During this Technical Week, we asked whether the new technology could provide a better way to gauge public opinion during the election. We also met a robot that promised to guide us to the airport and we heard of a battle between two large Silicon Valley Autonomous Car Technology.
The performance of polls has been strictly controlled over the past year – and has found desires. Investigators are confident that the other side will win the EU referendum and Hillary Clinton will beat Donald Trump as President of the United States, and they are wrong.
But that seems to have caused some interesting ideas on how to use alternative methods to gauge general mood and we talked with two technology experts who thought they could beat the house poll.
Rob Lancashire at Essencient believes that his company uses natural language processing to analyze moods on Twitter that may receive signs of shirking polls.
Now, when I became the BBC’s Digital Election Reporter in 2010, this kind of emotional analysis of social media is taking place – and the results are not impressive.
It turned out that Twitter is something of a democracy-liberal democracy movement and analysts estimate too many seats it will gain in the election.
However, Lancashire claims that his technology is far more complex when it comes to evaluating the deeper meaning of tweets – and what they express about the intentions of someone specifically voting. He said it had uncovered both the results of the EU referendum and the final change for Donald Trump in the US election.
And he says there is a flaw in the center of what pollsters do that avoids their approach. “The trouble with asking a question is that you are immediately giving a favoritism to the process,” he said.
But Qriously, another company we interviewed, is all about asking questions, which are set for global smartphone users. It buys advertising space to ask a simple question in the smartphone app.
If someone answers, then they will have more queries, perhaps on election priority or on a product – Qriously’s clients are mostly market research companies.
Founder Chris Kahler showed me a map with flashes of feedback from smartphone users across the UK. “It was a very simple idea,” he said. “The tricky part is deciding who will serve the questions and how to understand the results when you get them back.”
That’s where machine learning comes in, with the teaching system itself evaluating the ability of a person to agree to a survey. Like Essencient, Chris Kahler claims to have successfully predicted the outcome of the EU referendum and the presidential election.
And his technology seems to have convinced both investors and customers. Qriously was backed up by the venture capital firm, invested in Twitter and Oculus and was hired a few weeks ago to predict the outcome of the French presidential election.
Before the first round, it put a code online that will unlock a document containing its predictions once the final result is known. Surveys have done pretty well so far in assessing the mood of the French people – let’s see if this new approach can fit them.
Hitachi Robot – Here to help
The receptionist at the Hitachi office in London greeted me in Japanese when I visited, but quickly turned to English when I clarified my misunderstanding. This multilingual icon is in fact a robot named Emiew, on his first trip to Europe.
Emiew, like Pepper Softbank, is a friendly robot designed to interact with people in a variety of customer service functions. It has been tested at Haneda Airport in Tokyo, where it has provided information to visitors – and not just those who approach it.
The robot is connected to different surveillance cameras and the brain connects to its cloud which allows discovering people who look as though they might need help and reach them. I suggested to designer Rachel Jones that in the UK some people might not like this type of interaction.
She explained that an important part of the Emiew project was to assess the different cultural attitudes to the robot: “Not only language problems but etiquette but I think there is a question. Wider for society about where we want to go and how we see robots fit in”
During the interaction with Emiew, I was bored with a lengthy explanation of the various facilities available at the Hitachi office and was severely interrupted, something I certainly would not do with reception of people.
Seventy years ago, Isaac Asimov introduced three robotic codes – showing how the robots behaved.
Now, as they enter more areas in our lives, we may need to update that code and include instructions on how we should treat them.
By: Anna Lee