AI Biweekly: 10 Bits from Sept (Pt 2)

Sept. 18 – GE Using AI to Optimize Power Grids

General Electric is using AI technology to optimize electricity grids. The technology increases efficiency and apparently can help the industry save up to US$200 billion. The company says it is also applying machine learning capabilities to airplanes, trains, and medical instruments. It has acquired multiple AI startups befitting similar functionalities. meanwhile Siemens, its long-time German industrial rival, has recently signed several contracts in North America and Asia to build local AI hubs.

Sept. 18 – Intel Pumps US$1 Billion Into AI Startups

Intel announces a US$1 billion investment in AI. The company is optimistic about the future of AI, and will accelerate in-house research on neuromorphic computing and other new learning paradigms. To date its venture capital unit has invested US$1 billion into startup companies, including AI-powered analytics companies like Seattle-based Mighty AI, Boston-based Data Robot, and San Mateo’s Lumiata.

Sept. 19 – Salesforce’s Venture Unit Puts US$50 Million into AI Fund

The Sales AI Fund has US$50 million dedicated to AI, which the company claims is making smaller, secondary investments into AI startups. Salesforce Ventures Vice President Matt Garratt says US$50 million may not be an impressive number for a VC fund, but shows the company’s preliminary commitment to AI. On top of this, Salesforce’s own AI toolset — Salesforce Einstein — officially turned one year old this month. Over 7,000 developers are building their AI applications on the platform.

Sept. 20 – IBM Targets India’s 2.7 Billion Connected Devices for IoT Market

IBM is stepping up its game in India, placing emphasis on the IoT platform augmented with AI capabilities, targeting India’s 2.7 billion connected devices. In the past year the company reported a 15% growth in market infiltration and established local partnerships with companies such as Tata Steel, Reliance Group, Tech Mahindra, KPIT, and Schneider Electric India. IBM has also made deals with the Indian government on water and agriculture projects. So far 1,000 firms have signed up for the Watson IoT platform.

Sept. 21 – Baidu Sets Up a US$1.5 Billion Self-Driving Car Fund

Baidu’s Apollo open-source self-driving platform is one of the leading industry players in China. The company recently announced a CN¥10 billion (US$1.5 billion) fund for the self-driving car industry, aiming to invest in 100 projects over the next three years. One of the most important investments — in conjunction with Ford — is in LiDAR manufacturing company Velodyne, for US$150 million. To date Baidu has established over 70 global partnerships.

Sept. 21 – Google Cloud Using NVIDIA GPUs to Boost AI Capabilities

Google Cloud Platform (GCP) is using beta version NVIDIA’s P100 GPUs to accelerate AI functionalities like machine learning, training and inference. The P100 GPUs can accelerate workloads up to 10x faster compared to K80 GPUs. The latter are now widely available on Google Compute Engine. A Google statement says “Cloud GPUs provide an unparalleled combination of flexibilities, performance and cost-savings.”

Sept. 25 – NVIDIA GTC China Announces Partnerships with Alibaba and Huawei for AI Smart Cities

NVIDIA’s AI smart city platform Metropolis is now partnering up with China’s biggest e-commerce platform Alibaba and mobile manufacturer Huawei. The NVIDIA Metropolis program hopes to install one billion video cameras worldwide by 2020. The announcement was made at the annual NVIDIA GTA Conference in Beijing on September 26th, where the company also introduced deep learning engine TensorRT 3.0, the HGX-1 hyperscale GPU accelerator powered by Tesla V100 for AI cloud computing, and the all-new AI processor Xavier for self-driving cars.

Sept. 26 – JD.com Partners with NVIDIA for Drone Deliveries

NVIDIA is working with e-commerce giant JD.com on drones. The JDrone and JDrover are aerial and ground-based robots used for last-mile delivery and rescue purposes. Both will install NVIDIA’S Jetsom platform with AI navigation capabilities. JD says it will start testing the drones next year across university campuses. The drone can reportedly carry over 30 kg of cargo, ranging from food and cosmetic packages to rescue parcels. The company hopes to put one million of these robots to work in the next half-decade.

Sept. 27 – BMW Is Integrating Alexa Into Its Cars in 2018

Amazon’s voice assistant Alexa will be accessible in BMW cars by 2018. A new microphone system will also be installed inside the vehicle. This adds a comprehensive set of internet capabilities to the car. Drivers can even use third-party apps like Starbucks via voice command. Auto-manufacturer Ford has already added Alexa to its systems.

Sept. 29 – Intel’s Neuromorphic Chip Loihi Challenges Traditional GPUs and CPUs

Intel announces a new, experimental neuromorphic chip called Loihi. This new type of chip physically mimics the brain’s architecture, in contrast to conventional silicon-based GPUs and CPUs. The success of these types of chips can increase efficiency and allow AI functionalities to run smoothly in devices like smartphones, cars, and robots. However commercialization of such chips is still far away, and they wouldn’t work with some deep learning models.

---