GTC Data Science Presentations

Starting on October 5-9, This fall’s GTC will run continuously for five days, across seven time zones. The conference will showcase the latest breakthroughs in data science, and many other GPU technology interest areas. 

Attend live events in the time zone that works best for you, or browse an extensive catalog of on-demand content showcasing innovative uses of data science technology.

Here’s a preview of some of the data science sessions we’re working on for GTC.

Universal Scalable Custom Machine Learning Estimators: In this session, we will discuss how to write estimators that can scale from single-core to distributed cores and on NVIDIA GPUs; all while returning appropriate data structures for downstream dependencies.

Generating Ensemble-Based Adversarial Attacks And Defense Methods For Robust Online Fraud Detection: This talk introduces a cohort of domain-oriented adversarial attacks based on an ensemble of attack vector generation techniques, to optimize for discovery of sparse perturbation spaces yielding decisive changes in fraud adjudication.

Google Cloud Dataproc with GPUs Improves Latency and Reduces Costs for Select Spark Workloads: In this talk, we’ll discuss how Dataproc and NVIDIA GPUs support Spark workloads using RAPIDS Accelerator for Apache Spark, an open-source plugin that leverages columnar data processing with RAPIDS to accelerate ETL processing without changes to the user’s query code.

Register now>