An IntelVision Accelerator Design Product Intel Vision Accelerator Design with Intel Arria 10 FPGA 2019 Accelerate To The Future Powered by Open Visual Inference & Neural Network Optimization (OpenVINO) toolkit Ubuntu 16.04.3 LTS 64bit, CentOS 7.4 64bit (Windows & more OS are coming soon). Supports popular frameworks...such as TensorFlow, MxNet, and CAFFE. Easily deploy open source deep learning frameworks via Intel Deep Learning Deployment Toolkit . Provides optimized computer vision libraries to quick handle the computer vision tasks. Intel FPGA DL Acceleration Suite. A Perfect Choice for AI Deep Learning Inference Workloads OpenVINO toolkit Deep learning and inference Deep learning is part of the machine learning method. It allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction.Deep neural network and recurrent neural network architectures have been used in applications such as object recognition, object detection, feature segmentation, text-to-speech, speech-to-text, translation, etc.In some cases the performance of deep learning algorithms can be even more accurate than human judgement. Al Sense, learn, reason, act, and adapt to the real world without explicit programming Perceptual Data Analytics Machine Learning Build a representation, Understanding Computational methods that use learning algorithms to build a model from data query, or model that Detect patterns (in supervised, unsupervised, semi-supervised, or reinforcement mode) enables descriptive, in audio or interactive, or predictive visual data analysis over any amount Deep Learning of diverse data Algorithms inspired by neural networks with multiple layers of neurons that learn successively complex representations Convolutional Neural Networks (CNN) DL topology particularly effective at image classification In the past, machine learning required researchers and domain experts knowledge to design filters that extracted the raw data into feature vectors. However, with the contributions of deep learning accelerators and algorithms, trained models can be applied to the raw data, which could be utilized to recognize new input data in inference. Learning from existing data Predict new input data Training Inference Training Dataset New Data Backward Trained Model Forward Forward DOG DOG CAT 1