Search Images Maps Play YouTube News Gmail Drive More »
Advanced Patent Search | Web History | Sign in

Patents

Crowdsourcing is described, for example, where solutions to tasks such as designing a logo, writing a piece of code or answering a question are solicited by making open calls to large scale communities. In an example, a crowdsourcing node models a plurality of contests as all-pay auctions, each contest having a task and a reward. In examples, the crowdsourcing node is arranged to set rewards for the contests so that the optimal results are received for each contest owner, those owners having provided a budget and using a utility function for each of the contests. In examples, the crowdsourcing node is arranged to recommend contests to potential participants so that those participants can more easily decide which tasks to take on next.

Inventors: Milan Vojnovic, Dominic Daniel DiPalantino
Original Assignee: Microsoft Corporation
Current U.S. Classification: 705/7.13; 705/14.11; 705/300; 705/319; 705/348

View patent at USPTO
Search USPTO Assignment Database
Download USPTO Public PAIR data

Claims

1. A computer-implemented method at a crowdsourcing node in a communications network having a plurality of potential contest participants, the method comprising:

storing a data structure in memory the data structure holding a model of a plurality of contests offered by the crowdsourcing node to the potential participants, each contest having a task and a reward;

monitoring the plurality of potential contest participants and storing an estimate of a total number of those potential contest participants;

using the model and the estimated total number of potential contest participants to control the behavior of the crowdsourcing node.

2. A method as claimed in claim 1 wherein the data structure is stored to hold a model of the contests as all-pay auctions.

3. A method as claimed in claim 1 which further comprises storing at the data structure, for each contest, a contest budget and a utility function.

4. A method as claimed in 3 which further comprises determining a relative reward for each of the contests by using a processor at the node to optimize an objective which is related to the aggregated utility over the contests minus the aggregated cost over the contests.

5. A method as claimed in claim 1 which further comprises assessing whether the estimated total number of participants is above a specified threshold.

6. A method as claimed in claim 5 which further comprises determining a relative reward for each of the contests by using a processor at the node to optimize an objective which is independent of the total number of participants.

7. A method as claimed in claim 1 which further comprises receiving information about a distribution of skills in the population of potential participants.

8. A method as claimed in claim 1 which further comprises storing the data structure such that the model represents each player as having the same skill for each contest.

9. A method as claimed in claim 1 which further comprises storing the data structure such that the model represents each player as having contest-specific skills.

10. A method as claimed in claim 1 wherein the step of controlling the behavior of the crowdsourcing node comprises sending output comprising contest recommendations to potential contest participants.

11. A method as claimed in claim 10 which further comprises receiving skill information for a particular potential participant; identifying one of a plurality of skill levels for the particular potential participant on the basis of the received skill information; accessing a weighted mapping for the identified skill level to the contests and using that mapping to select contests for recommending to the particular potential participant.

12. A method as claimed in claim 10 which further comprises checking that the estimated total number of potential participants is above a threshold; receiving skill information for a particular potential participant; identifying one of a plurality of skill levels for the particular potential participant on the basis of the received skill information; selecting a plurality of contests on the basis of the identified skill level and recommending at least one of those contests to the particular potential participant.

13. A computer-implemented method at a crowdsourcing node in a communications network having a plurality of potential contest participants, the method comprising:

storing a data structure in memory the data structure holding a model of a plurality of contests offered by the crowdsourcing node to the potential participants, each contest having a task and a reward and where the model represents each contest as an all-pay auction;

monitoring the plurality of potential contest participants and storing an estimate of a total number of those potential contest participants;

storing at the data structure, for each contest, a contest budget and a utility function;

setting a relative reward for each of the contests by using a processor at the node to optimize an objective which is related to the aggregated utility over the contests minus the aggregated cost over the contests.

14. A method as claimed in claim 13 wherein the rewards are reputation points and wherein the step of optimizing the objective comprises determining a shadow demand and using that shadow demand to set the relative rewards using a relationship which is dependent on the total number of potential participants.

15. A method as claimed in claim 13 wherein the rewards are reputation points and the method comprises checking whether the estimated total number of potential participants is greater than a threshold and if so, wherein the step of optimizing the objective comprises determining a shadow demand and using that shadow demand to set the relative rewards using a relationship which is independent of the total number of potential participants.

16. A crowdsourcing system comprising: a crowdsourcing node arranged to be connected to a communications network having a plurality of potential contest participants;

a memory at the crowdsourcing node storing a data structure holding a model of a plurality of contests offered by the crowdsourcing node to the potential participants, each contest having a task;

an input arranged to monitor the plurality of potential contest participants and to store at the memory an estimate of a total number of those potential contest participants;

a processor arranged to use the model and the estimated total number of potential contest participants to set rewards for each of the contests;

a communications interface arranged to receive submissions from at least some of the participants in relation to at least some of the offered tasks;
the processor being arranged to select at least one winner for each contest on the basis of the submissions and to allocate the rewards to the appropriate participants.

17. A system as claimed in claim 16 wherein the memory stores the data structure modeling the contests as all-pay auctions.

18. A system as claimed in claim 16 wherein the communications interface is also arranged to provide recommended contests to potential participants.

19. A system as claimed in claim 16 wherein the memory stores, for each contest, a budget and a utility function.

20. A system as claimed in claim 19 wherein the processor is arranged to set the rewards by optimizing an objective which is related to the aggregated utility over the contests minus the aggregated cost over the contests.

Drawings