SYSTEMS AND METHODS FOR USING LARGE LANGUAGE MODELS FOR REAL-TIME NETWORK DATA STRUCTURE COMPARISON

1 . A system for responding to imprecise player-requests for game scenario comparisons by dynamically generating input contexts, the system comprising:

one or more processors coupled to non-transitory memory, the one or more processors to:

maintain an indexed data structure comprising a plurality of wager scenarios corresponding to a plurality of live events, wherein each scenario of the plurality of wager scenarios associated with a respective odds value of a respective wager scenario;

receive, from a client device during a communication session in which a first wager scenario of the plurality of wager scenarios is identified, a prompt comprising a request for a comparison to the first wager scenario;

generate an input context for a language model using the prompt, data of the first wager scenario, and a set of second wager scenarios selected from the plurality of wager scenarios by searching multiple indexes of the indexed data structure to retrieve the set of second wager scenarios;

generate, using the language model and the input context, an output message indicating the comparison between the first wager scenario and the set of second wager scenarios by providing the input context; and

provide the output message to the client device in response to the prompt.

2 . A method for generating game scenario comparisons, comprising:

maintaining, with a server system, an indexed data structure comprising a plurality of wager scenarios corresponding to a plurality of live events, wherein each scenario of the plurality of wager scenarios associated with a respective odds value of a respective wager scenario;

receiving, at the server system, from a client device during a communication session in which a first wager scenario of the plurality of wager scenarios is identified, a prompt comprising a request for a comparison to the first wager scenario;

generating, with the server system, an input context for a language model using the prompt, data of the first wager scenario, and a set of second wager scenarios selected from the plurality of wager scenarios by searching multiple indexes of the indexed data structure to retrieve the set of second wager scenarios;

generating, with the server system, using the language model and the input context, an output message indicating the comparison between the first wager scenario and the set of second wager scenarios by providing the input context by:

generating program code for the comparison based on the output message of the language model, and displaying a rendering of the program code to represent the comparison; and

providing the output message to the client device in response to the prompt.

3 . The method of claim 2 , further comprising retrieving the set of second wager scenarios based on the prompt by:

retrieving a set of query parameters based on the prompt;

searching, with the set of query parameters, multiple indexes of the indexed data structure to retrieve a first candidate wager scenario based on a first index of the multiple indexes and a second candidate wager scenario based on a second index of the multiple indexes;

determining a set of embedding representations by applying an embedding matrix associated with the language model to each respective scenario of the first wager scenario, the first candidate wager scenario, and the second candidate wager scenario; and

determining a ranking of distances in an embedding space between the first wager scenario, the first candidate wager scenario, and the second candidate wager scenario based on the set of embedding representations; and

assigning at least one of the first candidate wager scenario and the second candidate wager scenario as a scenario of the set of second wager scenarios in lieu of the second candidate wager scenario based on the ranking of distances.

4 . The method of claim 2 , further comprising:

obtaining an indication that a player has associated an alternative wager scenario with the first wager scenario; and

in response to receiving the indication, updating an index of the multiple indexes to associate the alternative wager scenario with the first wager scenario, wherein searching the multiple indexes comprises using the alternative wager scenario as a scenario of the set of second wager scenarios.

5 . The method of claim 2 , wherein each respective wager scenario of the plurality of wager scenarios is associated with a respective odds value of a plurality of odds values, and wherein the indication of the comparison comprises a comparison of odds values between the first wager scenario and at least one scenario of the set of second wager scenarios.

6 . The method of claim 2 , further comprising generating the set of second wager scenarios based on the prompt by:

retrieving a set of query parameters based on the prompt;

searching, with the set of query parameters, multiple indexes of the indexed data structure to retrieve a first candidate wager scenario based on a first index of the multiple indexes and a second candidate wager scenario based on a second index of the multiple indexes;

determining a tolerance boundary based on a first game parameter tolerance of the first candidate wager scenario and a second game parameter tolerance of the second candidate wager scenario;

selecting a tolerance modification value based on the tolerance boundary; and

modifying a wager property of the first wager scenario based on the tolerance modification value to generate properties of the set of second wager scenarios, wherein generating the input context comprises incorporating the properties of the set of second wager scenarios in the input context.

7 . The method of claim 6 , wherein generating the set of second wager scenarios comprises updating the first wager scenario with an additional conditional event by searching at least one of the first index or the second index to retrieve the additional conditional event, wherein generating the input context comprises incorporating the properties of the set of second wager scenarios in the input context.

8 . The method of claim 2 , wherein generating the input context comprises generating the input context to include data relating to a first wager type of the first wager scenario and data relating to a second wager type of the set of second wager scenarios by:

generating a first document comprising the first wager type using a machine learning summarization model;

generating a second document comprising the second wager type using the machine learning summarization model; and

updating the input context to incorporate the first document and the second document.

9 . The method of claim 2 , wherein the client device is associated with a player profile, the first wager scenario identified in the player profile, further comprising:

receiving, from the client device, a second request to update the player profile to replace the first wager scenario with a replacement scenario of the set of second wager scenarios during the communication session; and

updating the player profile to identify the replacement scenario in response to the second request by updating an in-memory data store to store parameters of the replacement scenario.

10 . The method of claim 9 , wherein:

the prompt is a first prompt;

receiving the second request comprises:

receiving a second prompt for the language model;

retrieving, from the language model a command character indicating an update to the player profile;

determining a result indicating that data characterizing the first wager scenario is stored in association with the player profile in the in-memory data store; and

updating data associated with the player profile in the in-memory data store based on the result.

11 . The method of claim 2 , wherein the client device is associated with a player profile indicating a plurality of historical wagers, further comprising:

determining that the request for the comparison relates to at least one historical wager of the plurality of historical wagers; and

generating the input context to further include data of at least one historical wager.

12 . The method of claim 2 , wherein the prompt comprises an event identifier stored in an in-memory storage, and wherein generating the input context comprises retrieving additional values associated with the event identifier in the in-memory storage.

13 . A system, comprising:

one or more processors coupled to non-transitory memory, the one or more processors to:

maintain a data structure comprising a plurality of wager opportunities corresponding to a plurality of live events;

receive, from a client device during a communication session identifying a first wager opportunity of the plurality of wager opportunities, a prompt comprising a request for a comparison to the first wager opportunity;

generate an input context for a language model using the prompt, data of the first wager opportunity, and at least one second wager opportunity selected from the plurality of wager opportunities;

generate, using the language model and the input context, an output message indicating the comparison between the first wager opportunity and the at least one second wager opportunity; and

provide the output message to the client device in response to the prompt.

14 . The system of claim 13 , wherein each respective opportunity of the plurality of wager opportunities is associated with a respective odds value, and wherein the indication of the comparison comprises a comparison of odds values between the first wager opportunity and the at least one second wager opportunity.

15 . The system of claim 13 , wherein the one or more processors are further configured to:

generate the at least one second wager opportunity based on the prompt.

16 . The system of claim 15 , wherein the one or more processors are further configured to:

generate the at least one second wager opportunity by updating the first wager opportunity with an additional conditional event.

17 . The system of claim 13 , wherein the one or more processors are further configured to:

generate the input context to include data relating to a first wager type of the first wager opportunity and data relating to a second wager type of the at least one second wager opportunity.

18 . The system of claim 13 , wherein the client device is associated with a player profile, the first wager opportunity identified in the player profile, wherein the one or more processors are further configured to:

receive, from the client device, a second request to update the player profile to replace the first wager opportunity with the at least one second wager opportunity; and

update the player profile to identify the at least one second wager opportunity in response to the second request.

19 . The system of claim 18 , wherein the one or more processors are further configured to:

receive the second request as indicated in a second prompt for the language model; and

update the player profile responsive to the language model processing the second prompt.

20 . The system of claim 13 , wherein the client device is associated with a player profile indicating a plurality of historical wagers, wherein the one or more processors are further configured to:

determine that the request for the comparison relates to at least one historical wager of the plurality of historical wagers; and

generate the input context to further include data of at least one historical wager.