This is what enables us to identify a user’s intent, allowing that user to utilize services in his ‘own language’. But also fine-grained personalization applications aided by face and image recognition are possibilities that we explore.
Cognitive services advance at a very quick pace. The amount of new services grows every day. Simultaneously, we notice that the quality of these services increases at a similar if not faster pace. Considering the rapid progress of cognitive services’ development, Mirabeau has been testing and comparing the most-used services of a number of prominent cloud providers. This way, we have been able to establish in which situations, which services offer the best solution. Straight to the results? All the details on our research can be downloaded at the bottom of this page.
We spot a great number of parallels between the arrival of cognitive services and the transition of desktop to smartphone. This period of mobile first has led to an enormous acceleration of the role digital plays in our day-to-day lives. The impact it’s had, has no way of escaping us. Mobile services, such as WhatsApp, mobile banking and Uber have become impossible to exclude, ignore or expel from our lives. Entirely new companies with innovative business models have arisen due to the unique possibilities of the smartphone. We stand at the dawn of a new revolution; in which we expect the same acceleration through the availability of AI (Artificial Intelligence) in the form of cognitive services.
The voice interfaces are a fine example of this. Apple’s Siri processes 2 billion voice commands every week and 20% of all mobile Google searches are done through voice search. Amazon expects to sell 10 million reproductions of their smart speaker ‘Echo’ in 2017, and they have already noticed that those who own an Echo speaker, are inclined to purchase more on Amazon on average.
Chatbots too, lean heavily on the foundation of artificial intelligence. Understanding the true intent of users –including recognizing emotion – is the greatest challenge yet in developing properly functioning bot applications. Making bots human. A challenge that, without AI, simply cannot be overcome.
Another example constitutes the area of image recognition. Which is not just useful in ensuring our safety but also of value to authentication services. For example, Apple currently deploys this technology for their login process on the iPhoneX, as a substitute for their Touch ID application. Nowadays, this is the technology that makes these applications possible.
That AI is currently experiencing the incredible developmental growth that it is – and genuinely becoming an inexpendable part of our daily lives – has everything to do with the fact that the algorithms that are in use (like language or emotion recognition) have considerably improved, and that there’s more and more data available for models to train with and improve on. But, perhaps more importantly, the cost of the required computer processing power nowadays has significantly decreased due to the rise of cloud computing. Suppliers such as Amazon Web Services, IBM, Google or Microsoft offer ready-made building blocks for relatively minute fees. Which are easily applicable for software developers without having to invest beforehand.
Cognitive services are able to comprehend data in a way that’s comparable with how a human would understand data. An application for image recognition for example, will ‘see’ the difference between a cat or a dog in the same way that a chatbot ‘understands’ what a consumer types, without a human intermediary. The development of such smart models is reasonably complex and time consuming. Cloud providers offer AI as a service. And because these services develop continuously, you instantly profit from improvements and new functionalities. Your application gradually becomes more intelligent, without requiring adjustments.
A few examples of inspiring applications that have been put into practice for the aforementioned services:
• The Transavia chatbot is being deployed to discover destinations and book plane tickets through Facebook Messenger. Cognitive language services ensure that the chatbot understands language in the right context. This way the bot understands the manner in which users ‘talk’ while at the same time relevant feedback is being given. For more information, click here.
• Our face recognition demo recognizes employees in real-time and displays all sorts of information, such as age, name and function but also facial features like whether there’s makeup applied or not. More information.
• Emotion recognition. Cognitive cloud services are capable of recognizing emotion, so that they can, for example, set the tone for a piece of text. This way you can determine whether a mood is positive or negative. A possibility we’ve supplemented with the detection of emoticon’s emotions. Emoticons have become an important part of chat conversations, because of which it’s relevant to consider them in the process of ascertaining which emoticons users feel are fitting substitutes for their conversations.
• Augmented Reality. Cognitive services are a welcome addition to the development of AR applications – for the HoloLens for example. Image recognition services aid us in the recognition of objects in the real world, to then project onto them relevant information through AR.
The amount of Cloud Cognitive Services on offer, grows rapidly – Microsoft for example, is already working on the broadening and deepening of its portfolio. But other providers too, are offering competitive services. Interested into a summary of relevant cognitive services, how we deploy these services and how the performance of these services of various different providers compare?
Cognitive Services Findings
Discover all possibilities cognitive services put on offer. Plan a workshop with us or send an e-mail to Edgard Beckand, our Strategic Director.