Wimbledon spots hot topics with cognitive command center

The Wimbledon organization has been partnering with IBM for more than 25 years, turning tennis into all kinds of data.  A beep can alert umpires into whether a ball is in or out, correlations can be drawn between the number of volleys in a game and the chances of winning, and social data can be mined and matched to players as their popularity grows as they move through the contest.

2016 sees the addition of a cognitive command center, a behind-the-scenes dashboard for the digital editorial team, ingesting feeds from a variety of social media channels to look for common topics and conversations around which to build content. This gives the editorial team a new level of confidence in the content they post and it’s ability to resonate with the audience.

According to PR Newswire:

“For example, the Cognitive Command Center could identify emerging conversations around a Swiss football game at the same time as a peak in interest around a moment in a Roger Federer match. Using these insights, Wimbledon will be able to make rapid content decisions to engage and inform sports fans during a summer filled with numerous major sporting events.”

As the media landscape continues to be redefined and businesses are increasingly becoming their own media houses, think how this model could help almost any business meet the content demands of their target audience.


Discerning narrow and general AI in cognitive applications

Are we close to having robots that can represent (and possibly replace) our intelligence? The answer is yes and no according to an article in the Washington Post.

One reason we don’t have full artificial intelligence (AI) is because the field is made up of many components, with some more advanced than others. As an example, while speech-to-text is well established at this point, we’re still at the embryonic stages of computers obtaining real meaning from video. The field of AI is not moving at uniform pace.

We are also more advanced at having AI be applied to set tasks (such as a chat bot that can book an Uber) rather than the overall bots that try and replace all functions of humanity. Development of these task-based bots equate more to the division of labor we see in business, especially in areas of low-skilled work where we humans perform very similar tasks within a bounded range, and it is these areas where advances in AI have so far had the biggest business impact.

So when thinking about design of cognitive solutions, it’s important to distinguish between applications that are task-based and rely on narrow AI compared to the more challenging general AI applications.



Artificial intelligence brings human touch to travel industry

There was a time when the travel industry used to be represented on the high street. You’d actually talk to travel advisors who would help you find a flight or hotel. But then the industry went digital.

Now we’re getting back closer to the human touch, ironically thanks to artificial intelligence. The NYT highlights how bots are helping travel assistants help you navigate around a new city, and hotels are trialling robots as concierges. The article also points to a growing ecosystem of startups that are accompanying the major tech players in the charge for this new generation of cognitive applications.

What does this mean for other service industries? Can we see the same moves towards using technology to redefine customer experience?

Want some Pepper with your pizza?

Would you really be comfortable with a robotic servant taking your order for pizza or a burger?

You can test this proposition for yourself thanks to a pilot between Pizza Hut and Mastercard launching in Asia. The Pepper robot, that is already undergoing tests as a robotic hotel concierge, will be deployed throughout the region, in an effort to improve customer service and reduce labor costs. The robot can take orders and process payments.

Interestingly, Vipul Chawla, managing director of Pizza Hut Restaurants Asia talks about Pepper offering a “fun, frictionless user experience.”  What are the possibilities here for extension of service? Think if cognitive robotics are deployed in fast food joints in airports. They can know specifically when your flight is going to leave and offer you a service that fits your schedule.

There is also a cultural component at play here. It may be no accident that many of these pilots are currently taking place in Asia, where the idea of robotics and automation in the service industry appears more palatable. There may still be inhibitions that need to break down in the West before we’re comfortable having our pizza served by a knowledgeable hunk of metal, plastic and wires, especially judging by this Business Insider title.


Can IBM Watson get you off the hook with conference calls?

For many of us conference calls can be the bane of modern working life. Focus can be tricky when you realize after 30 mins you’ve only seen the first 5 slides of a 30-slide presentation. Or there’s that minute you tune out and turn on the kettle or check the football scores, and yes… you’re specifically asked a question in front of a bunch of superiors and a 1-minute aberration appears like an hour of dereliction of duty.

Josh Newlan at Splunk felt the same way. In fact he felt strongly enough he turned to IBM Watson to build an app to address this problem for an internal hackathon.  The app kicks in on hearing his name, alerting him to what he’s missed, with a few other accoutrements thrown in for good measure. His HR department is now thinking of sanctioning the app for use company-wide.

There’s a couple of facets to this story to take into account.

Given that the Watson services he needed are available in the cloud, all Josh needed was a trial account. There was no need for a formal relationship between Splunk and IBM for Josh to build his app. Gone are the days when the IT department can specify what products are services can be used to build which tools. With the rise of the API Economy, the landscape has broadened considerably.

Innovation can now come from anywhere within the organization. It is no longer just the purview of one particular department. We are seeing the rise of the citizen developer. Innovative companies are helping to facilitate these not-so-random acts of development through initiatives like hackathons.

In this era of open innovation, what developers dream, they can build. The companies that can harness this powerful new paradigm will be the ones that win out.

Google sees cognitive apps as competitive differentiator

Many of the product announcements at the search giant’s I/O conference had a common theme running through them: an ‘ambient experience’ where the application uses context and what it knows about the user’s behavior to offer a more personalized experience. As ZDNet points out, “The glue for all of these play-from-behind items is artificial intelligence, context, personalization and sheer computing power”. Google will also offer these machine learning capabilities to developers through its cloud platform.

Artificial intelligence and cognitive computing is seen as an assistant to help technology better assist us humans, wherever we are, including at home.


The US elections: a cognitive computing pundit ‘speaks’

How does Donald Trump stack up against Hillary when it comes to empathy? Does Bernie Sanders hold himself in higher regard than Donald Trump? While these questions may be debated by (professional and armchair) pundits across the US, Alex Thompson from the NYT thought he would give the IBM Watson service a run for its money.

The results?

See what he found.

He used the IBM Watson Personality Insights service to explore social media and abstracts from books to model the personalities of the presidential contenders. What does this mean for business? The service is available through the cloud via an API call. So businesses can incorporate this service into their own apps and just pay for what they use. For instance an eCommerce site which asks users to login via Facebook or Twitter could use the service to understand their propensity to buy certain goods (eg. eco-friendly) and offer specific, personalized recommendations.

As another example, a business could also compare the profiles of the C-Suite to determine who is the most effective spokesperson. We also saw a group at Berkeley use this service to determine what kind of tweets from a sports team would most likely gain new followers.

Is your Teaching Assistant real or a bot?

Georgia Institute of Technology runs a massive artificial intelligence online course, with a forum with over 40k questions. To help manage student questions, this year computer science professor Ashok Goel enlisted the help of “Jill Watson”, an automated reply system developed using IBM Watson technology to help answer student questions. Most students were unable to tell that they were being aided by a bot rather than a human Teaching Assistant. One specific aspect of this bot is that it would not get involved in a discussion unless it assessed a 97% confidence rate of answering the question.

As The Wall Street Journal states, this is an interesting use case in how such bots could transform learning, especially as we see courses move to the digital realm.

Read more on this story on Science Daily. Note that this post was created by human.

Beware the bots: the missteps of Microsoft’s Tay

If you are going to engage in research in how computers can converse with humans, the internet is not always the best place to start. This is what Microsoft learned when it launched Tay, an artificial intelligence project, meant to converse on social networks like Twitter and Kik. Apparently a ‘coordinated effort’ to teach the bot by internet communities (it learned from conversations using machine learning) led to Tay spewing out increasingly racist and sexist remarks.  Within 24 hours Microsoft had to close down the project.

As TechCrunch notes, for organizations experimenting in this space, if you are developing cognitive apps for open social networks, you need to make sure you have sufficient anti-abuse measures in place to avoid bots picking up anti-social traits.

See the official response from Microsoft.