System and method for multi-task learning
A method of generating a multi-modal prediction is disclosed herein. A computing system retrieves event data from a data store. The event data includes information for a plurality of events across a plurality of seasons. Computing system generates a predictive model using a mixture density network, by generating an input vector from the event data learning, by the mixture density network, a plurality of values associated with a next play following each play in the event data. The mixture density network is trained to output the plurality of values near simultaneously. Computing system receives a set of event data directed to an event in a match. The set of event data includes information directed to at least playing surface position and current score. Computing system generates, via the predictive model, a plurality of values associated with a next event following the event based on the set of event data.
1. A computer-implemented method for generating a multi-modal prediction, the computer-implemented method comprising:
receiving, by one or more processors, match data for a sporting match;
extracting, by the one or more processors, a plurality of parameters associated with an event of a play-by-play event sequence from the match data;
generating, by the one or more processors, an input data set from the plurality of extracted parameters;
generating, by the one or more processors, a multi-modal prediction based on the input data set; and
generating, by the one or more processors, one or more graphical representations of the multi-modal prediction for the event of the play-by-play event sequence.
2. The computer-implemented method of claim 1 , wherein the plurality of parameters include a playing surface position, a subsequent play-by-play event sequence, a plurality of players, at least one team, a team in possession of a ball, or a game context.
3. The computer-implemented method of claim 2 , wherein generating, by the one or more processors, the multi-modal prediction based on the input data set comprises:
providing, by the one or more processors, the plurality of extracted parameters to a mixture density network.
4. The computer-implemented method of claim 1 , wherein generating, by the one or more processors, the multi-modal prediction based on the input data set includes generating predications for at least one of: one or more expected meters, an expected try tackle, an expected try set, a win probability, an expected play selection, or a final score line.
5. The computer-implemented method of claim 1 , the computer-implemented method further comprising:
receiving, by the one or more processors, the match data from one or more tracking systems.
6. The computer-implemented method of claim 1 , wherein generating the input data set includes:
transforming, by the one or more processors, the match data into one or more segmented data sets;
selecting, by the one or more processors, a subset of one or more segmented data sets; and
creating, by the one or more processors, a dense representation of the subset of one or more segmented data sets.
7. The computer-implemented method of claim 6 , wherein generating the input data set further comprises:
providing, by the one or more processors, each of the one or more segmented data sets to one or more embedding layers;
receiving, by the one or more processors, dense output from the one or more embedding layers; and
generating, by the one or more processors, the input data set by concatenating the dense output, one or more continuous features, and spatial information.
8. The computer-implemented method of claim 7 , wherein the one or more continuous features include a score difference, a remaining time, or a playing surface position.
9. A non-transitory computer readable medium comprising one or more sequences of instructions, that, when executed by one or more processors, causes a computing system to perform operations comprising:
receiving, by the computing system, match data for a sporting match;
extracting, by the computing system, a plurality of parameters associated with an event of a play-by-play event sequence from the match data;
generating, by the computing system, an input data set from the plurality of extracted parameters;
generating, by the computing system, a multi-modal prediction based on the input data set; and
generating, by the computing system, one or more graphical representations of the multi-modal prediction for the event of the play-by-play event sequence.
10. The non-transitory computer readable medium of claim 9 , wherein the plurality of parameters include a playing surface position, a subsequent play-by-play event sequence, a plurality of players, at least one team, a team in possession of a ball, or a game context.
11. The non-transitory computer readable medium of claim 10 , wherein generating, by the computing system, the multi-modal prediction based on the input data set comprises:
providing, by the one or more processors, the plurality of extracted parameters to a mixture density network.
12. The non-transitory computer readable medium of claim 9 , wherein generating, by the computing system, the multi-modal prediction based on the input data set includes generating predications for at least one of: one or more expected meters, an expected try tackle, an expected try set, a win probability, an expected play selection, or a final score line.
13. The non-transitory computer readable medium of claim 9 , the operations further comprising:
receiving, by the computing system, the match data from one or more tracking systems.
14. The non-transitory computer readable medium of claim 9 , wherein generating the input data set includes:
transforming, by the computing system, the match data into one or more segmented data sets;
selecting, by the computing system, a subset of one or more segmented data sets; and
creating, by the computing system, a dense representation of the subset of one or more segmented data sets.
15. The non-transitory computer readable medium of claim 14 , wherein generating the input data set further comprises:
providing, by the computing system, each of the one or more segmented data sets to one or more embedding layers;
receiving, by the computing system, dense output from the one or more embedding layers; and
generating, by the computing system, the input data set by concatenating the dense output, one or more continuous features, and spatial information.
16. A computer system comprising:
a processor; and
a memory having programming instructions stored thereon, which, when executed by the processor, causes a computing system to perform operations comprising:
receiving match data for a sporting match;
extracting a plurality of parameters associated with an event of a play-by-play event sequence from the match data;
generating an input data set from the plurality of extracted parameters;
generating a multi-modal prediction based on the input data set; and
generating one or more graphical representations of the multi-modal prediction for the event of the play-by-play event sequence.
17. The computer system of claim 16 , the operations further comprising:
receiving the match data from one or more tracking systems.
18. The computer system of claim 16 , wherein generating the input data set includes:
transforming the match data into one or more segmented data sets;
selecting a subset of one or more segmented data sets; and
creating a dense representation of the subset of one or more segmented data sets.
19. The computer system of claim 18 , wherein generating the input data set further comprises:
providing each of the one or more segmented data sets to one or more embedding layers;
receiving dense output from the one or more embedding layers; and
generating the input data set by concatenating the dense output, one or more continuous features, and spatial information.
20. The computer system of claim 19 , wherein the one or more continuous features include a score difference, a remaining time, or a playing surface position.