As a motivating example, imagine that every morning when leaving your home you make sure that the lights are turned off as well as other appliances and smart home technologies. Now imagine that you don’t have to repeat all of that every day. Instead, the electronic devices and tiny embedded systems take care of it by themselves. By being enabled to learn from your everyday habits, digital things can discover what should be done when you leave the home.
– We want to enable independently designed and deployed devices and systems to automatically interoperate and learn from interactions with humans and the environment, Blerim Emruli explains.
Create an associative memory
– One way to achieve this is by making use of an associative memory. In contrast to conventional memories used in most digital devices, an associative memory can recover patterns when supplied with a noisy or incomplete pattern as a cue. In addition, an associative memory can store sequences of patterns, which enables new interesting technical applications.
In his doctoral thesis Ubiquitous Cognitive Computing: A Vector Symbolic Approach, Blerim Emruli has addressed this problem by employing cognitive computing principles that offer a new way to represent information and design interoperable and adaptive devices and systems. Inspired by biology and cognitive functions Blerim has developed a mathematical model that enables the development of digital devices and especially miniature embedded systems with cognitive qualities.
Internet requires new solutions
Small and cheap electronic devices are increasingly being integrated with the Internet, creating the so-called Internet of Things. The prediction is that by 2020 there will 50 billion connected devices. Enabling the development of adaptive electronic systems with cognitive qualities is important in that perspective. Today many of the existing electronic systems are non-adaptive and depend to a large extent on manual system integration, configuration and maintenance.
– These devices need to be easily deployed and integrated, otherwise the resulting systems will be too costly to configure and maintain, says Blerim Emruli.
– The main purpose of my thesis work is to investigate brain-inspired approaches to computing that could potentially be useful in addressing these challenges by enabling the Internet of Things to learn, for example from interactions with humans and the environment.