If a baby was born in 2016, she would be around two years old now and would have just about enough balance to jump, climb stairs with help, and feed herself with basic accuracy – with of course a bit of chaos thrown in. For data, on the other hand, two years is a very long time. According to IBM, 90 percent of data online has been created since 2016. That is because in the past couple of years, the number of internet connections and connected devices has proliferated and generated data at a phenomenal rate. At two years old, data has reached adolescence! No wonder many see data as the world’s most valuable resource – not oil.
Data is the new oil for 2018
Technology behemoths such as IBM, Microsoft, Google and Amazon were quick to see the potential of data. Over 2016 and ’17, these companies invested aggressively in technologies such as Cognitive Systems – computing that replicates the human thought process by integrating data mining and self-learning. Cognitive Systems tries to mimic human capabilities that help us to navigate our day to day life by interacting with objects and environments, making decisions based on past experiences and gaining knowledge.
Extracting useful and actionable information from structured and unstructured data is a major driver for Cognitive Systems. Additionally, predictive analytics in B2C markets is another key driver. Researchers have been optimizing algorithms and reducing error rates much faster than ever before. Cognitive Systems made great strides in 2017 and according to analysts at IDC, spending reached a record $12.5 billion in 2017. As R&D reached record levels, Cognitive Systems and AI started moving out of labs and the experimental stage to live deployments.
These are three highpoints from 2017 for Cognitive Systems; but they also reveal the limitations:
- Autonomous vehicles: The likes of Tesla and Waymo had live street trials and demonstrated how their autonomous capabilities had advanced. They are however not yet at “Level 5” – ready to replace human drivers.
- Digital Assistants: In 2017 Google Home was launched in the UK and Amazon’s Echo was sold in the millions. Both devices have skills added almost each day – it yet cannot hold a meaningful conversation with a user.
- Speech recognition: In August 2017, Microsoft’s speech recognition hit a remarkable 5.1 percent error rate – that’s on par with human transcribers. The company acknowledged that the next step is to actually get their computers to understand the meaning and intent of words – not just recognize it.
Data, data everywhere
All these achievements provide a glimpse of what’s on the horizon for 2018 and beyond. The Deep Learning field has experienced rapid advances that will help data scientists this year to analyse large data sets at an unprecedented scale as never before. Yet there are challenges businesses will need to overcome if they are to effectively utilize the benefits of Cognitive Systems.
In spite of the significant progress made so far with Cognitive Systems, one of the major challenges businesses will have to overcome is how to wade through data – lightning fast – and find actionable, relevant information that can be used with AI. Another challenge is the misaligned expectations on what Cognitive Systems can actually deliver. So, in 2018, enterprises will need to address these areas to utilize AI in a variety of use cases where relevant and appropriate data is available.
Let’s get automated
Over the past two years, along with Cognitive Systems, AI has also been utilized for automation and robotics. Put simply, this has added an extra dimension of “intelligence” to workflows. Already a number of businesses including the technology titans have integrated their cognitive computing systems.
In 2018, enterprises will be looking to speed up the integration of Cognitive Systems and intelligent automation to maximize the business impact. This is a match made in heaven for two technologies that have one crucial element in common – the utilization of data that has reached maturity. So, how can Cognitive Systems and automation get hitched?
Let’s start the wedding plan with the cake – a three layer cake. Think of the first layer as Cognitive Systems which tries to mimic human behaviour by learning continuously. The next layer is automation which refers to the codification of a manual process into software, sometimes known as scripting.
The pièce de résistance of the cake would be the top layer where automation would combine AI and Machine Learning (ML) capabilities to create intelligent automation. Intelligent automation enabled by AI/Ml technologies could help to process vast quantities of data, actionable intelligence and provide recommendations.
Icing on the cake
This ‘three layer cake’ would not be possible if not for the key ingredients coming into place at the right time. There has been significant progress in computing power, advances in AI/Ml algorithms and an abundance of data ripe of mining.
Cognitive Systems and intelligent automation belong together and arguably, should not be treated as two separate things. They both have AI and ML in common and Intelligent Automation is fundamental to Cognitive Systems. Intelligent automation allows machines to do more than just process data. They can now analyse data sets, spot inconsistencies, check for correctness, and ultimately, make far more intelligent decisions.
Of course, some decisions still require final confirmation from a human operator, but the initial thinking has already been done by the software robot. This has profound implications across different industries. Here are just a few use cases:
- Transportation: This is one industry that is already facing dramatic transformation. Cognitive Systems and automation will touch every aspect of planes, trains and automobiles – from intelligent maintenance to reducing congestion.
- Online security: Cognitive and automated systems would not only detect and block intrusions to a network faster, it would also take steps to self-heal– and intelligently strengthen its defence to block future intrusions. All this would take place in a matter of seconds.
- eHealth: Machines can identify diseases and automatically classify them according to risk bands. Automated systems can assist healthcare professionals with personalised medical care for patient triaging
- Health and safety in energy: Intelligent, cognitive systems would detect problems and keep workers safe in hazardous environments such as mines and oil rigs.
These uses cases only scratch the surface of what’s really possible by marrying intelligent automation with Cognitive Systems. It sounds clichéd but the possibilities are endless for enterprises who now watch dotingly over their data. Data is all grown up.