Abstract: A unified network service that connects multiple disparate private networks and end user client devices operating on separate networks is described. The multiple disparate private networks and end user client devices connect to a distributed cloud computing network that provides routing services, security services, and performance services, and that can be controlled consistently regardless of the connection type. The unified network service provides uniform access control at the L3 layer (e.g., at the IP layer) or at a higher layer using user identity information (e.g., a zero-trust model). The disparate private networks are run on top of the distributed cloud computing network. The virtual routing layer of the distributed cloud computing network allows customers of the service to have private resources visible only to client devices (e.g.

Abstract: In an embodiment, a process for automatic model monitoring for data streams includes receiving an input dataset, using a machine learning model to determine a model score for each data record of at least a portion of the input dataset, and determining monitoring values. Each monitoring value is associated with a measure of similarity between model scores for those data records of the input dataset within a corresponding moving reference window and model scores for those data records of the input dataset within a corresponding moving target window. The process includes outputting the determined monitoring values.

Abstract: A unified network service that connects multiple disparate private networks and end user client devices operating on separate networks is described. The multiple disparate private networks and end user client devices connect to a distributed cloud computing network that provides routing services, security services, and performance services, and that can be controlled consistently regardless of the connection type. The unified network service provides uniform access control at the L3 layer (e.g., at the IP layer) or at a higher layer using user identity information (e.g., a zero-trust model). The disparate private networks are run on top of the distributed cloud computing network. The virtual routing layer of the distributed cloud computing network allows customers of the service to have private resources visible only to client devices (e.g.

Abstract: An IPsec tunnel request for establishing an IPsec tunnel from a customer router to an anycast IP address of a distributed cloud computing network is received. The same anycast IP address is shared among compute servers of the distributed cloud computing network. A handshake is performed with the customer router from a first compute server including generating security associations for encrypting and decrypting IPsec traffic. The security associations are propagated to each compute server and are used for encrypting and decrypting traffic.

Abstract: In an embodiment, a process for automatic model monitoring for data streams includes receiving an input dataset, using a machine learning model to determine a model score for each data record of at least a portion of the input dataset, and determining monitoring values. Each monitoring value is associated with a measure of similarity between model scores for those data records of the input dataset within a corresponding moving reference window and model scores for those data records of the input dataset within a corresponding moving target window. The process includes outputting the determined monitoring values.

Abstract: In an embodiment, a process for adaptive threshold estimation for streaming data includes determining initial positions for a set of percentile bins, receiving a new data item in a stream of data, and identifying one of the set of percentile bins corresponding to the new data item. The process includes incrementing a count of items in the identified percentile bin, adjusting one or more counts of data items in one or more of the percentile bins including by applying a suppression factor based on a relative ordering of items, and redistributing positions for the set of percentile bins to equalize respective count numbers of items for each percentile bin of the set of percentile bins. The process includes utilizing the redistributed positions of the set of percentile bins to determine a percentile distribution of the data stream, and calculating a threshold based at least in part on the percentiles distribution.

Abstract: A unified network service that connects multiple disparate private networks and end user client devices operating on separate networks is described. The multiple disparate private networks and end user client devices connect to a distributed cloud computing network that provides routing services, security services, and performance services, and that can be controlled consistently regardless of the connection type. The unified network service provides uniform access control at the L3 layer (e.g., at the IP layer) or at a higher layer using user identity information (e.g., a zero-trust model). The disparate private networks are run on top of the distributed cloud computing network. The virtual routing layer of the distributed cloud computing network allows customers of the service to have private resources visible only to client devices (e.g.

Abstract: In an embodiment, a process for automatic model monitoring for data streams includes receiving an input dataset, using a machine learning model to determine a model score for each data record of at least a portion of the input dataset, and determining monitoring values. Each monitoring value is associated with a measure of similarity between model scores for those data records of the input dataset within a corresponding moving reference window and model scores for those data records of the input dataset within a corresponding moving target window. The process includes outputting the determined monitoring values.

Abstract: In an embodiment, a process for explanation reporting based on differentiation between items in different data groups includes obtaining model scores from a first machine learning model and training a second machine learning model to learn how to differentiate between two groups based on at least one of: features and the model scores obtained from the first machine learning model. The process includes applying the second machine learning model to each data record in a first group of data records to determine a corresponding ranking score for each data record in the first group, and based on the corresponding ranking scores, determining a relative contribution of each of the data records in the first group to the differentiation between the first group of data records and a second group of data records.

Abstract: In an embodiment, a process for automatic model monitoring for data streams includes receiving an input dataset, using a machine learning model to determine a model score for each data record of at least a portion of the input dataset, and determining monitoring values. Each monitoring value is associated with a measure of similarity between model scores for those data records of the input dataset within a corresponding moving reference window and model scores for those data records of the input dataset within a corresponding moving target window. The process includes outputting the determined monitoring values.

Abstract: In an embodiment, a process for adaptive threshold estimation for streaming data includes determining initial positions for a set of percentile bins, receiving a new data item in a stream of data, and identifying one of the set of percentile bins corresponding to the new data item. The process includes incrementing a count of items in the identified percentile bin, adjusting one or more counts of data items in one or more of the percentile bins including by applying a suppression factor based on a relative ordering of items, and redistributing positions for the set of percentile bins to equalize respective count numbers of items for each percentile bin of the set of percentile bins. The process includes utilizing the redistributed positions of the set of percentile bins to determine a percentile distribution of the data stream, and calculating a threshold based at least in part on the percentiles distribution.

Abstract: In an embodiment, a process for semantic-aware feature engineering includes receiving semantic labels for data fields of training data. Each of the semantic labels is associated with a semantic meaning associated with a corresponding data field. The process includes automatically generating at least one new feature using at least a portion of the semantic labels.