Power/Performance Bits: April 16


Faster CNN training Researchers at North Carolina State University developed a technique that reduces training time for deep learning networks by more than 60% without sacrificing accuracy. Convolutional neural networks (CNN) divide images into blocks, which are then run through a series of computational filters. In training, this needs to be repeated for the thousands to millions of images... » read more

Combining SLAM And CNN For High-Performance Augmented Reality


Robotics and headsets or goggles are the most common hardware devices requiring AR/VR/mixed reality, and AR is coming to mobile phones, tablets, and automobiles as well. For hardware devices to see the world around them and add to that reality with inserted graphics or images, they need to determine their position in space and map the surrounding environment. Simultaneous localization and ma... » read more

The Automation Of AI


Semiconductor Engineering sat down to discuss the role that EDA has in automating artificial intelligence and machine learning with Doug Letcher, president and CEO of Metrics; Daniel Hansson, CEO of Verifyter; Harry Foster, chief scientist verification for Mentor, a Siemens Business; Larry Melling, product management director for Cadence; Manish Pandey, Synopsys fellow; and Raik Brinkmann, CEO ... » read more

The Winograd Transformation


Cheng Wang, senior vice president of engineering at Flex Logix, explains how the Winograd Transformation applies to convolutional neural networks. https://youtu.be/E7QJUby9x-I » read more

Impact Of IP On AI SoCs


The combination of mathematics and processing capability has set in motion a new generation of technology advancements with an entire new world of possibilities related to Artificial Intelligence. AI mimics human behavior using deep learning algorithms. Neural networks are what we define as deep learning, which is a subset of machine learning, which is yet a subset of AI, as shown in Figure 1. ... » read more

The Week In Review: Manufacturing


Trade wars China and the United States are in the midst of a trade war. Click here for the latest from CNN. Meanwhile, click here for a list of the winners and losers so far. Display Supply Chain Consultants, a research firm, provides more insights from a hi-tech perspective. Gary Shapiro, president and CEO of the U.S.-based Consumer Technology Association (CTA), issued a statement abo... » read more

Low-Power Deep Learning Implementation For Automotive ICs


Examples of automotive applications abound where high-performance, low-power embedded vision processors are used, from in-car driver drowsiness detection, to a self-driving car ‘seeing’ the road ahead with pedestrians, oncoming cars, or the occasional animal crossing the road. Implementing deep learning in these types of applications requires a lot of processing power with the lowest possib... » read more

Predictions: Manufacturing, Devices And Companies


Some predictions are just wishful thinking, but most of these are a lot more thoughtful. They project what needs to happen for various markets or products to become successful. Those far reaching predictions may not fully happen within 2018, but we give everyone the chance to note the progress made towards their predictions at the end of the year. (See Reflection On 2017: Design And EDA and Man... » read more

Software Framework Requirements For Embedded Vision


Deep learning techniques such as convolutional neural networks (CNN) have significantly increased the accuracy—and therefore the adoption rate—of embedded vision for embedded systems. Starting with AlexNet’s win in the 2012 ImageNet Large Scale Visual Recognition Challenge (ILSVRC), deep learning has changed the market by drastically reducing the error rates for image classification and d... » read more

The Efficiency Problem


Part one of this report addressed the efficiency problem in neural networks. This segment addresses efficiencies in training, quantization, and optimizing the network and the hardware. Minimize the Bits (CNN Advanced Quantization) Training a CNN involves assigning weight vectors to certain results, and applying adaptive filters to those results to determine the positives, false positives, a... » read more

← Older posts