Abstract: Embodiments provide methods and systems for merchant data cleansing in payment network. Method performed by server system includes accessing electronic payment transaction records from transaction database. Each electronic payment transaction record includes merchant data fields. Method includes determining set of electronic payment transaction records with ambiguous merchant data fields having matching probability scores less than predetermined threshold value computed by probabilistic matching model and identifying at least one issue for non-matching of each of set of electronic payment transaction records. Method includes determining data model based on at least one issue of each of set of electronic payment transaction records. Data model is one of: phone-to-city model, payment aggregator model, and merchant name normalization model.

Abstract: Methods and server systems for computing fraud risk scores for various merchants associated with an acquirer described herein. The method performed by a server system includes accessing merchant-related transaction data including merchant-related transaction indicators associated with a merchant from a transaction database. Method includes generating a merchant-related transaction features based on the merchant-related indicators. Method includes generating via risk prediction models, for a payment transaction with the merchant, merchant health and compliance risk scores, merchant terminal risk scores, merchant chargeback risk scores, and merchant activity risk scores based on the merchant-related transaction features. Method includes facilitating transmission of a notification message to an acquirer server associated with the merchant.

Abstract: Siamese neural networks (SNN) are configured to detect differences between financial transactions for multiple financial institutions and transactions for a target party. A first neural network of the SNN tracks transactions (target transactions) for a particular customer or financial institution over time and provides a target output vector. Similarly, a second neural network of the SNN tracks transactions (baseline transactions) for all or a plurality of financial institutions (e.g., within a region) over the same period of time and provides a baseline output vector. The transactions for all or a plurality of financial institutions act as a baseline of transactions against which potentially fraudulent or money laundering activity may be compared. Because Siamese neural networks account for temporal changes based on the baseline of transactions, sudden changes in target transactions will only trigger an alarm if such changes (e.g., deviations or drifts) are relative to a baseline of transactions.

Abstract: The disclosure herein describes a system and method for predictive identification of breached entities. Identification number and expiration date pairs associated with compromised records in a source file are analyzed to identify a set of candidate entities having records at least partially matching the source file data pairs having events occurring during a selected time period. Probability vectors are calculated for records associated with each identified entity. A divergence value is calculated which represents a distance between probability distribution vectors for each entity and probability distribution vectors for the source file. A predicted breached entity is identified based on the divergence values. The predicted breached entity is notified of the predicted breach. The notification can include an identification of the breached entity, identification of breached records, predicted time of breach, and/or a recommendation to take action to mitigate the predicted breach.

Abstract: Embodiments of present disclosure provide methods and systems for increasing transaction approval rate. Method performed includes accessing transaction features and determining via fraud model and approval model, first and second set of rank-ordered transaction features. Method includes computing difference in ranks of transaction features and determining set of utilized and unutilized transaction features and generating simulated authorizing model and computing simulated transaction approval rate and simulated fraud transaction rate for simulated authorizing model. Method includes generating plurality of proxy authorization models. Method includes computing transaction approval rates and fraud transaction rates for each of plurality of proxy authorization models and computing an increase in transaction approval rate and change in fraud transaction rate for each of plurality of proxy transaction approval models.

Abstract: Embodiments provide methods and systems for generating domain-specific text summary. Method performed by processor includes receiving request to generate text summary of textual content from user device of user and applying pre-trained language generation model over textual content for encoding textual content into word embedding vectors. Method includes predicting current word of the text summary, by iteratively performing: generating first probability distribution of first set of words using first decoder based on word embedding vectors, generating second probability distribution of second set of words using second decoder based on word embedding vectors, and ensembling first and second probability distributions using configurable weight parameter for determining current word. First probability distribution indicates selection probability of each word being selected as current word.

Abstract: Embodiments provide methods and systems for merchant data cleansing in payment network. Method performed by server system includes accessing electronic payment transaction records from transaction database. Each electronic payment transaction record includes merchant data fields. Method includes determining set of electronic payment transaction records with ambiguous merchant data fields having matching probability scores less than predetermined threshold value computed by probabilistic matching model and identifying at least one issue for non-matching of each of set of electronic payment transaction records. Method includes determining data model based on at least one issue of each of set of electronic payment transaction records. Data model is one of: phone-to-city model, payment aggregator model, and merchant name normalization model.

Abstract: A computing device for determining a new credit card number that is a continuation match with an old credit card number of a credit card account that has changed numbers comprises a processing element programmed to: receive transactional data for a plurality of credit card numbers, determine a plurality of old credit card numbers and a plurality of new credit card numbers, determine a plurality of clusters of new credit card numbers, convert the transactional data for each old credit card number and the associated cluster of new credit card numbers into snapshots with an image-like data format, train a modified siamese network with instances of snapshots of an old credit card number, a first new credit card number, and a second new credit card number, and use the modified siamese network to determine one new credit card number that is an upgrade of one old credit card number.

Abstract: Embodiments provide electronic methods and systems for improving edge case classifications. The method performed by a server system includes accessing an input sample dataset including first labeled training data associated with a first class, and second labeled training data associated with a second class, from a database. Method includes executing training of a first autoencoder and a second autoencoder based on the first and second labeled training data, respectively. Method includes providing the first and second labeled training data along with unlabeled training data accessed from the database to the first and second autoencoders. Method includes calculating a common loss function based on a combination of a first reconstruction error associated with the first autoencoder and a second reconstruction error associated with the second autoencoder. Method includes fine-tuning the first autoencoder and the second autoencoder based on the common loss function.

Abstract: Embodiments provide methods and systems for predicting overall account-level risks of cardholders. The method performed by server system includes accessing payment transaction data associated with a cardholder from a transaction database. Method includes generating a set of transaction features based on a set of transaction indicators. The method includes determining a plurality of network risk scores associated with the cardholder based on the set of transaction features and a set of trained machine learning models. The plurality of network risk scores includes a payment capacity risk score, a contactless payment risk score, and a set of account-level risk scores. The method includes aggregating the plurality of network risk scores to calculate an overall account risk score associated with the cardholder based on a statistical model. The method also includes transmitting a notification to the issuer server associated with the cardholder based on the overall account risk score.

Abstract: Embodiments provide artificial intelligence-based methods and systems for predicting account-level risk scores associated with cardholders. Method performed by server system includes accessing payment transaction data and cardholder risk data associated with cardholder. The payment transaction data includes transaction variables associated with past payment transactions performed at Point of Interaction (POI) terminals within a particular time window. Method includes generating cardholder profile data based on the transaction variables and the cardholder risk data. Method includes determining account-level risk scores associated with the cardholder based on cardholder profile data. Each account-level risk score of account-level risk scores is determined by a trained machine learning model. The account-level risk scores include a wallet reload risk score, an account reissuance risk score, and a transaction channel risk score.

Abstract: The disclosure herein describes a system and method for predictive identification of breached entities. Identification number and expiration date pairs associated with compromised records in a source file are analyzed to identify a set of candidate entities having records at least partially matching the source file data pairs having events occurring during a selected time period. Probability vectors are calculated for records associated with each identified entity. A divergence value is calculated which represents a distance between probability distribution vectors for each entity and probability distribution vectors for the source file. A predicted breached entity is identified based on the divergence values. The predicted breached entity is notified of the predicted breach. The notification can include an identification of the breached entity, identification of breached records, predicted time of breach, and/or a recommendation to take action to mitigate the predicted breach.

Abstract: Siamese neural networks (SNN) are configured to detect differences between financial transactions for multiple financial institutions and transactions for a target party. A first neural network of the SNN tracks transactions (target transactions) for a particular customer or financial institution over time and provides a target output vector. Similarly, a second neural network of the SNN tracks transactions (baseline transactions) for all or a plurality of financial institutions (e.g., within a region) over the same period of time and provides a baseline output vector. The transactions for all or a plurality of financial institutions act as a baseline of transactions against which potentially fraudulent or money laundering activity may be compared. Because Siamese neural networks account for temporal changes based on the baseline of transactions, sudden changes in target transactions will only trigger an alarm if such changes (e.g., deviations or drifts) are relative to a baseline of transactions.

Abstract: Embodiments provide methods and systems for processing unstructured and unlabelled data. A method includes generating, by a processor, a structured and unlabelled training dataset from an unstructured and unlabelled dataset. The method includes categorizing the structured and unlabelled training dataset into a plurality of clusters by executing an unsupervised algorithm. Each cluster of a selected set of clusters from the plurality of clusters is labelled with an applicable label from a set of labels. The method includes executing a supervised algorithm to generate a trained supervised model using a labelled training dataset including the set of labels and an input dataset generated from plurality of datapoints present in each cluster of the selected set of clusters. The method includes generating a Labelled Data1 (LD1) by executing the trained supervised model configured to assign applicable label from the set of labels to each datapoint of the structured and unlabelled training dataset.

Abstract: Embodiments provide methods and systems for generating domain-specific text summary. Method performed by processor includes receiving request to generate text summary of textual content from user device of user and applying pre-trained language generation model over textual content for encoding textual content into word embedding vectors. Method includes predicting current word of the text summary, by iteratively performing: generating first probability distribution of first set of words using first decoder based on word embedding vectors, generating second probability distribution of second set of words using second decoder based on word embedding vectors, and ensembling first and second probability distributions using configurable weight parameter for determining current word. First probability distribution indicates selection probability of each word being selected as current word.

Abstract: Embodiments provide methods and systems of predicting server failures. A method may include accessing distinct log clusters representing instructions executed in server, applying first density machine learning model over input vector of distinct log clusters, with length equal to number of distinct log clusters, for obtaining first prediction output, applying first sequential machine learning model over time length sequence of distinct log clusters for obtaining second prediction output, applying second density machine learning model over input vector for obtaining third prediction output, applying second sequential machine learning model over time length sequence of distinct log clusters for obtaining fourth prediction output, aggregating first, second, third and fourth prediction outputs by ensemble model, and predicting likelihood of next log clusters to have anomalous behavior based on the aggregating. First density and first sequential models are trained by normal logs.