Data Science Powers AI Breakthroughs in Financial Services and Beyond

AI experts discuss the latest industry developments at GTC.
by John Ashley

Data science is turning one of the world’s oldest industries inside out by applying accelerated computing to legacy financial data.

In an industry used to intensive computation, data scientists are the new financial engineers, generating insights that power fraud prevention models, recommendation systems, risk management and other financial services.

And they can stay on the cutting edge of the field by learning from the world’s leading developers, researchers, scientists and technologists at the GPU Technology Conference, running from October 5-9.

The Brightest Data Minds, Under a Single Virtual Roof 

In a unique pairing of ecosystems and technology, GTC brings together a global community of innovators to share the latest AI breakthroughs. Attendees can learn how AI enables digital enterprise transformation through robust data science environments, and how these technologies are being applied by everything from fintech startups to global financial institutions.

At GTC, AI experts will share their insights on real-time fraud prevention, AI chatbots, recommenders and more applications across the financial services industry.

NVIDIA CEO Jensen Huang will kick off the conference on October 5 with his highly anticipated keynote on the future of computing. Last spring, his keynote covered recent AI developments, including NVIDIA RAPIDS for data science, NVIDIA Merlin for recommendation systems and NVIDIA Riva for conversational AI.

Huang will also join a fireside chat with the CEOs of startups Babble Labs and CrowdAI to discuss AI trends in the startup ecosystem. Learn more about this conversation and register to attend here.

Next month, GTC will continue the conversation around AI with over 600 live and recorded talks, interactive demos, podcasts, webinars and Connect with the Experts sessions featuring live Q&A.

Here are a few can’t-miss sessions covering data science and financial services:

Accelerated Computing

AccernAdaptive Machine Learning Models That Evolve and Improve Over Time Without Expensive Retraining – Anshul Vikram Pandey, co-founder and CTO at Accern, covers the difference between adaptive machine learning and the broader umbrella of AI, and why machine learning is essential for systems that continuously process new data in fields such as healthcare, finance, logistics, retail and politics.

Amazon Web ServicesAccelerating Data Science with NVIDIA RAPIDS on AWS – Walk through the end-to-end process of running NVIDIA RAPIDS on AWS, starting with a single GPU instance. Learn how to run large-scale hyperparameter search experiments with RAPIDS and Amazon SageMaker, and run distributed machine learning using Dask clusters on Amazon EKS and Amazon ECS.

Capital OneUniversal Scalable Custom Machine Learning Estimators – Software engineering leaders from Capital One discuss how to build custom machine learning estimators that can scale from single-core to distributed cores using Dask, GPUs and NVIDIA RAPIDS.

NVIDIA State of RAPIDS: Bridging the GPU Data Science Ecosystem – Joshua Patterson, senior director of engineering at NVIDIA, discusses the latest engineering work, current benchmarks, new release features and future roadmap for NVIDIA RAPIDS, our end-to-end open-source data science accelerator.

Advancing Conversational AI

PayPal Optimized Dialogue Policy Learning for a Task-Oriented Chatbot Through Deep Reinforcement Learning – Using a Deep-Q dual network, PayPal has developed chatbots that automatically extract optimal dialogue policies from unannotated live chat logs. Rajesh Munavalli, distinguished data scientist at PayPal, explains why this dialogue policy manager plays a pivotal role in chatbots by guiding the conversation flow through a series of appropriate agent responses for successful completion of a given task.

Voca.ai – Delivering Human AI Experience Through Voice Interactions – Can robots talk like humans? Voca.ai shares its journey using conversational AI and deep learning to engineer a human-like, telephony-based voice bot. Alan Bekker, CTO and co-founder of voca.ai, joins Jackie Assa, the company’s head of AI, to discuss deep learning technologies for building voice bots.

Real-Time Fraud Prevention

HDFC Bank – Anomaly and Outlier Detection for Fraud Prevention Based on Customer Profiles Using GPU AccelerationHDFC built an anomaly and outlier detection solution for fraud prevention, based on customer profiles and GPU acceleration. Join Bharath Shasthri, head of data science at HDFC Bank, to learn about the details of building engineering workflows to detect outliers in features like age, income and credit history using the binning and bucketing technique.

PayPal – Generating Ensemble-Based Adversarial Attacks and Defense Methods for Robust Online Fraud Detection – PayPal neutralizes sophisticated ensemble-based adversarial attacks with generative reinforcement learning models to stop online payment fraud. Hear from Nitin Sharma, senior research scientist of the AI Research Group at PayPal, about how financial institutions can deploy generative defensive models in large-scale fraud detection systems to protect against such large-scale attacks.

Take AI Courses Taught by NVIDIA Experts

The NVIDIA Deep Learning Institute offers hands-on training in AI, accelerated computing and accelerated data science. Developers, data scientists, researchers, students and IT professionals can get practical experience powered by GPUs in the cloud. DLI sessions at GTC have already reached maximum capacity, but you can register for the waitlists below, and sign up to be notified when future DLI trainings are available:

Don’t Miss These Special Events at GTC

With five continuous days of programming across all time zones, GTC also offers a host of special events. Make lasting connections with fellow GTC attendees by going on a Braindate, either one-on-one or in small groups. Enjoy a meal at home while meeting remotely with your peers in our popular Dinner with Strangers program — moderated by NVIDIA experts. And, for the first time, explore GTC’s AI Art Gallery with virtual studio tours and a live art social. See the stories behind visual artists using AI to create their digital masterpieces.

AI Innovation Never Sleeps

Join GTC online to learn how AI, data science and other technologies will transform the way people work by experiencing demos, interactive training, breakthrough research and more.

Explore the full GTC catalog to find data science and financial services sessions. And register today using event code CMFN to save on your digital conference pass.