The cognitive Internet of Things: an interview with Raffaele Giaffreda

RE.WORK speaker Raffaele Giaffreda took some time out of his busy schedule to speak to Web Magazin about the cognitive internet of things and how this will impact our day to day lives in the future.

Raffaele is a research scientist whose work focuses on enabling a cognitive internet of things. He is head of the Smart IoT research group at CREATE-NET and also holds the position of Project Coordinator of EU FP7 iCore. He previously worked for Telecom Italia and British Telecom and has a leadership role in the innovation sector of the European Research Cluster on the Internet of Things.

Raffaele is interested in exploring the growing needs to easily set up the digital world of connected “things”. His talk at the RE.WORK will give an overview of the reasons behind using cognitive technologies for the management of connected objects and explore the quality of life that it will offer society in the future.

WM: Tell us a little bit about CREATE-NET and smaRt IoT (RIoT)

RG: CREATE-NET is an international non-profit research organisation located in Trento, Italy. It brings together some of the best research groups in all areas of networking and communications and is highly devoted to promoting technological transfer, with a number of spin-off companies that have been created exploiting technologies developed and incubated in EU collaborative projects. SmaRt IoT (RIoT) is one of CREATE-NET research areas where we focus mostly on cognitive IoT and associated applications as well as on how cloud computing and software defined networking can help achieve more robust and reliable IoT services. 

One of our main engagements in the past three years has been leading and working on an EU FP7 collaborative project called iCore, which brings together many EU industrial institutions, universities and research centres on the subject of how to empower IoT with cognitive technologies.

WM: At the most basic level how will cognitive computing and the internet of things assist day to day life?

RG: At the most basic level cognitive computing will be improving the plethora of IoT-based solutions already available to help us in so called non-value add activities, which today take some of our time and attention while we would rather concentrate on doing something else or which should be done while we are physically elsewhere.

So turning to automated systems for feeding our plants, pets, turning off lights in the house, ensuring the right HVAC levels, controlling appliances from remote etc. will become more and more common thanks to IoT.Besides time saving for people, more widespread and detailed IoT sensing will also foster an increasingly efficient usage of our resources (from roads to energy and utilities in general).

Cognitive computing will add to the adaptability of these solutions attracting more users than just the current early adopters. Unprecedented sensing capabilities together with easily configurable machine learning algorithms can greatly help achieve simple and adaptable automated systems.

WM: At an advanced level how do you predict cognitive computing will change our daily routine in years to come?

RG: While it is relatively easy to set environmental actuators to fit user-defined rules, the prediction / understanding of users intentions and needs is a considerably more complex problem but has greater potential for returns in terms of perceived usefulness and therefore in terms of impact to our future daily routines. Recent experiments like IBM Watson or chatterbot Eugene or more and more popular smartphone assistants (i.e. Siri) show us the advances made in this domain.Today we still spend considerable time “looking-up” the info we need, “finding” the app we would like to use and customise to our needs.

Chances are that with the commercialisations of products derived from the above mentioned experiments will put us in a more comfortable situation talking to our devices as if we were talking to a human assistant. This more natural way of interfacing to machines and apps will not only enlarge the user base beyond “geek world” but most importantly it will increase the wealth of data (i.e. such as live voice user feedback on running apps) that applications will be able to harvest.

Increasing amounts of feedback data means better training for machine learning algorithms leading to improved models of the real world, enabling applications to make better inferences and cause-effect predictions. With positive implications for health monitoring, disaster prevention, timely maintenance of machinery, traffic control etc.

WM: What advice would you give to computer engineers hoping to find work in the Internet of Things?

RG: We are living the world of apps and an unprecedented availability of miniaturised hardware for sensing, communicating, actuating. Given this contextual background, to keep their CV appealing, computer engineers must ensure they have something to contribute to the so called “makers” community. In other words, taking a deep cut across the entire OSI stack, it is highly desirable to have familiarity in one or more of the following: building apps for iOS and Android, turning real world objects into connected objects through the use of various sensors, open hardware projects and miniaturised computing platforms (i.e. from Arduino to Raspberry Pi), low power wireless sensor networks and communications. All these are ingredients that contribute to the IoT great potential.

Oh, being so enthusiastic about the opportunities ahead, I nearly forgot about what most technologists and IoT engineers are sweeping under the carpet or forgetting too, that is dealing with potential security threats that such more ubiquitous monitoring exposes users to. If you have expertise in the security domain, on how to foresee, mitigate and prevent the risk of having more and more of our private connected objects tampered with, then your future working life is rosy.

WM: IBM made five predictions for cognitive computing in 2014, do you agree with them?

RG: Overall I tend to agree with each of the predictions. Actually many of the points I made above do resonate well with those predictions. Technology to achieve those challenges is there or if it isn’t, it is not too far off. What I am skeptical about is how popular those services will be within the given 5 years timeframe since some of them need cultural changes and better trust in technology.

Many people are not yet impressed with the features of Siri or Watson and fail to appreciate the achievements of Eugene. The perception of what they gain from cognitive applications is counterbalanced by the perception of what they lose (time for personalising apps, control of data ownership, privacy, security in general). The latter has to be considerably reduced and overwhelmed by the perceived advantages of trusting cognitive technologies. Legacy and need for training also make me think that we might need more than 5 years before seeing those predictions widely fulfilled.

WM: We keep hearing about the vulnerabilities of IoT devices. Can you comment on security measures being taken to protect data?

RG: There are two security aspects hidden in this question. One relates indeed to the vulnerabilities of IoT devices which not only threaten to expose our private data but can cause more serious damage within the network to which these devices are connected. These vulnerabilities often relate to unencrypted wireless communications or unprotected usernames and passwords for gaining control of these devices. Ensuring users and engineers get to know about these is paramount. They should be relatively easy to uncover and fix with the help of security experts and/or with the replacement of non reconfigurable devices.

Protecting data and therefore the data owner’s right to keep it private is the other aspect and indeed a big issue we face with IoT. The EU has a law for data protection of personal information since 1995 (Directive 95/46/EC). A comprehensive reform to strengthen individual rights and tackle the challenges of globalisation and new technologies was proposed in 2012 and is likely to become law by 2015.

On the technology front there are means to ensure privacy and non-disclosure „the hard way”. The downside of enforcing security is that there is a cost associated with it, and that is often reduced user-friendliness of resulting applications (a bit like websites that require us to change periodically our passwords). To give an example in iCore we have adopted a solution based on sticky policies where basically IoT sensed data can be protected with policies for access which must be satisfied before data can be disclosed. A comprehensive report on security has also been recently produced by the European Cluster of IoT projects (IERC). So while we are at the stage of „enthusiastic tinkering” with IoT there is also consciousness that security measures must be in place to ensure wide adoption.

The RE.WORK Technology Conference is taking place in London on the 18th-19th of September. The conference will host a wide range of speakers, presenting on emerging technologies which will provide an unprecedented era of opportunity to create a more sustainable and equal society. Topics discussed at the conference will include the Internet of things, 3D Printing, Nanotech, Artificial Intelligence, Computing Systems, Sensors and Renewable Energy.


Feature image: The man thinking about work, family, house, mails, money; car; football and other things via Shutterstock. Copyright: wildfloweret

Unsere Redaktion empfiehlt:

Relevante Beiträge

Meinungen zu diesem Beitrag

- Gib Deinen Standort ein -
- or -