# US 2020/0394318 A1 · Privacy trustworthiness based api access

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

- Number: US 2020/0394318 A1
- Document ID: US20200394318A1
- Status: Published application
- Kind: A1
- Application number: 16/440,656
- Filed: 2019-06-13
- Published: 2020-12-17
- Earliest claimed: 2019-06-13
- Assignees: Fujitsu Limited
- Applicants: Fujitsu Limited
- Inventors: Mehdi BAHRAMI; Wei-Peng CHEN
- Web page: https://patentgrep.com/patent/US20200394318A1

## CPC

- G06F 21/62: Protecting access to data via a platform, e.g. using keys or access control rules
- G06N 3/08: Learning methods
- G06F 9/541: {via adapters, e.g. between incompatible applications}
- G06F 16/986: {Document structures and storage, e.g. HTML extensions}
- G06F 21/31: User authentication
- G06F 21/6245: {Protecting personal data, e.g. for financial or medical purposes}
- G06F 21/64: Protecting data integrity, e.g. using checksums, certificates or signatures
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/09: Supervised learning
- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A method may include providing access to a first application programming interface (API) provided by a first party and a second API provided by a second party. The method may also include collecting a first set of API data sources related to the first API and a second set of API data sources related to the second API. The method may additionally include using a deep learning model to predict a privacy trustworthiness level for the first API and the second API, and disabling access to the first API based on the privacy trustworthiness level of the first API being below a threshold level.

## Claims

### Claim 1

A method, comprising:
providing, by a developer, access to a first application programming interface (API) provided by a first party and a second API provided by a second party;
collecting a first plurality of API data sources related to the first API and a second plurality of API data sources related to the second API;
using a deep learning model, predicting a privacy trustworthiness level for the first API and the second API; and
disabling access to the first API based on the privacy trustworthiness level of the first API being below a threshold level.

### Claim 2 (depends on 1)

The method of claim 1, further comprising crawling a plurality of web pages related to the first API to collect the first plurality of API data sources.

### Claim 3 (depends on 2)

The method of claim 2, wherein the plurality of web pages include at least one of a documentation page of documents explaining the first API, a page including terms of service of the first API, a page including terms of service of the first party, a news page mentioning the first API or the first party, and a stock website including a stock price of the first party.

### Claim 4 (depends on 3)

The method of claim 3, wherein each of the plurality of web pages are accessed without providing login information related to the respective plurality of web pages.

### Claim 5 (depends on 1)

The method of claim 1, further comprising:
obtaining seed words related to the privacy trustworthiness level;
based on the seed words, deriving synsets of the seed words;
determining a set of top terms occurring most frequently in the first plurality of API data sources; and
quantifying an amount of overlap between the set of top terms and a combination of the seed words and synsets of the seed words;
wherein the deep learning model uses the amount of overlap in predicting the privacy trustworthiness level for the first API.

### Claim 6 (depends on 1)

The method of claim 1, further comprising:
storing multiple versions of the first plurality of API data sources, each of the multiple versions corresponding to a different point in time;
tracking changes in the first plurality of API data sources based on differences between the multiple versions; and
updating the privacy trustworthiness level of the first API based on the tracked changes to the first plurality of API data sources.

### Claim 7 (depends on 1)

The method of claim 1, further comprising:
receiving an identifier of a third API to be provided by the developer upon approval;
collecting a third plurality of API data sources related to the third API based on the identifier of the third API;
using the deep learning model, predicting the privacy trustworthiness level for the third API; and
enabling access to the third API based on the privacy trustworthiness level of the third API being above the threshold level.

### Claim 8 (depends on 7)

The method of claim 7, further comprising:
tracking historic values of the privacy trustworthiness level of the first API;
wherein the deep learning model uses the historic values of the privacy trustworthiness level of the first API to predict a future privacy trustworthiness level of the first API, the second API, or the third API.

### Claim 9 (depends on 1)

The method of claim 1, further comprising training the deep learning model, comprising:
providing extracted data from the first plurality of API data sources related to the first API for multiple points in time as training data;
obtaining rules related to types of data of the extracted data, the rules indicating how the types of data affect a privacy trustworthiness score;
performing bidirectional long short-term memory (bi-LSTM) deep learning on the extracted data for each of the multiple points in time to derive a first API privacy trustworthiness score for each of the multiple points in time;
combining the first API privacy trustworthiness scores for each of the multiple points in time; and
distributing the privacy trustworthiness scores across a plurality of privacy trustworthiness levels as representative privacy trustworthiness levels.

### Claim 10 (depends on 1)

The method of claim 1, further comprising:
deriving data collected by the first API; and
deriving data shared by the first API and to what entity the data shared is provided;
wherein the deep learning model uses the data collected by the first API, the data shared by the first API, and the entity to which the data shared is provided by the first API in predicting the privacy trustworthiness level of the first API.

### Claim 11

A non-transitory computer-readable medium containing instructions that, when executed by one or more processors, are configured to perform operations, the operations comprising:
providing access to a first application programming interface (API) provided by a first party and a second API provided by a second party;
collecting a first plurality of API data sources related to the first API and a second plurality of API data sources related to the second API;
using a deep learning model, predicting a privacy trustworthiness level for the first API and the second API; and
disabling access to the first API based on the privacy trustworthiness level of the first API being below a threshold level.

### Claim 12 (depends on 11)

The computer-readable medium of claim 11, the operations further comprising:
crawling a plurality of web pages related to the first API to collect the first plurality of API data sources;
wherein the plurality of web pages include at least one of a documentation page of documents explaining the first API, a page including terms of service of the first API, a page including terms of service of the first party, a news page mentioning the first API or the first party, and a stock website including a stock price of the first party.

### Claim 13 (depends on 12)

The computer-readable medium of claim 12, wherein each of the plurality of web pages are accessed without providing login information related to the respective plurality of web pages.

### Claim 14 (depends on 11)

The computer-readable medium of claim 11, the operations further comprising:
obtaining seed words related to the privacy trustworthiness level;
based on the seed words, deriving synsets of the seed words;
determining a set of top terms occurring most frequently in the first plurality of API data sources; and
quantifying an amount of overlap between the set of top terms and a combination of the seed words and synsets of the seed words;
wherein the deep learning model uses the amount of overlap in predicting the privacy trustworthiness level for the first API.

### Claim 15 (depends on 11)

The computer-readable medium of claim 11, the operations further comprising:
storing multiple versions of the first plurality of API data sources, each of the multiple versions corresponding to a different point in time;
tracking changes in the first plurality of API data sources based on differences between the multiple versions; and
updating the privacy trustworthiness level of the first API based on the tracked changes to the first plurality of API data sources.

### Claim 16 (depends on 11)

The computer-readable medium of claim 11, the operations further comprising:
receiving an identifier of a third API to be provided by a developer upon approval;
collecting a third plurality of API data sources related to the third API based on the identifier of the third API;
using the deep learning model, predicting the privacy trustworthiness level for the third API; and
enabling access to the third API based on the privacy trustworthiness level of the third API being above the threshold level.

### Claim 17 (depends on 16)

The computer-readable medium of claim 16, the operations further comprising:
tracking historic values of the privacy trustworthiness level of the first API;
wherein the deep learning model uses the historic values of the privacy trustworthiness level of the first API to predict a future privacy trustworthiness level of the first API, the second API, or the third API.

### Claim 18 (depends on 11)

The computer-readable medium of claim 11, the operations further comprising training the deep learning model, comprising:
providing extracted data from the first plurality of API data sources related to the first API for multiple points in time as training data;
obtaining rules related to types of data of the extracted data, the rules indicating how the types of data affect a privacy trustworthiness score;
performing bidirectional long short-term memory (bi-LSTM) deep learning on the extracted data for each of the multiple points in time to derive a first API privacy trustworthiness score for each of the multiple points in time;
combining the first API privacy trustworthiness scores for each of the multiple points in time; and
distributing the privacy trustworthiness scores across a plurality of privacy trustworthiness levels as representative privacy trustworthiness levels.

### Claim 19

A system comprising:
one or more processors;
one or more non-transitory computer-readable media containing instructions, which when executed by the one or more processors, cause the system to perform operations comprising:
providing access to a first application programming interface (API) provided by a first party and a second API provided by a second party;
collecting a first plurality of API data sources related to the first API and a second plurality of API data sources related to the second API;
using a deep learning model, predicting a privacy trustworthiness level for the first API and the second API; and
disabling access to the first API based on the privacy trustworthiness level of the first API being below a threshold level.

### Claim 20 (depends on 19)

The system of claim 19, the operations further comprising:
crawling a plurality of web pages related to the first API to collect the first plurality of API data sources;
wherein the plurality of web pages include at least one of a documentation page of documents explaining the first API, a page including terms of service of the first API, a page including terms of service of the first party, a news page mentioning the first API or the first party, and a stock website including a stock price of the first party.
