How Financial Leaders Are Solving Data Challenges with AI

The modern financial system is built on massive datasets and the lightning-fast ability to analyze information — and AI is shaping up to be its backbone.

At this year’s GPU Technology Conference in Silicon Valley, leading financial institutions such as UBS, Citi, S&P Global, Capital One and Bloomberg will present talks on how machine learning applied to data analysis is allowing them to derive valuable insights on-demand.

Here are several sessions that those in the financial industry should consider attending to put their understanding of what’s next with AI at the top of the charts.

Deep Learning for NLP on Small Data Sets — Much of the progress using deep learning for natural language understanding and natural language generation has relied on building training models with massive labeled datasets. Hanoz Bhathena and Raghavachari Madhavan, data scientists at UBS, discuss how to approach these problems when massive amounts of labeled data aren’t available.
Using AI Machine Learning to Explore Large Streaming Financial Data Sets to Improve Market Making — Instead of making a single static model, Peter Decrem, director at Citi, explores how multi-GPU setups and large streaming datasets can be used to set up an online machine learning environment where thousands of strategies can be monitored and pockets of available liquidity uncovered.
Deep Learning Extraction for Counterparty Risk Signals from a Corpus of Millions of Documents — China’s technology industry has experienced rapid acceleration, which has allowed firms and governments to track and record vast amounts of data. However, there’s a wide gap in the ability to provide transparency to the exposed importing firms. Moody Hadi, R&D and Innovation group manager at S&P Global, shares how his company used deep learning to address this problem and create better transparency in these markets for its customers.
A Massively Scalable Architecture for Learning Representations from Heterogeneous Graphs — Working with sparse, high-dimensional graphs can be challenging in machine learning. C. Bayan Bruss and Athanassios Kintsakis, machine learning engineers at Capital One, explain a proposed architecture for neural graph embeddings, which are used extensively in unsupervised dimensionality reduction of large graph networks.
Machine Learning @ Bloomberg — The Bloomberg Terminal provides data, analytics, news, information and communication for professionals in business, finance, government, and philanthropy. Ian Hummel and David Eis of Bloomberg discuss how internal machine learning platforms apply advanced AI and GPU-accelerated compute to dozens of domains such as NLP, computer vision, time-series analysis, and personalization.

Check out the full finance session lineup at GTC, including a talk on how Veterans United Home Loans designed and implemented a VDI project powered with NVIDIA vGPUs. And register for the conference today.