Introducing Cybercom Machinebook – An intelligent social IoT convergence


2016-10-17, 10:04 Posted by: Tero Keski-Valkama

Hellos from the Machinebook technical officer, Tero Keski-Valkama! I'd like to tell you about our new Machinebook concept.

By now, everyone has already seen the emergence of industrial internet of things in its various forms, and all kinds of platforms have been implemented for managing the flocks of devices and data streams. As the proverb goes, there are multiple methods to defur a feline, but there are at least a hundred ways to draw a pie diagram or a histogram of your distributed data sources by now.

If your business does not yet have a suitable IoT platform, we also offer help in selecting the best IoT solutions for your specific business, for more information please contact Tarmo.Pajunen@cybercom.com.

Machinebook is not a yet another IoT platform. It integrates with existing IoT platforms and adds a social dimension on top of them. Machinebook is a convergence between IoT, conversational paradigm and machine learning. In Industrie 4.0 terminology, Machinebook is a platform for Cyber-Physical Social Systems (CPSS).

Cybercom Machinebook is a convergence of IoT, conversational paradigm and machine learning.

As machine learning technologies are being developed now in an increasing pace, humans and machines are coming together in new kinds of conversational user interfaces. In the consumer services side, these are realized as customer service bots and personal assistants.

The inherent value in this new conversational paradigm of interaction is the accumulating context. During a long interaction, spanning possibly even the whole customer relationship, the conversation accumulates new facts and information. The meaning of the new interaction depends on the whole conversation up to that point. The capabilities of the automatic agent can also be incrementally added without requiring new user interface design or without the need for the user to learn new functions. The conversational interface is inherently extensible and self-describing. It can also be trivially projected to different terminals from laptops to mobile devices, from kiosks to being embedded to web sites, and also supporting speech recognition interfaces.

To support the implementation of all these new conversational agents for different types of businesses, big AI companies have started offering their conversational AI APIs for general use. These include all kinds of things from tools to build modern chatbots, to services extracting meaning from conversational data and sentences.

Machinebook is a way to adapt these new capabilities into the industrial environment. Indifferent machines are transformed with virtual avatars into social agents that function as helpful digital colleagues, or assistants to the human workers.

Machinebook architecture

In manufacturing and process industry, there is a constant pressure to increase the share of automation. The most easily automated tasks have already been automated. The old way of thinking automation has been about focusing on replacing human workers. However, there is another way to think about automation, and that is about focusing on collaborative automation that improves the human worker efficiency.

Increasing the share of automation in a traditional way has become non-trivial, because a complex robot requires complex planning to make it work effectively and with minimal supervision. The easiest way to increase the share of automation at this point requires collaborative way of working between humans and machines, leaving humans to do things they excel in, but allowing machines to fulfill their greatest potential in tasks that still require some human interaction, such as maintenance, training and flexible adaptation to new products or processes. Machinebook allows more collaboration in such tasks by bringing in the social, conversational layer on top of and on the borders of the indifferent M2M processes.

An indifferent, asocial machine

The social environment is a new way of looking at what has happened during the week in the factory. Instead of pie diagrams, we get stories of what has happened. This increases the visibility to the real day-to day work, allowing more informed management. It functions also as a good way to spread tacit knowledge about the processes and incidents, and about best practices and ways of working.

Machinebook UI

Machinebook avatars are independent agents who care about the job at hand, and can perform all kinds of assistance, recommending actions such as maintenance or after market sales, and even training and motivation. The avatar does not only represent the machine itself, but it also represents the device provider and the management of the site. Support calls might transform into simple discussions with the avatar. For incidents, the service guide information will be readily available enriched with information from similar previous incidents.

Implementing intelligent Industrie 4.0 factories is challenging in the fast changing world. A smart device might have half of its intelligent features obsolete by the time they are completed and deployed. Progressive deployment of intelligent capabilities makes business sense. Instead of large monoliths, it is smarter to implement smaller modules when needed.

To allow progressive deployment of intelligence to the factory floor, Machinebook includes a Marketplace for deploying external components to the platform. This is to allow device and solution providers to implement their own avatars and intelligent modules onto the platform.

Industry has come a long way, adding incremental intelligence is a way forward

Lastly, after showing the timeliness, feasibility and business value of our Machinebook concept, I'll tell you a quick summary of what it looks like inside. In practice, Machinebook is a sophisticated conversational bus and data store, a collection of modular Machinebook components deployed from the Machinebook Marketplace, integrations to IoT and AI APIs, and the social frontend with forum-like features designed specifically for the industrial context and collaboration between machine avatars and human workers. We are now authoring the Machinebook white paper 2.0 to define the technical composition of the platform in a more specific way.

Our Machinebook concept is currently in the process of finding a pilot partner with whom we will implement the first revision of the platform in a customer-focused fashion. Our focus is in Finland and Sweden at the moment. If you are interested in becoming our partner in building the future, please don't hesitate to contact our Machinebook marketing officer: Tarmo.Pajunen@cybercom.com.


comments powered by Disqus