Methodology to automatically incorporate feedback to enable self learning in neural learning artifactories
Approaches, techniques, and mechanisms are disclosed for generating, enhancing, applying and updating knowledge neurons for providing decision making information to a wide variety of client applications. Domain keywords for knowledge domains are generated from domain data of selected domain data sources, along with keyword values for the domain keywords, and are used to generate knowledge artifacts for inclusion in knowledge neurons. These knowledge neurons may be enhanced by domain knowledge data sets found in various data sources and used to generate neural responses to neural queries received from the client applications. Neural feedbacks may be used to update and/or generate knowledge neurons. Any ML algorithm can use, or operate in conjunction with, a neural knowledge artifactory comprising the knowledge neurons to enhance or improve baseline accuracy, for example during a cold start period, for augmented decision making and/or for labeling data points or establishing ground truth to perform supervised learning.
1. A computer-implemented method comprising:
receiving a neural feedback from a client computing device over one or more computer networks, wherein the neural feedback is formatted as a computer message sent from the client computing device to a remote server, wherein the computer message includes one or more message data fields containing numeric measurement data of temperatures collected from one or more sensors;
generating one or more feedback keywords from the neural feedback;
searching, based at least in part on a global neural schema and the one or more feedback keywords generated from the neural feedback, for one or more knowledge neurons in a repository of knowledge neurons;
wherein the one or more knowledge neurons include knowledge artifacts learned from documents retrieved from a plurality of web-based data sources through machine learning with a machine learning model implemented by a computing device;
incorporating neural feedback information, derived from the neural feedback, into the one or more knowledge neurons;
wherein incorporating neural feedback information comprises:
using a plurality of neural feedbacks including the received neural feedback to generate a plurality of historic feedback datapoints comprising a plurality of samples and a corresponding plurality of observed values;
generating a prediction mapping that maps the plurality of samples to the corresponding plurality of observed values;
updating at least one neuron in the repository of knowledge neurons with one or more specific prediction values predicted by the prediction mapping;
and
causing the one or more knowledge neurons to be used by a query processor in one or more computer devices to generate responses to query requests from client computing devices.
2. The method of claim 1 , wherein the neural feedback information as derived from the neural feedback, is recorded as one or more updated histories stored in the one or more knowledge neurons.
3. The method of claim 1 , wherein the neural feedback information derived from the neural feedback and other neural feedback information derived from other neural feedbacks is used to predict a preferred value for a property in one or more knowledge artifacts stored in the one or more knowledge neurons.
4. The method of claim 3 , wherein the preferred value for the property is generated using one or more other machine learning methods, and wherein the one or more other machine learning methods comprises one or more of: regression-based machine learning methods, classification-based machine learning methods, decision-tree-based machine learning methods, or random-forest-based machine learning methods.
5. The method of claim 3 , wherein the preferred value for the property is generated by minimizing an objective function measuring a quality of predicted values.
6. The method of claim 1 , wherein the neural feedback information derived from the neural feedback is used to update an existing knowledge neuron in the repository of knowledge neurons.
7. The method of claim 1 , wherein the neural feedback information derived from the neural feedback is used to generate a new knowledge neuron to be stored in the repository of knowledge neurons.
8. A non-transitory computer readable medium that stores computer instructions which, when executed by one or more computing processors, cause the one or more computing processors to perform:
receiving a neural feedback from a client computing device over one or more computer networks, wherein the neural feedback is formatted as a computer message sent from the client computing device to a remote server, wherein the computer message includes one or more message data fields containing numeric measurement data of temperatures collected from one or more sensors;
generating one or more feedback keywords from the neural feedback;
searching, based at least in part on a global neural schema and the one or more feedback keywords generated from the neural feedback, for one or more knowledge neurons in a repository of knowledge neurons;
wherein the one or more knowledge neurons include knowledge artifacts learned from documents retrieved from a plurality of web-based data sources through machine learning with a machine learning model implemented by a computing device;
incorporating neural feedback information, derived from the neural feedback, into the one or more knowledge neurons;
wherein incorporating neural feedback information comprises:
using a plurality of neural feedbacks including the received neural feedback to generate a plurality of historic feedback datapoints comprising a plurality of samples and a corresponding plurality of observed values;
generating a prediction mapping that maps the plurality of samples to the corresponding plurality of observed values;
updating at least one neuron in the repository of knowledge neurons with one or more specific prediction values predicted by the prediction mapping; and
causing the one or more knowledge neurons to be used by a query processor in one or more computer devices to generate responses to query requests from client computing devices.
9. The non-transitory computer readable medium of claim 8 , wherein the neural feedback information as derived from the neural feedback, is recorded as one or more updated histories stored in the one or more knowledge neurons.
10. The non-transitory computer readable medium of claim 8 , wherein the neural feedback information derived from the neural feedback and other neural feedback information derived from other neural feedbacks is used to predict a preferred value for a property in one or more knowledge artifacts stored in the one or more knowledge neurons.
11. The non-transitory computer readable medium of claim 10 , wherein the preferred value for the property is generated using one or more other machine learning methods, and wherein the one or more other machine learning methods comprises one or more of: regression-based machine learning methods, classification-based machine learning methods, decision-tree-based machine learning methods, or random-forest-based machine learning methods.
12. The non-transitory computer readable medium of claim 10 , wherein the preferred value for the property is generated by minimizing an objective function measuring a quality of predicted values.
13. The non-transitory computer readable medium of claim 8 , wherein the neural feedback information derived from the neural feedback is used to update an existing knowledge neuron in the repository of knowledge neurons.
14. The non-transitory computer readable medium of claim 8 , wherein the neural feedback information derived from the neural feedback is used to generate a new knowledge neuron to be stored in the repository of knowledge neurons.
15. An apparatus, comprising:
one or more computing processors;
a non-transitory computer readable medium that stores computer instructions which, when executed by the one or more computing processors, cause the one or more computing processors to perform:
receiving a neural feedback from a client computing device over one or more computer networks, wherein the neural feedback is formatted as a computer message sent from the client computing device to a remote server, wherein the computer message includes one or more message data fields containing numeric measurement data of temperatures collected from one or more sensors;
generating one or more feedback keywords from the neural feedback;
searching, based at least in part on a global neural schema and the one or more feedback keywords generated from the neural feedback, for one or more knowledge neurons in a repository of knowledge neurons;
wherein the one or more knowledge neurons include knowledge artifacts learned from documents retrieved from a plurality of web-based data sources through machine learning with a machine learning model implemented by a computing device;
incorporating neural feedback information, derived from the neural feedback, into the one or more knowledge neurons;
wherein incorporating neural feedback information comprises:
using a plurality of neural feedbacks including the received neural feedback to generate a plurality of historic feedback datapoints comprising a plurality of samples and a corresponding plurality of observed values;
generating a prediction mapping that maps the plurality of samples to the corresponding plurality of observed values;
updating at least one neuron in the repository of knowledge neurons with one or more specific prediction values predicted by the prediction mapping;
and
causing the one or more knowledge neurons to be used by a query processor in one or more computer devices to generate responses to query requests from client computing devices.
16. The apparatus of claim 15 , wherein the neural feedback information as derived from the neural feedback, is recorded as one or more updated histories stored in the one or more knowledge neurons.
17. The apparatus of claim 15 , wherein the neural feedback information derived from the neural feedback and other neural feedback information derived from other neural feedbacks is used to predict a preferred value for a property in one or more knowledge artifacts stored in the one or more knowledge neurons.
18. The apparatus of claim 17 , wherein the preferred value for the property is generated using one or more other machine learning methods, and wherein the one or more other machine learning methods comprises one or more of: regression-based machine learning methods, classification-based machine learning methods, decision-tree-based machine learning methods, or random-forest-based machine learning methods.
19. The apparatus of claim 17 , wherein the preferred value for the property is generated by minimizing an objective function measuring a quality of predicted values.
20. The apparatus of claim 15 , wherein the neural feedback information derived from the neural feedback is used to update an existing knowledge neuron in the repository of knowledge neurons.