Cloud-based Artificial Intelligence for Everyone

Introduction

Artificial Intelligence or also known as AI was invented by John McCarthy in 1979, which is also can be defined as “the science and engineering of making intelligent machines”. (Linthicum, November/December 2017).  Whereby, cloud computing is defined as delivering hosted services over the internet either in publicly or privately. In 1996, the phrase “cloud computing” appeared as the first known that was mentioned in a Compaq internal document (Regalado, 2011). By combining and integrating these two powerful technologies, it gives people more access to apps and interfaces that use AI functionality through public cloud providers. Cognitive response required can be easily obtained by allowing user’s applications or things to be uploaded either in structured or unstructured data format. As Artificial Intelligence is currently used by mostly dominated giant companies but having it as a cloud-based service, could be extensively available and moreover, giving the economy a boost. (Snow, 2018)

Technology Overview

With this technology, almost everyone can drive their very own AI application as these systems are affordable to operate. This would be feasibly the largest advantage of AI in the cloud and is really bringing back AI as a core enabling technology. Moreover, public clouds can be affordable and cost savings for data storage. Users also can control the storage systems or true databases as the input of the data into the AI enabled applications. In addition to that, AI functionality can be driven directly within the application using software developer kits (SDK) and application programing interface (API) as the AI systems offered by public cloud providers. Most programming languages would also be normally supported by the systems. Within this applications, real value of AI technologies is fully utilized. Given an example, fraudulent cases for loan application can be immediately determine and all the process to deal with the issues can be provided, so applicant can fix any errors and resubmit the application. From business user perspective, with this technology, clients get to discover new insights about the customers or any other information needed. They can make smarter decisions, recommend the next sales, marketing and service action by predicting the outcomes. This technology also can be used to improve sales, automating tasks, improve service and marketing and IT applications. Some examples of products for this technology would be Video Intelligence, Internet-of-Things Platform, Machine Learning, Data Refinery and many more.

Technology Stage

As new capabilities are emerging, and the market is changing rapidly, the AI trends need to be followed closely. AI early adopters among start-ups and tech giants are experimenting with these technology capabilities to challenge their rivals (Loucks, 2018). Most products from this technology have already been developed and widely used across the globe. Therefore, Cloud-based AI overall is at scale 9 for Technical Readiness Level (TRL) and Product Readiness Level because the system is already proven in operational environment and it is running in production. The readiness level is partly based on research and practical assessment across five essential areas which includes the strategy, people, data, infrastructure and ethics. As for per products example, giant company like Google have designed its own Artificial Intelligence specific chips to drive machine learning in the Internet of Things devices and data centres (Loucks, 2018) (Metz, 2016). Ever since the launch of Google Brain in 2011, they have been experimenting and analysing the technology of deep learning and uses it extensively from every possible aspects and possible capabilities like performing video analytics to cooling data centres (Loucks, 2018) (Marr, 2017). Amazon has used machine learning to drive recommendations for many years and using this technology to reform business processes and develop new product categories such as the very well-known virtual assistant nowadays, the Alexa (Loucks, 2018) (Levy, 2018).  SAP integrated AI into its cloud-based ERP system, to support sales, finance, procurement, and the supply chain throughout business processes (Loucks, 2018) (Ali, 2018). Meanwhile, Alibaba venturing heavily in AI while expanding into areas like autonomous vehicles, virtual assistants and chip design (Loucks, 2018).

Possible products

Any products or moreover new innovations will take take time to spread or diffuse and adopt into markets. After taking a few assessments through these three approaches; diffusion and adoption, exploration and learning, and triangulation for insights, there are four possible products that have been narrowed down with its own description and advantages accordingly.

The first suggested product is the Video Intelligence API. Using this API, content and information in videos can be easily searchable and obtained. By extracting the metadata of the video content, users can easily organize and gather insights to make it easy to organize and analyze without require any machine learning expertise.

Secondly is the Internet-of-Things API. This API is an IoT interface that can be used to connect IoT devices through networks and gateways. It is a cloud-hosted service that have functions to store data, do rapid visualization, device registration, connectivity and control. Getting real-time and edge predictions and analysis of users, environmental data and machines are some of the advantages of this API without having the use of machine learning or any cognitive APIs.

Next is Machine Learning Modelling is a tool where data scientist and developer get to construct and deploy machine learning models rapidly. There will be machine learning models that can be train and use for companies/business needs specifically without needing machine learning expertise to build it from scratch.

Lastly, would be the Data Refinery Catalog, whereby this product can be used by the data users to find, analyze, categorized information, curate and share the data with others.

Product Attributes

The challenge of assessing future markets for new technologies is to determine the demand for products that don’t exist from customers who don’t yet know about them and at the same time, the trajectory of technology development and speed of market acceptance are also uncertain (Day, 2000). The followings are the products attributes that have been recognized for potential product development and commercialization in each market segments respectively.

Video Intelligence API would be a very useful tool amongst students, researchers, young professional for searching video catalogues as the same way you search text documents. Whereas from the business sector view, the departments from the marketing, administration and their IT specialist can use this product for extract actionable insights from video files without requiring any machine learning or computer vision knowledge to gather insights for markets research and predictions. Apart from that, security and defense department in government sector can also utilize this product to automatically analyses video content to identify, get information on the entities in the video content with precision video analysis technology.

With Internet of Things Cloud API, IT specialist, predictive maintenance and real time asset tracking in the business sector can use this product to accelerate business agility and decision making with IoT data. Young professional and researcher from private sector can also use this product to run IoT solutions with machine learning capabilities both locally and in the cloud. They can also use this product to extend AI capabilities to deliver deep insights faster from locally generated data.

For Machine Learning API, the market segment would be mainly focusing in the education, research and business sector area where this product would be the tool to provides every developer and researchers the ability to use and embed machine learning models with minimum machine learning knowledge. Other than that, this API can also generate high-quality training data for business sector for their IT specialist, marketing, administration departments. This product can also be a platform to train custom machine learning models that are specific to the business needs, with minimum effort and machine learning expertise and providing customers with a consistent method of access across the entire Cloud service line, store training data in cloud storage and generate a prediction on the trained model to produce high-quality model enables developers with less machine learning knowledge to have high-quality machine learning models specifically to their business needs.

Researchers, young professional and any data users can use this product; Data Refinery APIas a tool to quickly gather and analyze data for further research. IT experts like data engineers, data stewards or scientists and business analysts in business sector have the capability to shop for data easily.

PAMM

Product Attribute Market Matrix helps to identify the key marketing atributes for a successful product. Based from the analysis, the business sector would be the highest potential user for most of the products in this cloud-based AI technology. Products like the Internet of Things platform and Data Refinery API can be predicted to be one of the main factors that influence the boost of the company’s economy and industries. Video Intelligence API would be likely useful for the security and defense departments in the government and also among researchers for gathering insights.

References

• Ali, F., 2018. SAP sticking to its 2025 deadline, s.l.: Enterprise Resource Planner.
• Day, G. S., 2000. Wharton on Managing Emerging Technologies. First ed. New Jersey: John Wiley and Sons.
• Levy, S., 2018. Inside Amazon’s artificial intelligence flywheel, s.l.: Wired.
• Linthicum, D. S., November/December 2017. Making Sense of AI in Public Clouds. IEEE Cloud Computing, 4(6), pp. 70-72.
• Loucks, J., 2018. Artificial intelligence – From expert-only to everywhere. [Online]
  Available at: https://www.deloitte.co.uk/tmtpredictions/predictions/artificial-intelligence/#top
  [Accessed 10 3 2019].
• Marr, B., 2017. The amazing ways Google uses deep learning AI, s.l.: Forbes.
• Metz, C., 2016. Google Built Its Very Own Chips to Power Its AI Bots. [Online]
  Available at: https://www.wired.com/2016/05/google-tpu-custom-chips/
  [Accessed 10 5 2019].
• Regalado, A., 2011. Who Coined ‘Cloud Computing’?. [Online]
  Available at: https://www.technologyreview.com/s/425970/who-coined-cloud-computing/
  [Accessed 31 July 2013].
• Snow, J., 2018. TOP BREAKTHROUGH TECHNOLOGIES FOR 2018 : AI FOR EVERYBODY. [Online]
  Available at: https://citi.io/2018/03/08/top-breakthrough-technologies-for-2018-ai-for-everybody/
  [Accessed 15 February 2019].

Appendix

Video Intelligence API

Product: Internet of Things Service

Product: Machine Learning API

Product: Data Refinery API