Does Google really lead the world in artificial intelligence? The search giant leads in commercial search and its technical expertise in machine learning and deep learning is unmatched – but in the broader world of artificial intelligence that includes the smart home device space underpinned by NLP & speech recognition, self-driving technology space (Google’s facing stiff competition from all corners, especially NVIDIA), cloud-based technologies, next battleground for AI – Google is the no. 3 US cloud services provider and smartphone market, now on the cusp on mobile AI revolution, the company is still playing catch-up.
But when it comes to the enterprise AI market – Google has cracked the code for success, thanks to open sourcing a slew of ML tools in Google Cloud that could translate into big leads for the company. By offering a complete AI ecosystem, Google looks to corner a substantial market share. Besides, Google’s strong suit had always been data and immense computing power on which it built its enterprise AI market.
Google was never a hardware-focused enterprise — is now baking AI into the devices
Let’s face it – when it comes to building hardware, Google has met with limited success. The company was ostensibly late to market with social networking, the market for its AR glasses, Google Glass is still unproven, Android TV still lags behind in popularity to Apple TV (Chromecast is an exception), their robotics division (Boston Robotics sold off to SoftBank) has been sold off and in the smart home space, Amazon definitely is winning in device sales at the moment. Given how cloud computing services play a huge role in the smart home device market, Amazon is still in a great position to lead the market.Reports state that Amazon has sold more Echo smart speakers than Google Homes.
It’s most recent hardware success – Google’s AI driven phone Pixel phones which offers excellent photo-enhancement features and deeper hardware-software integration underpinned by AI-based technology isn’t pegged as a commercial success either. In fact, analysts suggest Pixel is toeing the line of iPhone when it comes the look and design. And given how smartphones can be the key to build mobile AI focused innovations, it’s hard to overlook the fact that iPhone still rules the market.
Even though the Pixel series is nipping at the heels of Samsung & Apple, a recent report points out that the company still needs a year to sell as many Pixels as Apple’s iPhone and gain a larger footprint. An IDC report states that Google Pixel sales doubled last year to 3.9 million, but still represents “a tiny portion of the 1.5 billion market size.” In the VR headset segment, Samsung Gear VR is still most widely accepted than Google Daydream VR headset.
However, the company is plowing billions of dollars in building consumer-facing AI-powered devices such as Pixel Buds, wireless AI powered headphones but it is still years away from making them mainstream.
But there’s one area where the search giant leads – building open ecosystems & profiting from it
It’s true open ecosystems lead to unprecedented innovation and Google has built its empire on open source – just the way Android created innovation and became a gamechanger in the smartphone space. According to NetMarketShare, Android is the dominant mobile platform with a 58.75 percent share of the worldwide mobile and tablet operating systems market. With mobile becoming the dominant global platform, Android’s global OS share was 37.93 percent compared to 37.91 percent for Windows, as per StatCounter report. Interestingly, Android tore down barriers in the mobile ecosystem and spurred developers to build a mobile software ecosystem without having to do heavy engineering. When it came to web browser, Google’s Chrome effectively killed Microsoft’s Internet Explorer and choked the competition, thereby profiting from collecting user data.
Google, a trendsetter in open sourcing AI technologies
According to a Capgemini State of AI report, one of the most defining characteristics that led to the growth of AI technologies is the open sourcing of key technologies by Google, Microsoft & Amazon. All the major tech companies are keen to have more developers on their platforms. It is here that Google proved to be a trendsetter by open sourcing its TensorFlow Platform in 2016 (Facebook then open-sourced Caffe, its flexible deep learning framework, and Amazon did the same with MXNET).
Most consumer-facing, traditional organizations that lacked the technology were willing to find real-world applications for their business challenges and found the open-source platforms an interesting avenue for investment.
Google clearly automated data science by enabling enterprises to build and train custom machine learning models using APIs to add machine learning capabilities to their applications, thereby implementing AI at scale.
1) First mover advantage with TensorFlow: As machine learning increased in importance, Google open sourced Tensorflow and Kubeflow to open up the machine learning ecosystem. It was a good attempt to help companies run like Google, in return they wish enterprises will turn to Google Cloud Platform to run more of their workloads.
2) In the same vein, Google has lowered the bar for machine learning adoption with Cloud AutoML: When it comes to the enterprise AI market, Google clearly has a well-laid out strategy and has cracked the code of winning over the developer community by equipping them with the tools and tech which leads to mainstream enterprise adoption.
According to Google Cloud CEO Diane Greene, it is only a matter of time before AI offerings become “more and more self-help”. For long, Greene’s pitch to accelerate Google Cloud adoption has been — you don’t need a Ph.D. in machine learning, but you can still build a highly accurate machine learning model, she said reportedly. A key advantage of CloudML is that enables developers, besides data scientists to customize a model without having in-depth ML knowledge and Google has made the interface more drag-and-drop.
Google Cloud offers APIs that provide machine learning capabilities to companies to build tools. Another key advantage over other providers AWS & Microsoft Azure is that companies can train their algorithms on enterprise data directly and Cloud AutoML ingests data assets and cranks out the model accordingly. Unlike other major providers, AWS which has a suite of other AI/ML services, and Microsoft’s Azure machine learning, Cloud AutoML demonstrated an ease of use.
This is also Google’s attempt to even out the cloud market and catch up with rivals AWS & Microsoft in the cloud computing market.
3) Last year, Google’s AutoML project built a computer vision system: Earlier last year in May, Google Brain researchers announced the creation of AutoML, an artificial intelligence (AI) that’s capable of generating its own AIs. In December 2017, the company announced a new computer vision system built by AutoML that outperformed man made state-of-the-art-models. The computer vision system can improve how autonomous vehicles and next-generation AI robots “see” better and can be leveraged by companies that have no experience to build a computer vision system, tailored to their needs. For eg. Radiologists can use CT scans to train a computer algorithm that identifies signs of lung cancer.
4) Google brings AI to healthcare: For several years, Google’s DeepMind has been making a play in the healthcare sector. Recently, London headquartered DeepMind announced it is taking the kidney failure prediction algorithms to the US Department of Veteran Affairs to combat AKI, acute kidney failure. Through the collaboration, DeepMind will have access to 700,000 plus medical records and will use its deep learning algorithms to predict AKI. Earlier last month, a Google Brain team released a research paper demonstrating the use of computer vision to detect heart disease, a major cause of death across the world.
Try deep learning using MATLAB