US 12,748,796 B2Granted

Uncertainty-aware sequence modeling

AssigneeThemis AI, Inc.

InventorsIaroslav Elistratov, Qi Yang, Ege Demir, Fynn Schmitt-Ulms, Elaheh Ahmadi, Alejandro Perez

Life of this patent

Estimating the term…

Loading patent…

Abstract

Described herein are systems and methods for improving accuracy of model output generation. A method can include obtaining a risk-aware model and a user input, applying the risk-aware model to the user input, receiving, based on the applying, model output and corresponding risk values, comparing the corresponding risk values to a threshold risk value, and regenerating the user input based on the comparing. The method can also include iteratively performing the applying, receiving, comparing, and regenerating using the regenerated user input until one or more processing conditions is met. The user input can be regenerated in response to determining that the corresponding risk values are greater than the threshold risk value. The model output can include one or more sequences in a train-of-thought (TOT) of the risk-aware model.

Claims26

Dependent claims

A method for improving accuracy of model output generation, the method comprising:

receiving a user model and a user input, wherein the user input comprises a prompt;
transforming the user model into a risk-aware variant of the user model based on applying a wrapper-specific algorithm to the user model, wherein transforming the user model comprises modifying one or more components of the user model to integrate an uncertainty estimation into the user model such that the risk-aware variant of the user model is configured to compute risk values during inference;
iteratively processing the prompt by the risk-aware variant of the user model until the risk values computed during inference are less than a threshold risk value, wherein each iteration of the iterative processing comprises:
applying the risk-aware variant of the user model to the prompt;
receiving, based on the applying, model output and corresponding risk values;
determining whether one or more of the corresponding risk values is less than the threshold risk value;
modifying the prompt based on a determination that the one or more of the corresponding risk values is greater than or equal to the threshold risk value;
and
setting the modified prompt as the prompt for subsequent iterations of the iterative processing based on the determination that the one or more of the corresponding risk values is greater than or equal to the threshold risk value;
selecting the modified prompt having the one or more corresponding risk values that is less than the threshold risk value;
and
returning the selected prompt for runtime execution with the risk-aware variant of the user model.

9 more elements

A system for improving accuracy of model output generation, the system comprising:

a computer system comprising one or more processors and memory storing instructions that, when executed by the one or more processors, cause the computer system to perform a process comprising:
receiving a user model and a user input, wherein the user input comprises a prompt;
transforming the user model into a risk-aware variant of the user model based on applying a wrapper-specific algorithm to the user model, wherein transforming the user model comprises modifying one or more components of the user model to integrate an uncertainty estimation into the user model such that the risk-aware variant of the user model is configured to compute risk values during inference;
iteratively processing the prompt by the risk-aware variant of the user model until the risk values computed during inference are less than a threshold risk value, wherein each iteration of the iterative processing comprises:
applying the risk-aware variant of the user model to the prompt;
receiving, based on the applying, model output and corresponding risk values;
determining whether one or more of the corresponding risk values is less than the threshold risk value;
modifying the prompt based on a determination that the one or more of the corresponding risk values is greater than or equal to the threshold risk value;
and
setting the modified prompt as the prompt for subsequent iterations of the iterative processing based on the determination that the one or more of the corresponding risk values is greater than or equal to the threshold risk value;
selecting the modified prompt having the one or more corresponding risk values that is less than the threshold risk value;
and
returning the selected prompt for runtime execution with the risk-aware variant of the user model.

10 more elements

A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations for improving accuracy of model output generation, the operations comprising:

receiving a user model and a user input, wherein the user input comprises a prompt;
transforming the user model into a risk-aware variant of the user model based on applying a wrapper-specific algorithm to the user model, wherein transforming the user model comprises modifying one or more components of the user model to integrate an uncertainty estimation into the user model such that the risk-aware variant of the user model is configured to compute risk values during inference;
iteratively processing the prompt by the risk-aware variant of the user model until the risk values computed during inference are less than a threshold risk value, wherein each iteration of the iterative processing comprises:
applying the risk-aware variant of the user model to the prompt;
receiving, based on the applying, model output and corresponding risk values;
determining whether one or more of the corresponding risk values is less than the threshold risk value;
modifying the prompt based on a determination that the one or more of the corresponding risk values is greater than or equal to the threshold risk value;
and
setting the modified prompt as the prompt for subsequent iterations of the iterative processing based on the determination that the one or more of the corresponding risk values is greater than or equal to the threshold risk value;
selecting the modified prompt having the one or more corresponding risk values that is less than the threshold risk value;
and
returning the selected prompt for runtime execution with the risk-aware variant of the user model.

9 more elements

Dependencies are inferred from the claim text.