Using cloud-based, GPU-accelerated AI for financial risk management – HPCwire | Region & Cash

There are strict rules for financial institutions, with a number of global regulators issuing financial compliance requirements. Financial institutions face many challenges and legal responsibilities for risk management, compliance violations and failure to detect financial fraud.

Fraud in financial systems can occur in unstructured data or communications such as audio, images and chat. It is difficult to identify this type of scam as the data contains minimal markers for identification, which requires advanced analysis and techniques to uncover discrepancies.

Banks and other financial institutions face hefty fines for breaches of regulations and compliance. The Financial Times reports that “analysis by Behavox found that in 2021, only 0.0024% of voice-based communications intercepted by Behavox and 0.0002% of text analyzed by Behavox were deemed to be of concern for regulatory and compliance fines.” are high. Banks worldwide were fined $15 billion for such violations in 2020 alone.”

A large number of financial regulatory updates are released each year. Compliance officers and finance professionals seek to accurately and quickly identify criminal activities such as credit card fraud, insider trading, market manipulation, money laundering and trade violations. However, the sheer volume of transactions and changes in regulations make this an almost impossible task.

Importance of AI in financial risk management

Artificial intelligence (AI) is increasingly being used as a financial fraud risk management and compliance tool to combat a variety of financial regulatory violations. Models for AI, machine learning (ML) or deep learning (DL) offer new possibilities for regulators and compliance officers. AI models perform analytical tasks autonomously, ingesting large amounts of data and then spotting patterns in that data that indicate fraud or non-compliance. AI solutions help reduce the likelihood of human error that can lead to missing out on financial fraud, valid transactions being flagged as fraudulent, or costly penalties for non-compliance.

A 2022 State of AI in Financial Services survey sponsored by NVIDIA found that 78 percent of financial services firms say their organization is using accelerated computing to deliver AI-enabled applications through machine learning, deep learning, or high-performance computing. According to the survey, “With more than 70 billion real-time payment transactions being processed worldwide in 2020, financial institutions need robust systems to prevent fraud and reduce costs. Accordingly, fraud detection in payments and transactions was the most important AI use case for all respondents at 31 percent, followed by conversational AI at 28 percent and algorithmic trading at 27 percent.”

How cloud-based, GPU-accelerated AI meets risk management needs

Running AI models used for fraud prevention and risk management requires massive computing resources that are often unavailable in data centers. NVIDIA graphics processing units (GPUs) provide processing power for AI, ML or DL ​​models that cannot be achieved by central processing units (CPUs). For example, American Express uses NVIDIA GPUs and AI to detect anomaly scams and processes tens of millions of transactions every day. With the NVIDIA solution, American Express saw a 50x improvement over CPU processing.

Microsoft and NVIDIA have a long history of working together to help financial institutions detect fraud, manage risk, and meet financial compliance requirements. Using the Microsoft Azure cloud, NVIDIA GPUs and NVIDIA AI solutions provide scalable, accelerated resources needed to run AI/DL algorithms, routines, and libraries.

The partnership between Microsoft and NVIDIA makes NVIDIA’s powerful GPU acceleration available to financial institutions. The Azure Machine Learning service integrates NVIDIA’s open-source RAPIDS software library, which enables machine learning users to accelerate their pipelines with NVIDIA GPUs. Added NVIDIA TensorRT acceleration library to ONNX Runtime to accelerate deep learning inference. Azure supports NVIDIA’s T4 Tensor Core Graphics Processing Units (GPUs), optimized for cost-effective deployment of machine learning inference or analytics workloads.

Risk Management and Compliance Solutions

Organizations need help with financial management, locating financial risk and meeting compliance regulations. Microsoft provides governance checklists and best practices to help financial institutions manage risk and comply. The Microsoft Azure Security website provides valuable information about Microsoft Azure security and framework settings. A Microsoft Purview Compliance Manager software tool simplifies compliance and reduces risk by providing multicloud and continuous regulatory assessments, updates on new regulations, and a compliance assessment for the organization.

Another useful tool for detecting and managing financial fraud is the FIS Memento solution. This tool is a cross-channel solution that provides real-time protection using a combination of rules, AI, ML and statistical techniques to detect and prevent fraud in banking and payment channels. FIS Memento is offered as an on-premise or Azure cloud-based solution.


There is a massive increase in financial fraud and demands on financial organizations to provide risk management and to comply with financial regulations. Enterprises cannot adequately meet these demands with the infrastructure of many data centers.

Finance organizations need transformative technology to address the complexities of managing financial risk and meeting regulatory requirements. GPU-accelerated AI solutions from Microsoft and NVIDIA running on the Microsoft Azure cloud provide the technology to automate and streamline the risk management process.

Leave a Comment