Explainable and automated decisions in computer-based reasoning systems
The techniques herein include using an input context to determine a suggested action. One or more explanations may also be determined and returned along with the suggested action. The one or more explanations may include (i) one or more most similar cases to the suggested case (e.g., the case associated with the suggested action) and, optionally, a conviction score for each nearby cases; (ii) action probabilities, (iii) excluding cases and distances, (iv) archetype and/or counterfactual cases for the suggested action; (v) feature residuals; (vi) regional model complexity; (vii) fractional dimensionality; (viii) prediction conviction; (ix) feature prediction contribution; and/or other measures such as the ones discussed herein, including certainty. In some embodiments, the explanation data may be used to determine whether to perform a suggested action.
1. A method comprising:
receiving a request for a suggested action based on an input context, in a case-based reasoning system, wherein the case-based reasoning system includes a case-based reasoning model;
determining two or more candidate cases based on the input context in the case-based reasoning system, wherein each of the two or more candidate cases has at least one respective candidate action and a candidate context;
determining an action probability for each of the respective candidate actions associated with the two or more candidate cases based at least in part on a function of a distance of each particular candidate context to the input context and an aggregation of all distances of all candidate contexts to the input context;
determining the suggested action and an action probability for the suggested action based on the respective two or more candidate actions and their action probabilities;
responding to the request for suggested action with the suggested action and at least the action probability for the suggested action, wherein the action probability for the suggested action comprises a categorical action probability associated with the suggested action;
when the action probability for the suggested action is beyond a certain threshold, causing control of a controllable system based on the suggested action;
when the action probability for the suggested action is not beyond the certain threshold:
determining one or more explanation factors for the suggested action determined based at least in part on the input context, wherein the one or more explanation factors comprise the categorical action probability for the suggested action;
providing the one or more explanation factors in response to the request for the suggested action;
wherein the method is performed by one or more computing devices.
2. The method of claim 1 , wherein the input context is a current context for an image labeling system during operation of the image labeling system, and causing control of the controllable system based on the suggested action comprises causing a suggested labeling to be performed by the image labeling system.
3. The method of claim 1 , wherein the input context is a current context for a self-driving car during operation of the self-driving car, and causing control of the controllable system based on the suggested action comprises causing the suggested action to be performed by the self-driving car.
4. The method of claim 1 , wherein the input context is a current context for a manufacturing control system during operation of the manufacturing control system, and causing control of the controllable system based on the suggested action comprises causing the suggested action to be performed by the manufacturing control system.
5. The method of claim 1 , wherein determining the action probability for the suggested action comprises determining the action probability for the suggested action based at least in part on a number of times that each action appears in the respective two or more candidate actions.
6. The method of claim 1 , wherein determining the action probability for the suggested action comprises determining the action probability for the suggested action based at least in part on a function of a distance of the suggested action to the input context.
7. The method of claim 1 , wherein determining the action probability for the suggested action comprises determining the action probability for the suggested action based at least in part on a weighted function of a distance of the suggested action to the input context.
8. The method of claim 1 , wherein determining the action probability for an action comprises determining the action probability for the action based at least in part on a confidence interval of the action for a specific tolerance.
9. A non-transitory computer readable medium storing instructions which, when executed by one or more computing devices, cause the one or more computing devices to perform a process of:
receiving a request for a suggested action based on an input context, in a case-based reasoning system, wherein the case-based reasoning system includes a case-based reasoning model;
determining two or more candidate cases based on the input context in the case-based reasoning system, wherein each of the two or more candidate cases has at least one respective candidate action and a candidate context;
determining an action probability for each of the respective candidate actions associated with the two or more candidate cases based at least in part on a function of a distance of each particular candidate context to the input context and an aggregation of all distances of all candidate contexts to the input context;
determining the suggested action and an action probability for the suggested action based on the respective two or more candidate actions and their action probabilities;
responding to the request for suggested action with the suggested action and at least the action probability for the suggested action, wherein the action probability for the suggested action comprises a categorical action probability associated with the suggested action;
when the action probability for the suggested action is beyond a certain threshold, causing control of a controllable system based on the suggested action;
when the action probability for the suggested action is not beyond the certain threshold:
determining one or more explanation factors for the suggested action determined based at least in part on the input context, wherein the one or more explanation factors comprise the categorical action probability for the suggested action;
providing the one or more explanation factors in response to the request for the suggested action.
10. The non-transitory computer readable medium of claim 9 , wherein the input context is a current context for an image labeling system during operation of the image labeling system, and causing control of the controllable system based on the suggested action comprises causing a suggested labeling to be performed by the image labeling system.
11. The non-transitory computer readable medium of claim 9 , wherein the input context is a current context for a self-driving car during operation of the self-driving car, and causing control of the controllable system based on the suggested action comprises causing the suggested action to be performed by the self-driving car.
12. The non-transitory computer readable medium of claim 9 , wherein the input context is a current context for a manufacturing control system during operation of the manufacturing control system, and causing control of the controllable system based on the suggested action comprises causing the suggested action to be performed by the manufacturing control system.
13. The non-transitory computer readable medium of claim 9 , wherein determining the action probability for the suggested action comprises determining the action probability for the suggested action based at least in part on a number of times that each action appears in the respective two or more candidate actions.
14. The non-transitory computer readable medium of claim 9 , wherein determining the action probability for the suggested action comprises determining the action probability for the suggested action based at least in part on a function of a distance of the suggested action to the input context.
15. The non-transitory computer readable medium of claim 9 , wherein determining the action probability for the suggested action comprises determining the action probability for the suggested action based at least in part on a weighted function of a distance of the suggested action to the input context.
16. The non-transitory computer readable medium of claim 9 , wherein determining the action probability for an action comprises determining the action probability for the action based at least in part on a confidence interval of the action for a specific tolerance.
17. A system for performing a machine-executed operation involving instructions, wherein said instructions are instructions which, when executed by one or more computing devices, cause performance of a process comprising:
receiving a request for a suggested action based on an input context, in a case-based reasoning system, wherein the case-based reasoning system includes a case-based reasoning model;
determining one or more candidate cases based on the input context and the case-based reasoning system, wherein the one or more candidate cases include respective one or more candidate actions;
determining the suggested action based on the respective one or more candidate actions;
determining a regional model of two or more cases in the case-based reasoning model near a suggested case containing the suggested action;
determining for each feature in the input context, whether a value for that feature is outside a range of values for the feature in the cases in the regional model;
responding to the request for suggested action with the suggested action and an indication of any input features outside the range of values for the feature in the cases in the regional model;
determining whether to perform the suggested action based on the indication of input features outside the range of values in the regional model;
when it is determined to perform the suggested action, causing performance of the suggested action;
when it is determined not to perform the suggested action:
determining one or more explanation factors for the suggested action determined based at least in part on the input context, wherein the one or more explanation factors comprise the indication of any input features outside the range of values for the feature in the cases in the regional model;
providing the one or more explanation factors in response to the request for the suggested action.
18. The system of claim 17 , wherein the input context is a current context for a self-driving car during operation of the self-driving car, and causing performance of the suggested action comprises causing the suggested action to be performed by the self-driving car.
19. The system of claim 17 , wherein the input context is a current context for a manufacturing control system during operation of the manufacturing control system, and causing the suggested action comprises causing the suggested action to be performed by the manufacturing control system.