# US 2019/0377897 A1 · Hybrid cognitive system for ai/ml data privacy

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0377897 A1
- Document ID: US20190377897A1
- Status: Published application
- Kind: A1
- Application number: 16/002,934
- Filed: 2018-06-07
- Published: 2019-12-12
- Earliest claimed: 2018-06-07
- Assignees: Cisco Technology, Inc.
- Applicants: Cisco Technology, Inc.
- Inventors: Keith Griffin; Jonathan Rosenberg
- Web page: https://patentgrep.com/patent/US20190377897A1

## CPC

- G06F 21/6245: {Protecting personal data, e.g. for financial or medical purposes}
- G06Q 10/06: Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06F 9/453: {Help systems}
- G06F 16/9535: Search customisation based on user profiles and personalisation
- G06N 20/00: Machine learning
- H04L 67/1097: for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

Systems, methods, and devices are disclosed for cognitive collaboration systems on a hybrid node. A query is received by a virtual assistant running on a public cloud, and it is determined whether the query pertains to data available on a public cloud resource, or the query pertains to data available on a private cloud resource. When it is determined that the query pertains to the data available on the public cloud resource, the query is interpreted by using a first model trained on at least one machine learning technique on data from the public cloud. When it is determined that the query pertains to the data available on the private cloud resource, the query is interpreted by using a second model trained on at least one machine learning technique on the data from the private cloud.

## Claims

### Claim 1

A non-transitory computer readable medium comprising instructions that, when executed, cause a computing system to:
receive a query by a virtual assistant running on a public cloud;
determine whether the query pertains to data available on a public cloud resource, or the query pertains to data available on a private cloud resource;
interpret the query by using a first model trained on at least one machine learning technique on data from the public cloud when it is determined that the query pertains to the data available on the public cloud resource; and
interpret the query by the using a second model trained on at least one machine learning technique on data from the private cloud when it is determined that the query pertains to the data available on the private cloud resource.

### Claim 2 (depends on 1)

The non-transitory computer readable medium of claim 1, wherein the instructions to determine whether the query pertains to the data available on the public cloud resource, or the query pertains to the data available on the private cloud resource, includes instructions to cause the computing system to:
recognize an intent included in the query; and
identify that the intent is handled by the first model or the second model.

### Claim 3 (depends on 1)

The non-transitory computer readable medium of claim 1, wherein the data available on the private cloud resource is confidential data that does not leave the private cloud resource, and wherein the second model trained on the confidential data is executed on the public cloud.

### Claim 4 (depends on 3)

The non-transitory computer readable medium of claim 3, the instructions cause the computing system to:
request an answer to the query from the data on the private cloud resource;
receive a response to the query from the private cloud resource; and
deliver the response using a virtual assistant running on the private cloud.

### Claim 5 (depends on 1)

The non-transitory computer readable medium of claim 1, wherein the query is an initial query in a dialog between the virtual assistant and a user, and the initial query is interpreted using the first model, and a subsequent query is interpreted by the second model.

### Claim 6 (depends on 1)

The non-transitory computer readable medium of claim 1, wherein the first model and the second model share at least one common functionality, and wherein the first model determines, based on the at least one common functionality, that the query pertains to the second model.

### Claim 7 (depends on 1)

The non-transitory computer readable medium of claim 1, wherein the second model is one of a plurality of second models, each second model being specific to a respective domain from a plurality of domains, the instructions causing the computing system to:
identify that the query originates from a first location;
determine access rights to the plurality of second models based on the first location; and
select at least one of the plurality of second models to interpret the query from the at least one of the plurality of second models to which access rights are permitted.

### Claim 8

A hybrid cognitive system comprising:
at least one private cloud node having confidential data stored thereon, the at least one private cloud configured to train a second model using at least one machine learning technique on the confidential data, the at least one private cloud configured to export the second model to a public cloud;
a public cloud node configured to run a first model trained on at least one machine learning technique on data from the public cloud, and configured to run a second model, the public cloud further including computer readable instructions effective to cause the public cloud to:
receive a query by a virtual assistant running on the public cloud;
determine whether the query pertains to data available on the public cloud, or the query pertains to the confidential data available on the private cloud;
interpret the query by using the first model on the public cloud when it is determined that the query pertains to the data available on the public cloud; and
interpret the query by the using the second model when it is determined that the query pertains to the confidential data available on the private cloud.

### Claim 9 (depends on 8)

The hybrid cognitive system of claim 8, wherein the public node configured to determine whether the query pertains to the data available on the public cloud, or the query pertains to the data available on the private cloud, includes computer readable instructions effective to cause the public cloud to:
recognize an intent included in the query; and
identify that the intent is handled by the first model or the second model.

### Claim 10 (depends on 8)

The hybrid cognitive system of claim 8, wherein the data available on the private cloud is confidential data that does not leave the private cloud, and wherein the second model trained on the confidential data is executed on the public cloud.

### Claim 11 (depends on 10)

The hybrid cognitive system of claim 10, the public cloud node configured to:
request an answer to the query from the data on the private cloud;
receive a response to the query from the private cloud; and
deliver the response using a virtual assistant running on the private cloud.

### Claim 12 (depends on 8)

The hybrid cognitive system of claim 8, wherein the query is an initial query in a dialog between the virtual assistant and a user, and the initial query is interpreted using the first model, and a subsequent query is interpreted by the second model.

### Claim 13 (depends on 8)

The hybrid cognitive system of claim 8, wherein the first model and the second model share at least one common functionality, and wherein the first model determines, based on the at least one common functionality, that the query pertains to the second model.

### Claim 14 (depends on 8)

The hybrid cognitive system of claim 8, wherein the second model is one of a plurality of second models, each second model being specific to a respective domain from a plurality of domains, the computer readable instructions effective to cause the public cloud to:
identify that the query originates from a first location;
determine access rights to the plurality of second models based on the first location; and
select at least one of the plurality of second models to interpret the query from the at least one of the plurality of second models to which access rights are permitted.

### Claim 15

A method comprising:
receiving a query by a virtual assistant running on a public cloud;
determining whether the query pertains to data available on a public cloud resource, or the query pertains to data available on a private cloud resource;
interpreting the query by using a first model trained on at least one machine learning technique on data from the public cloud when it is determined that the query pertains to the data available on the public cloud resource; and
interpreting the query by the using a second model trained on at least one machine learning technique on data from the private cloud when it is determined that the query pertains to the data available on the private cloud resource.

### Claim 16 (depends on 15)

The method of claim 15, wherein determining whether the query pertains to the data available on the public cloud resource, or the query pertains to the data available on the private cloud resource, comprises:
recognizing an intent included in the query; and
identifying that the intent is handled by the first model or the second model.

### Claim 17 (depends on 15)

The method of claim 15, wherein the data available on the private cloud resource is confidential data that does not leave the private cloud resource, and wherein the second model trained on the confidential data is executed on the public cloud.

### Claim 18 (depends on 17)

The method of claim 17, the method further comprising:
requesting an answer to the query from the data on the private cloud resource;
receiving a response to the query from the private cloud resource; and
delivering the response using a virtual assistant running on the private cloud.

### Claim 19 (depends on 15)

The method of claim 15, wherein the query is an initial query in a dialog between the virtual assistant and a user, and the initial query is interpreted using the first model, and a subsequent query is interpreted by the second model.

### Claim 20 (depends on 15)

The method of claim 15, wherein the second model is one of a plurality of second models, each second model being specific to a respective domain from a plurality of domains, the method for comprising:
identifying that the query originates from a first location;
determining access rights to the plurality of second models based on the first location; and
selecting at least one of the plurality of second models to interpret the query from the at least one of the plurality of second models to which access rights are permitted.