Intel on AI

In this Intel on AI podcast episode: There are many individuals (makers) creating artificial intelligence (AI) enabled solutions and many of which may help drive AI to the next level. Yet, there is a challenge for makers when moving from an inexpensive board used for development to large scale production that requires much more specialization and infrastructure. Aling Wu, the AI and IoT Alliance Director at AAEON, joins the Intel on AI Podcast to discuss how the AAEON UP Board Computer Board family for Professional Makers helps bridge the gap from maker market to large scale production. She highlights how AAEON works directly with makers to take their design and actually produce and ship in volume production in 6-12 month period of time. Aling also discusses how AAEON provides makers with easy shopping solutions for development boards with UP shop and a forum where makers can discuss and refine their solution called UP Community. She talks about how important AI on the edge is for the development of the AI industry and how AAEON uses Intel’s broad portfolio of technologies including FPGAs (Field Programmable Gate Array) and Intel Core process to achieve performance on the edge. To learn more, visit: aaeon.com Visit Intel AI Builders at: builders.intel.com/ai

19 Sep 201912min

In this Intel on AI podcast episode: As artificial intelligence (AI) technologies rapidly advance many companies would like to take advantage of the potential of AI but can’t because they do not have the resources or skills for modeling and engineering. Yuan Shen, Founder and CEO of OneClick.AI, joins Intel on AI to talk about how OneClick enables businesses with no knowledge of AI and coding to solve complex business problems and drive real ROI with only a few clicks. He discusses how the OneClick.ai Automated Deep Learning (AutoDL) platform automates every step of AI model development supporting many different use cases including numeric prediction, classification, forecasting, recommendation, and image object detection. Yuan talks about how OneClick.AI has worked with the Intel AI Builders program to optimize deep learning on Intel Xeon Scalable processors using Intel optimization libraries and how they have been able to see great performance increases due to using those optimizations. To learn more, visit: oneclick.ai Or: builders.intel.com/ai/membership/oneclickai

13 Sep 201911min

In this Intel on AI podcast episode: Today athletes are always looking for new ways to analyze and improve their performance through video analysis and sports enthusiasts love to experience sports in new and innovative ways, but 3D capture technology can be difficult to program and expensive to access. Paul Kruszewski, the CEO and Founder of wrnch AI, joins the Intel on AI podcast to talk about the markerless motion capture real-time 3D tracking that wrnch is enabling which is helping to democratize that motion capture technology. He discusses how the wrnch solution can enable any 2D camera to become a motion capture device by using a deep learning trained model that tracks 23 skeletal points to do real-time analysis on human movement. Dr. Paul highlights how this technology can be used not just for sports training and in-home exercise guidance, but can even span many healthcare and even autonomous driving applications. He expresses how wrnch heavily benefits from breadth of hardware options that Intel delivers and how utilizing the Intel OpenVINO toolkit has enabled wrnch solutions to be available to developers around the world to do real-time AI on Intel Architecture. To learn more, visit: wrnch.ai Visit Intel AI Builders at: builders.intel.com/ai

9 Sep 201911min

In this Intel on AI podcast episode: Regular inspection of underground water and sewage pipelines is essential in the water utility industry to prioritize maintenance tasks and avoid pipe leakage, breakage, and blockage, which might result in property damage or safety hazards. Traditional inspection with videos being captured by a remotely operated rover and then manually analyzing footage is cumbersome, time-consuming, and error prone. Sundar Varadarajan, Consulting partner for AI/ML at Wipro Technologies, joins the Intel on AI podcast to discuss how Wipro is using artificial intelligence (AI) to automatically process video scans and identify, grade, and score overall pipe quality. Utilizing the Intel Optimization for TensorFlow and Intel OpenVINO Toolkit, Sundar explains how Wipro enables their customers to perform image inferencing in real time allowing pipeline inspections to be incredibly more efficient, consistent, and accurate. Sundar also talks about how Wipro is working with the Intel AI Builders program to optimize their solutions for the latest Intel technologies and lists several other solutions focused on concrete surface crack detection and medical imaging. To learn more, visit: wipro.com Visit Intel AI Builders at: builders.intel.com/ai

8 Sep 201910min

In this Intel on AI podcast episode: Many large enterprises need more accuracy and atomization for planning and purchasing. Yet, when generating and analyzing such a massive volume of data, the algorithms use an enormous amount of memory and processing can be slow, problematic, and often not calculate correctly at all? Dionisio Agourakis, the CEO at J!Quant, joins the Intel on AI podcast to talk about how J!Quant has a diverse portfolio of products involving deep learning (DL) and time-series prediction for stock optimization, demand forecast, and profit forecast to help their customers. He talks about how J!Quant helps enable a data-driven culture within their customers’ decision-making processes in order to stay relevant, profitable and open to new opportunities. Dionisio also discusses a specific use case utilizes the 2nd Generation Intel Xeon Scalable processors to tackle a memory bounded algorithm that a customer had and were able to successfully process the inference using Intel processors and Intel Optimizations for Tensorflow. To learn more, visit: jquant.com.br builders.intel.com/ai/membership/jquant software.intel.com/en-us/frameworks Visit Intel AI Builders at: builders.intel.com/ai

29 Aug 201911min

In this Intel on AI podcast episode: Deep neural networks (DNNs), which are arguably the most powerful form of AI today, are difficult to build, run, and explain. Such challenges constitute significant roadblocks for their adoption in their enterprise. Ironically, AI itself can be used to assist data scientists and developers to build and evaluate DNNs. Sheldon Fernandez, CEO of DarwinAI, joins us to talk about how DarwinAI is using this ‘AI building AI’ method in their Generative Synthesis platform. Sheldon explains how their technology reduces the complexity in designing high-performance deep learning solutions and also facilitates explainable deep learning, which allows a user to understand why a network makes the decisions it does. Finally, he describes a recent analysis conducted by the Intel AI Builders team, where Darwin-generated networks coupled with Intel Optimizations for TensorFlow were able to deliver up to 16.3X performance increase on ResNet50 and up to 9.6X on NASNet workloads. To learn more, visit: darwinai.ca Visit Intel AI Builders at: builders.intel.com/ai

21 Aug 201915min

In this Intel on AI podcast episode: Millions of dollars are lost by retailers worldwide each year as a result of out of stock products, overstocking, and shrinkage. Out of stock products also incur other costs, such as the reduced impact of in-store promotions and the time employees use dealing with customer requests about missing items. Ceyhun Burak Akgül, the Co-Founder and CTO of Vispera, joins the Intel on AI podcast to discuss how Vispera image recognition and retail real-time monitoring solutions are enabling businesses to use cameras to track product stocks and customer behavior to analyze and react to better serve the customer and business in real time. He describes the Vispera ShelfSight solution that is powered by the Intel Movidius Neural Compute Stick and optimized using Intel Distribution of OpenVINO Toolkit to allow brick and mortar stores to better compete with the online shopping experiences of today. Ceyhun talks about how the Intel AI Builders program has helped Vispera to better enable and benchmark their systems while helping connect them with potential prospective customers. To learn more, visit: vispera.co Visit Intel AI Builders at: builders.intel.com/ai

7 Aug 201912min

In this Intel on AI podcast episode: Justin Norman, Director of Data Science Research and Services at Cloudera, joins the Intel on AI podcast to talk about the challenges that enterprises face when executing AI solutions like scalability, compliance, and security. He illustrates how Cloudera’s experience and deep portfolio, including the Cloudera Data Science Workbench, provides the tools and expert guidance needed to help customer’s end-to-end transition into an AI first world. Justin highlights how Cloudera Fast Forward Labs is the foundation to provide strategic vision and expertise to facilitate customers in planning and business strategy when utilizing AI. Justin also describes how Cloudera and Intel has had a long strategic partnership working to provide customers both the cutting-edge hardware and software to meet all of their business needs. To learn more about the Cloudera Fast Forward Labs offering, the Cloudera Science Workbench tool, or machine learning, visit: cloudera.com/about/machine-learning.html Visit Intel AI Builders at: builders.intel.com/ai

7 Aug 201911min