Currently, the future of smart homes relies on connected devices that communicate with each other, in order to create customizable user experiences for added efficiency and comfort in our lives. Now a new breed of camera-enabled devices within the connected home is adding the important social element into the mix.
Social robots and companion bots alike are on a remarkable rise; and when coupled with intelligent vision-based face analytics and emotion recognition technology, the future of the connected home is sure to be one that we have long been waiting for.
At the RE•WORK Connected Home Summit
, Modar Alaoui, CEO of Eyeris
, will reveal how the integration of deep learning into the next generation of camera-enabled social and companion robots will change the connected home landscape forever. I asked him a few questions to learn more.
How will intelligent vision-based face analytics change the way we interact with our homes?
Most of today’s smart devices still require action / command-based user interactions. The future of connected home devices relies for the most part on their ability to anticipate users’ needs and respond accordingly through invisible adaptive interfaces for seamless interaction. When camera-enabled devices at home include intelligent face analytics algorithms, these interactions take AI to a whole new level: a human-like type of interaction.
In fact, it is widely consensual that 93 percent of human communication is non-verbal, most of which is through facial expressions followed by gesture and posture. At the same token, humans also process most – if not all – of the communicated information emotionally by default.
These forms of communication should translate well into today’s consumer Human-Robot Interaction. When integrated into Social Robotics, face analytics including facial expression recognition deem critical elements that enable the promising generation of intelligent companion machines; ones that can recognize the users and their emotions to warrantee an all-around level of understanding and human-like social connections through Ambient Intelligence (AmI). Intelligent vision-based face analytics hold the same promise for other camera-enabled devices within our homes; from face recognition-enabled smart doorbells, to emotionally-intelligent smart avatars and virtual assistants to gaming consoles and other appliances.
How can connected devices and deep learning combine to create smarter homes?
To create the next generation of smarter homes, the role of connected devices should not be limited to one that connects to an Internet network and exchange processed data with other compatible connected devices within predetermined protocols, but also should be one that leverages the many advantages of Deep Learning. With the recent advancements in processing power at increasingly reduced costs, incrementally faster connectivity speeds and the abundance of generated data that can be collected for training, the future of smart homes that leverage artificially intelligent algorithms and Deep Learning Neural Networks can only get brighter with virtually unlimited sophistication.
In fact, home connected devices that learn continuously from their immediate environment – and most importantly register and track users’ preferences along with their behavior over time – contribute critically as the building blocks towards creating the world of smarter homes that we have long been waiting for. Furthermore, when connected devices are enabled with Over The Air (OTA) architecture capabilities, the quest for technology self-improvements towards more and more customizable user experiences will be all within reach.
What are the main challenges to creating next generation smart homes currently?
The building blocks of the next generation smart homes require delivery on the most fundamental challenges it faces today. Inadequate security and privacy protections for user data and lack of standardized design protocols seem to be the biggest hurtles that face the rapid evolution of the connected home. There is currently no consensus on how to implement security on domestic IoT, and while it is unfortunate that most device providers treat security with a lower priority in their design process to prevent misuse and safeguard users personal data, there is no silver bullet that can mitigate the inherent threats that will potentially arise.
On the other hand, while these are important challenges, they are predicated on the notion that the smart home components – which may have been acquired at different times, from different vendors, and which were created under different design constraints and considerations – are able to interoperate at all. In fact, impromptu interoperability, defined as the ability to interconnect with little or no advance planning or implementation, is implicit in much of the current literature of ubiquitous computing in the home. With fluid interoperability, individual technologies have the potential to create a fabric of complementary functionality. Without improvised interoperability, the smart home of the future is likely to be characterized by islands of functionality, as the sets of devices that were explicitly built to recognize each other can interoperate, but other sets of devices simply cannot, due to software upgrades, version mismatches, driver installations challenges and other related matters. The chief obstacle here is that, in general, every device or software service that targets the smart home must be explicitly written to understand every other type of device or software that it may encounter within the home. Without this a priori agreement on both syntax and semantics, strong security protocols, connected home device interoperability remains one of the chief obstacles to the future of the connected homes.
What advancements excite you most in the field?
An altruistic smart home with devices “doing it on their own” through AI-enabled sensors and invisible interfaces; a smart home that understands user behavior on human-like and emotional levels; a smart home that offers new levels of social interaction, ambiance, comfort, entertainment, safety and energy efficiency; are all area that are certainly thrilling. But what excites me the most is the broader context where platforms will also connect the home to various other locations as other types of actors in the ecosystem such as cars, online stores, offices, schools, etc. What excites me the most is the potential for the enablement of a technologically smarter society as a whole.
Modar Alaoui will be speaking at the RE•WORK Connected Home Summit in Boston on 12-13 May 2016. Other speakers include Blade Kotelly, VP of Design at Jibo; David Isbitski, Chief Evangelist of Alexa and Echo and Amazon; Rahul Bhattacharyya, Research Scientist at MIT; Chris Jones, VP of Technology, iRobot; and Ted Booth, Senior Director of User Experience, Honeywell. The Connected Home Summit is taking place alongside the Deep Learning Summit, with speakers from Facebook, MIT, Salesforce, Mitsubishi, Université de Montréal, Affectiva and more. For more information and to register please visit the event website here.
Deep Learning Algorithms
Connected Home Summit