offline auctions.
DISCRIMINATORY AUCTIONS – AUGUST 5, 200525
![Page 87: Machine Learning & Mechanism Design](https://reader035.vdocuments.mx/reader035/viewer/2022071600/613d1d07736caf36b759786c/html5/thumbnails/87.jpg)
Conclusions
1. Used machine learning techniques for auction design/analysis.
2. Prior-free discriminatory optimal mechanism design.
(a) distinguishing between products (and selecting products to sell).
(b) price discriminate based on observable customer features.
3. Similar bounds for offline and online auctions.
4. Retrospective analysis for offline auctions.
5. Open: ε-cover arguments for online auctions?
6. Open: limited supply?
7. Open: general cost function on outcomes?
DISCRIMINATORY AUCTIONS – AUGUST 5, 200525