Here we go again. The last two years felt like living in a perpetual deja-vu of the 90's. Virtual Reality is back (…and still disappoints), Pokemon is a hit, and tech stocks are booming. I was about to write a “and a Clinton in the Whitehouse” but well that played out otherwise.
It was 2001 when I built my first cluster-computer back in Honolulu, Hawaii using an array of old Sun Sparcs to support my research on dnalib (molecular computing using DNA) occasionally working with Sandia National Labs in Livermore, California to fix bugs in the server software. Overall it was a tedious, manual, and error prone way of doing things.
Fast forward to 2017, clusters, load-balancing, and high-performing distributed computers are an off-the-shelf product. Highly available, distributed Computing infrastructure has become like water. We take it for granted. Back in 2001 I could only dream about my watch having the computing power of a super computer in the back-end. Now it is a reality.
What an amazing time we live in.
We are now living in an era where a natural language processing system is just a (long) press away on your phone and facial recognition is a standard feature to tag people in images on popular social platforms (if we want it or not).
Big Platform Support
This revolution, and that’s the only word to properly describe it, has been build on large-scale cloud-based services. An impressive accomplishment by the the four horsemen of technology (Amazon, Google, IBM, and Microsoft) who have invested Billions in large-scale distributed cloud services, data analytics services, and business intelligence solutions and are now developing feverishly the new playing field inspiring engineers worldwide and effectively unites these technological areas and moves their product offerings to a new level.
Big platforms give computing power a price point. Need a few terabyte of hard-drive space ? Or a few petaflops of computing power? All of this is swiftly available to the right credit-card holders. No need to buy your own hardware. Instantly scalable to a million users.
Most cognitive services offerings hover around the topic areas of computer vision, computer speech, knowledge discovery, and language processing. All of these are AI-related techniques to train computers to recognize objects in photos and understand human language. Some of them aim to mimick the functionality of human brain and they use synthesis of not just information sources but also of influences, contexts, and insights. For drop!in we combine geospatial event data with data mapping functions and customer sentiment feedback gathered from the app to make the app provide more relevant results.
Drilling for oil on Venice Beach
The Data Rush
A lot of companies continually gather data about the various aspects of their business (commerce data, transaction data, or increasingly social data). Rarely, companies collect biometric data, real-world geospatial or telematic data, or health/fitness related data.
We all heard of the stunning statistics that about 90% of all the data in the world has been generated in the past two years. But that will only increase with the rise of IoT technology. The more we gather this kind of data the value of it degrades quickly over time. Who cares about your Tweet from last year or your lost In Shadows game from 2013?
IoT technology will transfer short-term data about the health of humans, but also billions of machinery like fridges, cars, lifts, escalators, clocks, and other machinery and so enable “predictive maintenance”. Who has this manpower ? You guessed right. Cognitive services.
IoT devices and especially wearables will solve of the main problems of data collection. With smartphone-based tracking it was required for users to start/stop their activities with wearables this tracking is continuous.
Data is moving faster than ever before. Already between 50% and 70% of all trades on US stock exchanges were performed by algorithms created by financial engineers reacting within milliseconds to changes of information and market trends.
And that’s the game-changer right there. Developers have public access to this high-performing infrastructure at a negligible cost compared to 16 years ago.
Strong Open Source and Developer Support
Whether it is Amazon Alexa, Amazon AWS, Google’s SyntaxNet /Tensorflow , IBM Watson, or Microsoft Cognitive Services, for every major computing platform there is a sophisticated fully featured developer network with API’s, documentations, and discussion forums.
All of these companies are releasing these AI software toolkits (SDK’s) for free so engineers learn how to use these tools and build even better solutions in the future .
I was invited in March 2017 to IBM Connect a huge conference here in Singapore hosted by IBM mainly because of my work on drop!in and especially the Twitter based Event Bot @DropinIn , I built prototypes with Microsoft’s Face recognition API for a large financial services company in Asia.
Developers are direly needed to expand use cases and companies are saving no cost to keep you engaged in their solution. This developer support strengthens the existing AI-related algorithms and lead to new more powerful algorithms, which further leverages the utilization of the large-scale cloud infrastructure and improves model quality.
But more importantly, this human component redefines the relationship between the people and their increasingly pervasive digital environment. The research field of Human / Robot interaction will be one of the most interesting areas in the years to come. Especially since in a working environment Siri, as an example, are far from flawless.
The future
Cognitive Services will change significantly how we will interact with systems. Albeit most Organizations have just barely begun to scratch the surface of cognitive computing capabilities and people are just starting to realize how they influence our daily lives.
Companies like tenqyu and other forward looking developers are already seeing the writing on the wall rushing quickly to adapting their service offerings to this new world. Apps are moving out of the walled gardens of Apple and Google towards messaging platforms to provide one integrated face to the customers of large social platforms.
The cognitive era has begun and with it money to be earned. We started teaching systems how to understand inputs without explicit instructions. Greater availability of key skills and certifications (like certified Watson Application Developers) will be critical in the evolution and adoption of the technology towards the larger populace. Because we humans need to understand how to interact with this new intelligence. It is like in the hitch-hikers guide to the Galaxy’s. The answer is 42. But what was the question?
This article was brought to you by tenqyu, a Singaporean startups making urban living more fun, healthy, inclusive, and thriving using big data, machine learning, and LOTs of creativity.
[2]https://www.fool.com/investing/general/2016/06/02/for-under-armour-data-is-the-new-oil.aspx
[4]http://www.theverge.com/2014/7/7/5878069/why-facebook-is-beating-the-fbi-at-facial-recognition
[5] https://www-01.ibm.com/software/data/bigdata/what-is-big-data.html
[6] https://www.ibm.com/watson/developercloud/services-catalog.html