Salesforce Com, Inc.

Ketch Kloud, Inc.

Salesforce, Inc.

Spectrum Labs, Inc.

Switchbit, Inc.

Truthset, Inc.

Yacov Salomon

Vivek Vaidya

Seth Yates

Tom A Chavez

Anton Winter

Maxwell Anderson

Samuel Alexander

Kexin Xie

Wanderley Liu

David Yourdon

Jonathan Thomas Purnell

Shrestha Basu Mallick

Brian Brechbuhl

Nathan Irace Burke

Michael Jones

Sheng Loong Su

Trent Albright

Jonathan Budd

Maria Dimakopoulou

Indrajit Haridas

Alexander Greene

Zhao Jin

Todd Swardenski

Armita Peymandoust

Kathryn Ward Barnitt

Nawid Sayed

Aditya Chaturvedi

Theodore Jacob Kornish

Josh Newman

Scott Matthew McKinley

Methods, systems, and devices for data processing are described. Some systems may support data processing permits and cryptographic techniques tying user consent to data handling. By tying user consent to data handling, the systems may comply with data regulations on a technical level and efficiently update to handle changing data regulations and/or regulations across different jurisdictions. For example, the system may maintain a set of data processing permits indicating user consent for the system to use a user's data for particular data processes. The system may encrypt the user's data using a cryptographic key (e.g., a cryptographic nonce) and may encrypt the nonce using permit keys for any permits applicable to that data. In this way, to access a user's data for a data process, the system may first verify that a relevant permit indicates that the user complies with the requested process prior to decrypting the user's data.

Data processing permits system with keys

Methods, systems, and devices for data processing are described. Some database systems may implement data processing permits to manage data access. A database system may use encryption schemes to tie permits to data (e.g., cryptographically ensuring that the system follows data regulations). To support queries for a database implementing such encryption schemes, the database may implement a proxy. When the system receives a query, the database proxy may intercept and transform the query based on the encryption schema of the database. The database proxy may execute the transformed query at the database, receive encrypted query results in response, and decrypt the results for use by the querying application. Additionally, the system may access relevant data processing permits to support querying operations. For example, the system may use permits when transforming the query, executing the query in the database, preparing query results for the querying application, or any combination thereof.

Managing queries with data processing permits

This disclosure describes one or more implementations of a data verification system that provides accurate validations of user trait data for data providers. For example, in various implementations, the data verification system generates and utilizes data verification models and approaches to determine the probability that user trait data obtained by data providers is accurate and correct. In this manner, the data verification system can independently evaluate the accuracy of both individual user records as well as collective segments of user records for data providers.

Generating user attribute verification scores to facilitate improved data validation from scaled data providers

Methods, systems, and devices for displaying a user interface for frequent pattern (FP) analysis are described. In some cases, data stored at a multi-tenant database server may be analyzed to understand various interactions and patterns between data attributes associated with multiple users, or determine one or more attributes associated with a characterization of an individual (e.g., a persona). The multi-tenant database server may effectively cluster and/or perform calculations on attributes of the data to understand user patterns and determine common personas. The results may then be displayed by a user interface at a user device (e.g., associated with the user).

User interface for frequent pattern analysis

Methods, systems, and devices supporting epsilon (&#x3b5;)-closure for frequent pattern (FP) analysis are described. Some database systems may analyze data sets to determine FPs. In some cases, the FP set may include a large number of semi-redundant patterns, resulting in significant memory or processing overhead. To reduce the redundancy of these patterns, the database system may implement pre-configured or dynamic threshold occurrence differences (e.g., &#x3b5; values) to test against related patterns. For example, the database system may calculate the difference between the data objects covered by a sub-pattern and a super-pattern (e.g., where the super-pattern includes all the same data attributes of the sub-pattern, plus one additional attribute). This difference may be compared to a corresponding &#x3b5; value, and if the difference is less than the &#x3b5; value, the database system may remove one of the patterns (e.g., the sub-pattern) from the set of valid FPs to limit redundancy.

Epsilon-closure for frequent pattern analysis

Methods, systems, and devices for data processing are described. Some database systems may support differential privacy for encrypted data. For example, a database may store user data as ciphertext. A system may receive a statistical query for the user data and may identify a relevant differential privacy mechanism. The system may transform the query to operate on encrypted data while including a noisification function based on the mechanism. The system may execute the transformed query at the database, involving adding noise to the query result according to the noisification function without decrypting the data. For example, the system may leverage homomorphic encryption techniques to inject the noise while the data remains encrypted. The database may return the noisified, encrypted query results, which the system may decrypt for statistical analysis. By applying differential privacy on the encrypted data, the system may avoid exposing any private user information throughout the process.

Differential privacy for encrypted data

Methods, systems, and devices for data attribution using frequent pattern analysis are described. In some cases, data stored at a multi-tenant database server may be analyzed to understand various interactions and patterns between data attributes associated with multiple users. The multi-tenant database server may effectively cluster and/or perform calculations on attributes of the data to understand user patterns. In some examples, the multi-tenant database server may determine a change (e.g., a probability change) in the user patterns by removing one or more attributes from the data set and re-performing the analysis. By re-performing the analysis, the multi-tenant database server may attribute a value to individual pieces and combinations of the data in order to indicate the effect that each piece of data has on the analysis.

Data attribution using frequent pattern analysis

Methods, systems, and devices supporting differential support for frequent pattern (FP) analysis are described. Some database systems may analyze data sets to determine FPs of data attributes within the data sets. However, if data distributions for different types of data attributes vary greatly, more frequent data attribute types may skew the FPs away from the less frequent types. To reduce the noise of common attributes while maintaining sensitivity to the less common attributes, the database system may implement multiple minimum support (e.g., frequency) thresholds. For example, the database system may adaptively categorize the different data attribute types into data categories based on their distributions and may dynamically determine support thresholds for the categories. Using different minimum support thresholds for different data categories allows the system to filter out data attribute patterns based on the distributions of the data attribute types in the pattern.

Differential support for frequent pattern analysis

Methods, systems, and devices for data processing are described. In some systems, data pipelines may be implemented to handle data processing jobs. To improve data pipeline flexibility, the systems may use separate pipeline and policy declarations. For example, a pipeline server may receive both a pipeline definition defining a first set of data operations to perform and a policy definition including instructions for performing a second set of data operations, where the first set of data operations is a subset of the second set. The server may execute a data pipeline based on a trigger (e.g., a scheduled trigger, a received message, etc.). To execute the pipeline, the server may layer the policy definition into the pipeline definition when creating an execution plan. The server may execute the execution plan by performing a number of jobs using a set of resources and plugins according to the policy definition.

Policy handling for data pipelines

Methods, systems, and devices for language mapping are described. Some machine learning models may be trained to support multiple languages. However, word embedding alignments may be too general to accurately capture the meaning of certain words when mapping different languages into a single reference vector space. To improve the accuracy of vector mapping, a system may implement a supervised learning layer to refine the cross-lingual alignment of particular vectors corresponding to a vocabulary of interest (e.g., toxic language). This supervised learning layer may be trained using a dictionary of toxic words or phrases across the different supported languages in order to learn how to weight an initial vector alignment to more accurately map the meanings behind insults, threats, or other toxic words or phrases between languages. The vector output from this weighted mapping can be sent to supervised models, trained on the reference vector space, to determine toxicity scores.

Toxic vector mapping across languages

A database system performs lookalike analysis on a data set including a plurality of user identifiers, which are associated with one or more attribute records. The database system classifies the user identifiers into one or more segments of user identifiers based on the attribute records. The database system performs Linear Discriminant Analysis (LDA) to calculate a measure of importance of the attribute records relative to the one or more segments. The database system auto-correlates the attribute records based on the numbers of attribute records in the user identifier population and the one or more segments. The database system identifies a set of user identifiers relative to one or more segments using the measures of importance and the auto-correlated parameters.

Method and system for classifying user identifiers into similar segments

Methods, systems, and devices for determining device associations are described. Some database systems may store information related to device characteristics. Each of these devices may be operated by one or more users, and each user may operate one or more devices. In some cases, information about users may be more valuable than information about devices. As such, a system may determine probable associations between devices, where an association can correspond to operation by a same user. To determine device associations, the system may perform a machine-learning process (e.g., using probabilistic soft logic (PSL) and a hinge-loss Markov Random Field (HL-MRF) model) on input device characteristics and connection information to generate a probability density function. The probability density function may indicate associations between devices within the system. Based on one or more thresholds, the system may determine sets of associated devices and may transmit this association information for analysis or display.

Probabilistic framework for determining device associations

Growing community of inventors

Danville, CA, United States of America

Yacov Salomon