Presentation is loading. Please wait.

Presentation is loading. Please wait.

Approximating Optimal Social Choice under Metric Preferences Elliot Anshelevich Onkar Bhardwaj John Postl Rensselaer Polytechnic Institute (RPI), Troy,

Similar presentations


Presentation on theme: "Approximating Optimal Social Choice under Metric Preferences Elliot Anshelevich Onkar Bhardwaj John Postl Rensselaer Polytechnic Institute (RPI), Troy,"— Presentation transcript:

1 Approximating Optimal Social Choice under Metric Preferences Elliot Anshelevich Onkar Bhardwaj John Postl Rensselaer Polytechnic Institute (RPI), Troy, NY

2 Voting and Social Choice m candidates/alternatives A, B, C, D, … n voters/agents: have preferences over alternatives Elections Recommender systems Search engines Preference aggregation

3 Voting and Social Choice m candidates/alternatives A, B, C, D, … n voters/agents: have preferences over alternatives Usually specify total order over alternatives Voting mechanism decides outcome given these preferences (e.g., which alternative is chosen; ranking of alternatives; etc) 1. A > B > C 2. A > B > C 3. A > B > C 4. B > A > C 5. B > A > C 6. C > A > B 7. C > A > B 8. C > A > B 9. C > A > B

4 Voting Mechanisms m candidates/alternatives A, B, C, D, … n voters/agents: have preferences over alternatives Usually specify total order over alternatives Majority/ Plurality does not work very well: C wins even though A pairwise preferred to C. E.g., Bush-Gore-Nader 1. A > B > C 2. A > B > C 3. A > B > C 4. B > A > C 5. B > A > C 6. C > A > B 7. C > A > B 8. C > A > B 9. C > A > B B A C

5 Voting Mechanisms m candidates/alternatives A, B, C, D, … n voters/agents: have preferences over alternatives Usually specify total order over alternatives Majority/ Plurality does not work very well: C wins even though A pairwise preferred to C. E.g., Bush-Gore-Nader 1. A > B > C 2. A > B > C 3. A > B > C 4. B > A > C 5. B > A > C 6. C > A > B 7. C > A > B 8. C > A > B 9. C > A > B B A C

6 Voting Mechanisms Condorcet Cycle 1. A > B > C 2. B > C > A 3. C > A > B B A C

7 Voting Mechanisms Condorcet Cycle So, what is “best” outcome? All voting mechanisms have weaknesses. “Axiomatic” approach: define some properties, see which mechanisms satisfy them 1. A > B > C 2. B > C > A 3. C > A > B B A C

8 Arrow’s Impossibility Theorem (1950) No mechanism for more than 2 alternatives can satisfy the following “reasonable” properties Formally, no mechanism obeys all 3 of following properties o Unanimity (if A preferred to B by all voters, than A should be ranked higher) o Independence of Irrelevant Alternatives (how A is ranked relative to B only depends on order of A and B in voter preferences) o Non-dictatorship (voting mechanism does not just do what one voter says) Common approaches o “Axiomatic” approach: analyze lots of different mechanisms, show good properties about each o Make extra assumptions on preferences (Nobel prize in economics)

9

10 Our Approach: Metric Preferences Metric preferences o Also called spatial preferences Additional structure on who prefers which alternative

11 Example: Political Spectrum Left Right BAC

12 Example: Political Spectrum

13

14 xkcd

15 Example: Political Spectrum xkcd Downsian proximity model (1957): Each dimension is a different issue

16 Our Model Voters and candidates are points in an arbitrary metric space Each voter prefers candidates closer to themselves Best alternative: min Σ d(i,A) A i B AC

17 Our Model Voters and candidates are points in an arbitrary metric space Each voter prefers candidates closer to themselves Best alternative: min Σ d(i,A) A i B AC B > A > C

18 Our Model Voters and candidates are points in an arbitrary metric space Each voter prefers candidates closer to themselves Best alternative: min Σ d(i,A) A i B AC

19 Our Model Voters and candidates are points in an arbitrary metric space Each voter prefers candidates closer to themselves Best alternative: Finding best alternative is easy min Σ d(i,A) A i B AC

20 Our Model Voters and candidates are points in an arbitrary metric space Each voter prefers candidates closer to themselves Best alternative: Usually don’t know numerical values! min Σ d(i,A) A i B AC

21 Our Model Given: Ordinal preferences of all voters These preferences come from an unknown arbitrary metric space Goal: Return best alternative 1. A > B > C 2. A > B > C 3. A > B > C 4. B > A > C 5. B > A > C 6. C > A > B 7. C > A > B 8. C > A > B 9. C > A > B............

22 Our Model Given: Ordinal preferences of all voters These preferences come from an unknown arbitrary metric space Goal: Return provably good approximation to the best alternative 1. A > B > C 2. A > B > C 3. A > B > C 4. B > A > C 5. B > A > C 6. C > A > B 7. C > A > B 8. C > A > B 9. C > A > B........ B = OPT AC Σ d(i,C) i Σ d(i,B) i small

23 Model Summary Given: Ordinal preferences p of all voters These preferences come from an unknown arbitrary metric space Want mechanism which has small distortion: 1. A > B > C 2. A > B > C 3. A > B > C 4. B > A > C 5. B > A > C 6. C > A > B 7. C > A > B 8. C > A > B 9. C > A > B........ Σ d(i,winner) i i max d ϵ D(p) A min Σ d(i,A) Approximate median using only ordinal information

24 Easy Example: 2 candidates 2 candidates o n-k voters have A > B o k voters have B > A

25 Easy Example: 2 candidates 2 candidates o n-k voters have A > B o k voters have B > A BA k n-k B may be optimal even if k=1

26 Easy Example: 2 candidates 2 candidates o n-k voters have A > B o k voters have B > A BA k n-k B may be optimal even if k=1 But, if use majority, then distortion ≤ 3

27 Easy Example: 2 candidates 2 candidates o n/2 voters have A > B o n/2 voters have B > A BA n/2 B may be optimal even if k=1 But, if use majority, then distortion ≤ 3 Also shows that no deterministic mechanism can have distortion < 3

28 Our Results SumMedian Plurality2m-1Unbounded Borda2m-1Unbounded k-approval2n-1Unbounded Veto2n-1Unbounded Copeland55 Uncovered Set55 Lower Bound35 Σ d(i,winner) i i max d ϵ D(p) A min Σ d(i,A) Sum Distortion =Median Distortion = replace sum with median

29 Copeland Mechanism Majority Graph: Edge (A,B) if A pairwise defeats B Copeland Winner: Candidate who defeats most others B A C E D

30 Copeland Mechanism Majority Graph: Edge (A,B) if A pairwise defeats B Copeland Winner: Candidate who defeats most others B A C E D Tournament winner: has one or two-hop path to all other nodes Always exists, Copeland chooses one such winner

31 Our Results SumMedian Plurality2m-1Unbounded Borda2m-1Unbounded k-approval2n-1Unbounded Veto2n-1Unbounded Copeland55 Uncovered Set55 Lower Bound35 Σ d(i,winner) i i max d ϵ D(p) A min Σ d(i,A) Sum Distortion =Median Distortion = replace sum with median

32 Distortion at most 5 Tournament winner W Optimal candidate X XW Distortion ≤ 3 XW B Distortion ≤ 5

33 Our Results SumMedian Plurality2m-1Unbounded Borda2m-1Unbounded k-approval2n-1Unbounded Veto2n-1Unbounded Copeland55 Uncovered Set55 Lower Bound35 Σ d(i,winner) i i max d ϵ D(p) A min Σ d(i,A) Sum Distortion =Median Distortion = replace sum with median

34 Our Results SumMedian Plurality2m-1Unbounded Borda2m-1Unbounded k-approval2n-1Unbounded Veto2n-1Unbounded Copeland55 Uncovered Set55 Lower Bound35 med d(i,winner) max d ϵ D(p) A min med d(i,A) Median Distortion = Median instead of average voter happiness i i

35 Bounds on Percentile Distortion Percentile distortion: happiness of top α -percentile with outcome α=1: minimize maximum voter unhappiness α=1/2: minimize median voter unhappiness α=0: minimize minimum voter unhappiness

36 Bounds on Percentile Distortion Percentile distortion: happiness of top α -percentile with outcome α=1: minimize maximum voter unhappiness α=1/2: minimize median voter unhappiness α=0: minimize minimum voter unhappiness Lower Bounds on Distortion α 01 Unbounded 5 3 2/3

37 Bounds on Percentile Distortion Percentile distortion: happiness of top α -percentile with outcome α=1: minimize maximum voter unhappiness α=1/2: minimize median voter unhappiness α=0: minimize minimum voter unhappiness Lower Bounds on Distortion α 01 Unbounded 5 3 2/3 Upper Bounds on Distortion α 01 Unbounded (Copeland) 5 (Plurality) 3 (m-1)/m

38 Our Results SumMedian Plurality2m-1Unbounded Borda2m-1Unbounded k-approval2n-1Unbounded Veto2n-1Unbounded Copeland55 Uncovered Set55 Lower Bound35 Σ d(i,winner) i i max d ϵ D(p) A min Σ d(i,A) Sum Distortion =Median Distortion = replace sum with median

39 Conclusions and Future Work Closing gap between 5 and 3 Randomized Mechanisms can do better: Get distortion ≤ 3, but lower bound becomes 2 Multiple winners, k-median, k-center Manipulation by voters or by candidates Special voter distributions (e.g., never have many voters far away from a candidate)

40 Conclusions and Future Work Closing gap between 5 and 3 Randomized Mechanisms can do better: Get distortion ≤ 3, but lower bound becomes 2 Multiple winners, k-median, k-center Manipulation by voters or by candidates Special voter distributions (e.g., never have many voters far away from a candidate) What other problems can be approximated using only ordinal information?


Download ppt "Approximating Optimal Social Choice under Metric Preferences Elliot Anshelevich Onkar Bhardwaj John Postl Rensselaer Polytechnic Institute (RPI), Troy,"

Similar presentations


Ads by Google