Money Laundering Detection with AI

A method for detecting anomalous activity, like fraud or money laundering, using a dynamic approach that combines transaction analysis and social network analysis. The method involves calculating anomaly scores for user transactions using unsupervised and supervised algorithms trained on transaction attributes. It also calculates network anomaly scores based on interconnected user profiles. The potential for anomalous activity is determined by combining the transaction and network scores. This provides dynamic detection from multiple angles rather than just transaction analysis alone.

Source

Anti-money laundering suspicious transaction monitoring system that uses real-time collection of transaction and personal data to improve accuracy and reduce false positives compared to just relying on bank transaction data. The system collects details like account balances, property, assets, insurance, and income for monitored individuals. It associates this with their basic info and continuously monitors for suspicious activity. This comprehensive data allows more accurate detection of money laundering beyond just bank transactions.

Source

Anti-money laundering suspicious transaction identification method and device that combines rule-based and machine learning models to improve accuracy and coverage compared to using just one approach. The method involves using a rule model and a machine learning model in parallel, weighting their results, and using the weighted average to determine if a transaction is suspicious. The models are iteratively trained with updated data to keep up with evolving money laundering techniques.

Source

Determining the likelihood of a connection between funds entering and exiting a cryptocurrency tumbler, to identify laundered funds. The method involves using classical computers to formulate an optimization problem representing the tumbler's flows. This is converted to a quantum-computable quadratic unconstrained binary optimization (QUBO) form. A quantum computer solves the QUBO to find matrices representing the tumbler's internal flows. Classical post-processing generates a probability matrix showing likelihoods of connections between input and output funds.

Source

Detecting money laundering activities hidden in large volumes of legitimate transactions using neural networks to compare financial transaction patterns. The method involves generating target and baseline vectors representing transaction activity for a specific party and overall region, respectively. A Siamese neural network compares the vectors to detect deviations indicative of potential money laundering. A drift score between the vectors is calculated, constrained in a learned space between non-money laundering and money laundering transactions. An alarm is triggered if the drift score indicates potential money laundering.

Source

An anti-money laundering (AML) suspicious transaction identification method using reinforcement learning to improve coverage and accuracy compared to traditional methods like rule engines and supervised learning. The method involves using reinforcement learning to train a model that takes customer and environmental data as input and outputs whether the transaction is suspicious for money laundering. The model learns by minimizing loss based on true labels, allowing it to adapt to changing risks and respond faster than offline training. The method involves obtaining transaction data, generating a state vector, training a loss prediction network, selecting actions using a search strategy, and updating parameters based on true labels.

Source

System for more effectively and efficiently identifying potentially suspicious cash structuring activity that may be related to money laundering or other financial crimes. The system uses a deep learning model with convolutional layers and non-linear activation functions to analyze daily transaction patterns and global transaction patterns separately. It extracts daily patterns from the daily transactions, aggregates them, and combines them with global transaction patterns. This aggregated input is fed into a classifier to generate scores for the account. If the score exceeds a threshold, it alerts for potential suspicious activity. The model uses techniques like dilated convolutions, L2 weight decay, Adam optimization, and ensemble teacher distillation to improve performance.

Source

Identifying suspicious activity using a multidimensional approach that goes beyond just transaction data to improve accuracy and adaptability compared to existing systems. The method involves preprocessing monitoring data using a plurality of vectors, applying rules to each vector, and triggering actions based on the rules. The vectors capture context beyond just transactions, such as party details, location, etc. This multidimensional approach allows identifying suspicious activity that would be missed by single-dimensional transaction-only analysis.

Source

Money laundering structure detection using hypergraph heterogeneous graph hierarchical representation learning to improve accuracy and applicability of existing technology for detecting money laundering structures. The method involves modeling transaction networks as heterogeneous graphs with meta-paths representing money laundering scenarios. Heterogeneous graph hierarchical representation learning is used to mine structure features from transaction, user, and association layers. This hierarchical representation greatly improves money laundering structure detection compared to homogeneous network methods.

Source

Automatically detecting suspicious money laundering transactions using artificial intelligence models. The method involves using two AI models - one to identify transactions as potential money laundering targets, and another to score the likelihood of money laundering. Transactions flagged by the first model are further analyzed by the second model to determine a suspicion level. If the level exceeds a threshold, the transaction is reported as suspicious. The AI models are trained using financial transaction data to learn patterns indicative of money laundering.

Source

Using AI to predict the growth of money laundering networks in financial systems and prioritize action against them. The method involves monitoring networks already identified as potential money laundering networks, predicting their future growth size based on past changes, and prioritizing networks with high growth potential for intervention. The prediction is done using AI models on historical network data.

Source

Using machine learning to improve anti-money laundering (AML) analysis by accurately identifying accounts and account holders with money laundering risk. The method involves obtaining a dataset of accounts and applying a trained algorithm to generate money laundering risk scores for each account holder. Accounts with high risk scores are flagged for investigation. The algorithm uses a dataset of aggregated financial data from multiple sources to analyze account variables like transactions, characteristics, and relationships. It provides actionable recommendations and prioritization for compliance teams.

Source

Tree-based user identification for real-time anti-money laundering monitoring. The method involves using tree models, like decision trees, to identify potential money laundering activities in real-time transactions. It leverages offline daily models trained on historical data to predict money laundering risks for users. This daily model identifies high-risk users for review. The daily model's feature importance is used to build a lightweight real-time model that can quickly identify potential money laundering in real-time transactions. The real-time model is more lightweight and fast compared to the daily model due to feature selection.

Source

Machine learning techniques for reviewing alerts in regulatory settings like anti-money laundering (AML) that provide insights and a flexible interface to improve efficiency and accuracy. The techniques involve calculating context-aware representations of entities like customers and transactions using self-supervised graph neural networks. These representations can be used to derive insights like clustering, anomaly scoring, and period detection to aid in AML alert review. The representations capture entity behavior based on surrounding context from a bipartite graph.

Source

Using machine learning to identify potential criminal activity involving financial institutions' customers by analyzing news alerts. The method involves extracting entities from news alerts, resolving them to known customers, enriching the alerts with customer data, predicting the likelihood of a suspicious activity report, and transmitting the enriched alerts to the appropriate investigator.

Source

Improving anti-money laundering detection by identifying money laundering rings instead of just individual accounts. The method involves clustering accounts based on their transaction behavior, then identifying money laundering gangs among the clusters. This leverages group-level transaction patterns to improve detection accuracy compared to analyzing accounts independently.

Source

Detecting financial crimes like money laundering using machine learning models that leverage network effects between financial entities. The method involves generating network features by applying risk indicators to a graph model of financial entities and their relationships. These network features are fed into machine learning models trained on both network and non-network features to predict financial crimes. Alerts are generated when crimes are predicted, identifying the involved entities. The network representation helps reveal hidden connections and improve crime detection compared to traditional non-network features.

Source

Detecting fraud in financial networks without relying on hardcoded rules by using graph neural networks to identify suspicious transaction patterns. The method involves generating a transaction graph representing the financial network with synthetic suspect transactions. Subgraphs are extracted and used to train a graph neural network model to classify subgraphs as suspect. This model is then applied to new financial networks to locate potentially fraudulent transaction patterns. The subgraph extraction dynamically selects nodes based on their transaction history to capture related activity.

Source

Anti-money laundering (AML) identification method, device, system and medium based on fusion models that combines decision models and feature extraction models to improve AML detection accuracy and efficiency. The method involves preprocessing transaction samples, extracting features, screening important variables using a decision model, and feeding them to a feature extraction model for AML classification. This fusion approach allows automated identification without manual screening.

Source

Automated optimization of anti-money laundering model thresholds using data analysis and trial calculations to improve accuracy, reduce false alerts, and increase timeliness compared to manual tuning. The optimization involves extracting discrete characterization parameters from transaction rules based on indicator values, training the model with these parameters, and finding optimal threshold combinations through massively parallel trial calculations.

Source

Fast anti-money laundering detection using graph neural networks to accurately monitor dynamic transaction graphs in real time and improve detection accuracy compared to single transaction analysis. The method involves a two-step process: pre-access and post-monitoring. In pre-access, blacklists and money laundering rules are used to directly block high-risk transactions. In post-monitoring, graph neural networks predict the money laundering probability of transactions based on network relationships. This allows detecting complex laundering patterns beyond individual transactions.

Source

A machine learning engine with a trained model for detecting potentially fraudulent or illegal transactions like money laundering, along with a narrative generation component that explains why the transaction was flagged. The model uses features like transaction type, accounts, amounts, addresses, sources, etc. to identify potentially suspicious transactions. If a transaction is flagged, the narrative generation component provides a detailed explanation of the reasons behind the flag. This helps human reviewers understand the rationale for the flagged transaction.

Source

Detecting transaction laundering in the payment ecosystem using machine learning models to flag potentially fraudulent merchants and transactions. The system uses multiple machine learning models to analyze merchant and transaction data for indicators of transaction laundering. If the models collectively indicate a high likelihood of laundered transactions, further investigation is triggered. The models can also have decision trees and weighted incorrect classifications to improve accuracy. The approach involves using multiple models in conjunction to improve detection of transaction laundering compared to relying on a single model.

Source

A system for detecting money laundering activity in bank accounts using machine learning transformations. The system identifies sets of related transactions within a time period, applies machine learning transformations to those sets, and identifies anomalous activity based on the transformed results. This allows more accurate and frequent detection of money laundering patterns compared to traditional short-term window analysis.

Source

Detecting illicit activity like money laundering that escapes rule-based systems by training a neural network generator to create synthetic illicit activity and a discriminator to distinguish between real and synthetic activity. The generator is trained using feedback on how to avoid triggering rule-based systems, and a money laundering objective. The discriminator learns to distinguish between real and synthetic activity based on the tensor representation of connected graphs of transactions. By focusing on generating illicit activity that avoids rules and maximizes money laundering goals, the generator can identify weaknesses in rule-based systems.

Source

Detecting fraudulent financial activity involving cryptocurrencies and money mules using AI. The method involves analyzing user banking account activity to identify potential money mule accounts. It combines metrics like frequency of cryptocurrency transactions and technology adaptation with risk scores to more accurately detect money mule accounts compared to just using risk scores. The metrics are analyzed using machine learning models.

Source

Generating suspicious activity reports using machine learning that can efficiently analyze large transaction histories to identify potential financial crimes. The system compresses transaction data using lossy compression methods and inputs it into a trained neural network to score the likelihood of criminal activity. If the score exceeds a threshold, it triggers generation of a suspicious activity report. This allows leveraging historical transaction data without analyzing all variables simultaneously, reducing false positives and improving efficiency.

Source

Using AI to validate security events like fraud or money laundering by collecting user information related to specific events and analyzing it with trained AI models. The models compare user data against baselines to generate risk metrics. This allows investigating security events in context rather than relying solely on thresholds or anomalies.

Source

Automatically detecting money laundering gangs in financial transactions using community detection algorithms and regularization engines. The method involves dividing a network of customer transactions into communities using a community discovery algorithm. It then screens for suspicious money laundering gangs by looking for suspicious cases and screening results from regularization engines within each community. This improves efficiency and accuracy compared to manually merging suspicious cases across customers.

Source

Detecting abnormal financial transactions by analyzing low-level transaction data using artificial intelligence. The method involves converting financial transaction data into low-level data formats like assembly language, hex code, binary, ASCII, and EBCO. This low-level data is then analyzed by AI models to predict probabilities of abnormal transactions. By converting and analyzing the raw transaction data at a low level, it becomes harder to manipulate and evade detection.

Source

An abnormal financial transaction detection system that uses low-level data analysis to effectively identify suspicious transactions and automate detection with artificial intelligence. The system converts financial transaction data into low-level representations like assembly language, hexadecimal, binary, ASCII, and EBCO. It analyzes these low-level forms using AI models to detect abnormal transactions. This provides a more robust and automated fraud detection compared to analyzing high-level transaction data.

Source

Transaction data processing method for financial analysis using graph convolutional networks (GCNs) and transformers to extract spatial-temporal features from transaction graphs. The method involves obtaining transaction data for an account over multiple time windows, extracting spatial features using GCNs and temporal features using transformers, and generating a feature representation for the account based on the combined spatial-temporal information. This representation can be used for downstream analysis tasks like credit risk modeling, fraud detection, and money laundering detection.

Source

Detecting money laundering using a machine learning system that prioritizes alerts and links accounts to improve detection accuracy. The system uses a profile-based feature construction technique to create features from entity data like KYC info and transaction history. It calculates an AML Threat Score using supervised learning on labeled cases to prioritize alerts. The score is calibrated over time using self-calibrating outlier detection. Accounts are efficiently linked through behavior-sorted lists that capture frequent transactions. This helps propagate risk information between associated accounts. The system can also be extended to analyze emerging payment systems like cryptocurrencies by leveraging KYC data and legal exchange information.

Source

A method and system for identifying money laundering transactions using a multi-layer perceptron neural network algorithm. It aims to improve efficiency and reduce false positives in identifying suspicious transactions compared to rule-based systems. The method involves extracting features from transaction data, standardizing them, reducing dimensionality, and training a multi-layer perceptron neural network model using a labeled dataset of known money laundering transactions. The trained model can then be used to predict if new transactions are suspicious.

Source

Method and system for identifying money laundering transactions using machine learning to improve efficiency and reduce false positives compared to rule-based methods. The method involves training a machine learning model using XGBoost algorithm to analyze features of transaction data and predict the risk of money laundering. The features are extracted, standardized, and dimensionally reduced before training. The trained model can then provide a money laundering risk score for new transactions.

Source

Simulating realistic transaction data using reinforcement learning to train financial crime detection models when real customer data is limited. The simulation involves generating artificial customer profiles and having intelligent agents learn to behave like "standard" customers by imitating their transaction patterns. The agents observe past behavior, then generate their own transactions that match the standard. This allows creating a large set of simulated transactions to train the detection model without using actual customer data. The simulation involves clustering real customer transactions to find patterns, then having agents learn those patterns to generate similar transactions.

Source

A machine learning based anti-money laundering behavior recognition system that uses data processing and neural networks to accurately identify accounts involved in money laundering. The system takes transaction data from multiple accounts over a time period and processes it to obtain features like number of transactions, max transaction amount, abnormality index, etc. These features are then input into a pre-trained neural network to identify if an account is suspicious for money laundering. The system uses clustering to group accounts based on similarity and then determines risk levels for each group.

Source

Identifying sets of financial transactions in a network that are likely to involve fraud or money laundering. The method involves selectively analyzing subsets of transaction sets based on properties of their initial messages. If the initial message is from an untrusted source (e.g. to a new account) or involves rapid transfers between unrelated accounts, it's more likely to be fraud. The method generates scores for each node in a set based on their behavior and identifies sets with high scores as potential fraud.

Source

Using logistic regression to identify abnormal transactions and customers for anti-money laundering purposes. The method involves loading customer transaction data into a trained logistic regression model to obtain a score indicating the suspiciousness of the transactions. This score is compared against a threshold to categorize as high or low suspicion. The logistic regression model is built using feature selection, data standardization, dimensionality reduction, and resampling techniques.

Source

Real-time detection of financial transactions suspicious for money laundering by processing high-speed streaming financial data using an incremental clustering algorithm. The algorithm, called Fused-Density (FD), continuously learns and adapts clusters as data arrives. It reads financial data points, maintains grids and provisional clusters, associates points, systems clusters, trims low-weight ones, and forms shape devise clusters after a threshold time. This enables real-time detection of evolving money laundering trends in streaming financial data.

Source

Detecting anomalous financial behavior like money laundering using AI to learn from traces of observed behavior. The method involves receiving behavior data, determining behavior traces based on states and actions, and classifying the traces into behavior categories. An AI model called CABBOT learns to classify behavior based on traces. It uses a relational kNN classifier that compares traces to find closest matches and classifies by mode. This allows on-line classification of behavior as it unfolds. The AI can also simulate financial behavior using automated planning. It generates richer, more realistic traces with features like interleaved standard activity and network structures.

Source

Blockchain-based method and device for detecting money laundering using reinforcement learning on blockchain transaction networks. The method involves dividing blockchain transaction data into time slices, applying reinforcement learning algorithms to find abnormal money laundering structures in each slice, and determining the laundering type from the identified structures. This allows finding potential laundering structures across time intervals, avoiding misidentifying long gaps.

Source

Accurately and timely identifying money laundering risk in transaction behavior using artificial intelligence techniques. The method involves steps like acquiring normalized transaction record samples, increasing positive sample count through data enhancement, selecting features using CNN, LGBM, Catboost models, and finally identifying money laundering risk using the selected features.

Source

Detecting unauthorized or suspicious financial activity using machine learning techniques on graph representations of financial data. The method involves preparing financial data as a graph with nodes representing entities like accounts, people, devices, etc, and edges representing relationships like transactions. Neural networks process the graph to calculate risk scores for entities indicating likelihood of unauthorized activity. This allows direct calculation of risk for entities without relying on user-defined heuristics. The graph representation is trained using neural networks like graph convolutional networks to extract features from the graph structure. The trained model can then process new graph representations to detect suspicious entities.

Source

Detecting fraudulent transactions in virtual assets using artificial intelligence to prevent money laundering and abnormal transactions. The system extracts transaction data and labels from virtual asset platforms, trains models to classify transactions into codes representing fraud types like money laundering or abnormal activity, and derives the specific fraud type for a given transaction based on matching codes. It assigns weights to codes for likely fraud and presents the results to users. This provides objective, accurate fraud detection without human error.

Source

A method for monitoring anti-money laundering data that combines big data and AI to effectively detect and prevent money laundering in financial institutions. The method involves using AI models to analyze large volumes of transaction data and customer information. It involves features extraction, model selection, and classification of suspicious cases into categories. The method aims to improve interpretability, model updating, and scalability compared to traditional manual methods.

Source

Detecting potential money laundering financial transactions using graph databases and adaptive AI techniques. The method involves receiving financial activity data for users, creating temporal knowledge graphs representing user interactions, encoding the graphs into vector representations, applying machine learning to predict money laundering, and flagging suspicious clusters for further analysis. By proactively identifying potential money laundering, it allows near real-time alerting to prevent transactions. The graph embedding and clustering techniques leverage the connected nature of financial networks to improve detection accuracy.

Source

Detecting potential money laundering financial transactions using graph databases and adaptive AI techniques to predict suspicious transactions in near real-time. The method involves creating temporal knowledge graphs representing users and their financial activities, encoding the graphs into vector representations, applying unsupervised machine learning algorithms to predict money laundering transactions, and flagging suspicious clusters for further investigation.

Source

Anti-money laundering monitoring method using migration learning and custom rules to improve accuracy and overcome the challenges of small data sets in financial risk assessment. The method involves training a target anti-money laundering monitoring model using a combination of custom rules and migration learning. This is done by inputting a target domain sample set into the target model to obtain target labels. The model consists of a custom rule model based on risk warnings and a migration learning model trained on source domain data. The weights of the custom and migration models are determined based on a maximum mean difference. This hybrid approach leverages custom rules for domain knowledge and migration learning for data distribution differences to overcome small data limitations.

Source

An anti-money laundering system using self-repairing machine learning models to improve accuracy and reduce manual intervention. The system has a money laundering recognition device that extracts high-dimensional full features from transactions. It uses these features as a consistent input for self-repairing model optimization. After each production run, feedback data is processed to generate new models without manual feature selection. This enables automatic and fast self-repair of the models as behavior characteristics change.

Source