Sift Science, Inc.

Sift Sciences, Inc.

Wei Liu

Hui Wang

Nicholas Benavides

Yanqing Bao

Kostyantyn Gurnov

Yuan Zhuang

Chang Liu

Helen Marushchenko

Ralf Gunter Correa Carvalho

Kevin A Lee

Fred Sadaghiani

Rishabh Kothari

Pradhan Bagur Umesh

Vera Dadok

Irving Chen

Cat Perry

Jintae Kim

Michael Legore

Rachel Mitrano

Duy Tran

Arjun Krishnaiah

Logan Davis

Volha Leusha

James Volz

Louie Zhang

Andy Cai

Yong Fu

Liz Kao

Amey Farde

Mohammed Jouahri

Gary Lee

A system and method for quantile-based assessment and handling of digital events in a digital threat mitigation platform includes receiving, via an application programming interface (API), a request from a subscriber to assess a threat of a digital event, computing, using one or more threat scoring machine learning models, a digital threat inference based on one or more corpora of feature vectors associated with the digital event, wherein the digital threat inference includes an uncalibrated digital threat score, retrieving, from a database, a T-Digest data structure of historical digital threat scores of the subscriber, computing, using the T-Digest data structure of historical digital threat scores, a percentile-based threat score based on the uncalibrated digital threat score computed for the digital event, and executing an automated disposal decision computed for the digital event based on at least the percentile-based threat score satisfying automated decisioning instructions of the digital threat mitigation platform.

Systems and methods for digital threat assessment and mitigation using t-digest score distribution representations and percentile-based threat scoring in a digital threat mitigation platform

A method for machine learning-based detection of an automated fraud or abuse attack includes: identifying, via a computer network, a digital event associated with a suspected automated fraud or abuse attack; composing, via one or more computers, a digital activity signature of the suspected automated fraud or abuse attack based on digital activity associated with the suspected automated fraud or abuse attack; computing, via a machine learning model, an encoded representation of the digital activity signature; searching, via the one or more computers, an automated fraud or abuse signature registry based on the encoded representation of the digital activity signature; determining a likely origin of the digital event based on the searching of the automated fraud or abuse signature registry; and selectively implementing one or more automated threat mitigation actions based on the likely origin of the digital event.

Systems and methods for machine learning-based detection of an automated fraud attack or an automated abuse attack

A system and method for accelerated anomaly detection and replacement of an anomaly-experiencing machine learning-based ensemble includes identifying a machine learning-based digital threat scoring ensemble having an anomalous drift behavior in digital threat score inferences computed by the machine learning-based digital threat scoring ensemble for a target period; executing a tiered anomaly evaluation for the machine learning-based digital threat scoring ensemble that includes identifying at least one errant machine learning-based model of the machine learning-based digital threat scoring ensemble contributing to the anomalous drift behavior, and identifying at least one errant feature variable of the at least one machine learning-based model contributing to the anomalous drift behavior; generating a successor machine learning-based digital threat scoring ensemble to the machine learning-based digital threat scoring ensemble based on the tiered anomaly evaluation; and replacing the machine learning-based digital threat scoring ensemble with the successor machine learning-based digital threat scoring ensemble.

Systems and methods for accelerated detection and replacement of anomalous machine learning-based ensembles and intelligent generation of anomalous artifacts for anomalous ensembles

Systems and methods for accelerated detection and replacement of anomalous machine learning-based digital threat scoring ensembles and intelligent generation of anomalous artifacts for anomalous ensembles

A system and method for accelerating an automated labeling of a volume of unlabeled digital event data samples includes identifying a corpus characteristic of a digital event data corpus that includes a plurality of distinct unlabeled digital event data samples; selecting an automated bulk labeling algorithm based on the corpus characteristic associated with the digital event data corpus satisfying a bulk labeling criterion of the automated bulk labeling algorithm; evaluating a subset of the plurality of unlabeled digital event data samples, wherein evaluating the subset includes attributing a distinct classification label to each digital event data sample within the subset; and in response to the selection, executing the selected automated bulk labeling algorithm against the digital event data corpus, wherein the executing includes simultaneously assigning a classification label equivalent to the distinct classification label to a superset of the digital event data corpus that relates to the subset.

Systems and methods for automated labeling of subscriber digital event data in a machine learning-based digital threat mitigation platform

Systems and methods for detecting digital abuse or digital fraud that involves malicious account testing includes implementing a machine learning threat model that predicts malicious account testing using misappropriate accounts, wherein a subset of a plurality of learnable variables of an algorithmic structure of the machine learning threat model includes one or more learnable variables derived based on feature data indicative of malicious account testing; wherein implementing the machine learning threat model includes: (i) identifying event data from an online event that is suspected to involve digital fraud or digital abuse, (ii) extracting adverse feature data from the event data that map to the one or more learnable variables of the subset, and (iii) providing the adverse feature data as model input to the machine learning threat model; and computing, using the machine learning threat model, a threat prediction indicating a probability that the online event involves malicious account testing.

Systems and methods for configuring and implementing a malicious account testing machine learning model in a machine learning-based digital threat mitigation platform

A system and method for fast-detection and mitigation of emerging network fraud attacks includes sourcing digital event data samples associated with one or more online services; executing graph-rendering computer instructions that automatically construct a backbone graph using a subset of features extracted from the sourced digital event data samples, wherein the constructing includes: identifying, as graphical nodes, a first plurality of distinct features of the subset of features; identifying, as graphical edges, a second plurality of distinct features of the subset of features; generating a graphical edge between distinct pairs of graphical nodes comprising a same type of feature of the subset of features based on feature values associated with at least one distinct feature of the second plurality of distinct features; and mitigating, via a digital threat mitigation action, if one or more emerging network fraud attacks is identified based on an assessment of a cluster of networked nodes.

Systems and methods for intelligently constructing a backbone network graph and identifying and mitigating digital threats based thereon in a machine learning task-oriented digital threat mitigation platform

A machine learning-based system and method for content clustering and content threat assessment includes generating embedding values for each piece of content of corpora of content data; implementing unsupervised machine learning models that: receive model input comprising the embeddings values of each piece of content of the corpora of content data; and predict distinct clusters of content data based on the embeddings values of the corpora of content data; assessing the distinct clusters of content data; associating metadata with each piece of content defining a member in each of the distinct clusters of content data based on the assessment, wherein the associating the metadata includes attributing to each piece of content within the clusters of content data a classification label of one of digital abuse/digital fraud and not digital abuse/digital fraud; and identifying members or content clusters having digital fraud/digital abuse based on querying the distinct clusters of content data.

Systems and methods for machine learning-based digital content clustering, digital content threat detection, and digital content threat remediation in machine learning-based digital threat mitigation platform

A system and method for adaptively sampling a corpus of data samples for improving an accuracy of a predictive machine learning model includes: identifying the corpus of data samples, wherein each data sample of the corpus of data samples is associated with a machine learning-derived threat inference value; stratifying the corpus of data samples into a plurality of distinct strata based on the machine learning-derived threat inference value associated with each data sample of the corpus of data samples; adaptively sampling the plurality of distinct strata; constructing a machine learning training corpus comprising a plurality of data samples based on the adaptive sampling of the plurality of distinct strata; and training the predictive machine learning model based on the machine learning training corpus.

Systems and methods for an adaptive sampling of unlabeled data samples for constructing an informative training data corpus that improves a training and predictive accuracy of a machine learning model

Systems and methods for configuring and implementing a card testing machine learning model in a machine learning-based digital threat mitigation platform

Systems and methods for machine learning-based digital content clustering, digital content threat detection, and digital content threat remediation in machine learning task-oriented digital threat mitigation platform

A machine learning-based system and method for identifying digital threats includes a threat service that: implements a unified threat model that produces a unified threat score that predicts both of: a level of threat of a piece of online content, and a level of threat that a target user will create a harmful piece of online content; wherein: implementing the unified threat model includes: receiving event data comprising historical content data for the target user and content data of the pending piece of online content and historical user digital activity data and real-time user activity data; and providing input of content feature data and user digital activity feature data to the unified threat model; and the unified threat model produces the unified threat score based on the content and the user digital activity data; and computes a threat mitigation action based on an evaluation of the threat score.

Systems and methods configuring a unified threat machine learning model for joint content and user threat detection

Growing community of inventors

San Francisco, CA, United States of America

Wei Liu