This half-day workshop features six mini-sessions covering software, frameworks, and hardware.
NVIDIA AI Tech Workshop at NEURIPS
Sunday, December 2, 2018 | 8:00 AM – 12:30 PM
Palais des Congrès de Montréal, Room 511C Montréal CANADA
GPU-accelerated computing is transforming the landscape of computational science and AI. Come connect with the leading researchers and engineers from NVIDIA as we explore the future of GPU computing, provide the latest analysis of NVIDIA’s AI platforms, and demo our software.
TOPICS
Session One:
Improvements to the Core
NVIDIA® CUDA® Graphs, WMMA
NVIDIA CUDA Deep Neural Network library (cuDNN)
- Performance Over-Time Tensor Cores
- FP32 Fast Math
- Small Batch Improvement
- Attention Support
Session Two:
Productivity Tools for Your Research
Precision Training and Success Stories
Open Seq2Seq: NLP and Speech Toolkit for TensorFlow
Session Three:
Inference and Quantization
Quantized Inference
NVIDIA TensorRT™ 5 and TensorFlow Integration
Inference Server
Session Four:
Machine Learning with RAPIDS
Session Five:
Applied Deep Learning
NVIDIA Research
Building Speech Recognition with Synthetic Speech
Optimization of training neural collaborative filtering systems
Session Six:
Open Source Contributions
NV Rich in OSS contributions: DLA, CUTLASS, DALI, NCCL, AMP, OS2S, FWs, Research, Models: MaskRCNN, Transformer, GNMT
NVIDIA PRESENTERS
BRYAN CATANZARO
NVIDIA
VP, Applied Deep Learning
STEPHEN JONES
NVIDIA
Principal Software Architect
DILIP SEQUEIRA
NVIDIA
Principal Engineer, TensorRT
KARI BRISKI
NVIDIA
Senior Director, Accelerated Computing Software and AI Product
JOSH PATTERSON
NVIDIA
Director, AI Infrastructure
CHIP HUYEN
NVIDIA
Deep Learning Engineer
CARL CASE
NVIDIA
Architect
LEVS DOLGOVS
NVIDIA
Developer
TING-CHUN WANG
NVIDIA
Research Scientist
UJVAL KAPASI
NVIDIA
Senior Director, Software & Engineering
EVENT IS FREE AND OPEN TO THE PUBLIC BUT REQUIRES REGISTRATION