Microsoft Technology Licensing, LLC

Dawn Antonette Burns

Hani Hana Neuvirth

Gueorgui Chkodrov

Yotam Livny

Elad Yom-Tov

Andrey Karpovsky

Dotan Patrich

Julian Federico Gonzalez

Lucius B Fleuchaus

Zhipeng Zhao

Marek Jedrzejewicz

Russell Biles

Vidhi Agarwal

Andre Oneal Alfred

Scott Elliott Gorlick

Justin Minaker

Owen Joseph Davis

According to examples, an apparatus may include a processor and a memory on which is stored machine-readable instructions that when executed by the processor, may cause the processor to identify a first set of entities corresponding to a security incident, identify anomalies associated with the first set of entities that occurred around a predefined time period with respect to the incident, identify a second set of entities associated with the identified anomalies, identify a set of incidents that share a common entity from the second set of entities, determine a probability of likelihood that the set of incidents normally share the common entity, determine whether the determined probability of likelihood falls below a predefined threshold, and based on the determined probability of likelihood falling below the predefined threshold, output an indication that the security incident and the set of incidents are likely related.

Determination of likely related security incidents

An indication of a security alert and a context for the security alert is received. The context includes one or more entities related to the context and a timestamp for the security alert. Data sources for the one or more entities are searched during a time window around the timestamp. One or more anomaly detection models are executed to identify anomalies that are related to the security alert based on the context. Identified anomalies for investigation of the security alert are output.

Detection of anomalous count of new entities

The automated creation of a dataflow graph of a standing query. Once the standing query dataflow graph is created, events may be flowed into the dataflow graph to execute the standing query. In execution, a store query is accessed. The store query is structured in accordance with a store query language. A syntax graph (such as an abstract syntax tree) of the store query may then be generated. Then, using the syntax graph and a set of rules of the store query language, the dataflow graph is automatically generated. This significant speeds up and makes more easy and efficient the conversion of a store query into a standing query.

Standing query creation using store query

Performing anomaly lookup on data sources that include an entity related to an alert. One or more entities related to an alert and a date when the alert occurred are received. The alert may indicate that an anomaly in data collected from a various data sources may be present in at least one of the data sources. The various data sources are searched for the one or more entities around the alert date to determine which of the data sources include the one or more entities. For those data sources including the one or more entities, an anomaly lookup procedure is performed on the data sources during a first time window to determine an initial set of suspicious anomalies.

Anomaly lookup for cyber security hunting

Using a set of anomalies indicative of a malicious pattern of behavior collected from data to determine new alerts for anomalies included in subsequently collected data. A set of anomalies found in data collected from data sources is accessed. The set of anomalies is determined by a prior analysis to be indicative of a malicious pattern of behavior by entities associated with the set of anomalies. Data that is subsequently collected from the data sources is searched to determine if any of the data includes the set of anomalies. Alerts are generated for any of the subsequently collected data that includes the set of anomalies.

Automatic generation of detection alerts

Methods, systems, apparatuses, and computer program products are provided for evaluating security detections. A detection instance obtainer obtains detection instances from a pool, such as a security detections pool. The detection instances may be obtained for detections that meet a predetermined criterion, such as detections that have not been onboarded or rejected, or detections that have generated detection instances for a threshold time period. The detection may be onboarded or rejected automatically based on a volume thresholder and/or a detection performance evaluator. For instance, the volume thresholder may be configured to automatically onboard the detection if the volume of the detection instances is below a first threshold, and reject the detection if the volume is above a second threshold. The detection performance evaluator may be configured to onboard or reject the detection based on an efficacy of the detection (e.g., based on a true positive rate of the detection instances).

Automated onboarding of detections for security operations center monitoring

Growing community of inventors

Lynnwood, WA, United States of America

Dawn Antonette Burns