Machine Learning, Artificial Intelligence, Neural Networks… These are not just hot “buzzwords” anymore. 2017 has seen the deployment of very real solutions that are no longer confined to the imagination of innovators of the likes of Elon Musk, and before the year wraps up we wanted to take a look at some of the main technology breakthroughs of 2017.
Just last month, Google announced the launch of TensorFlow Lite, a lightweight version of TensorFlow, Google’s open source library for machine learning. TensorFlow Lite is a solution specifically aimed at mobile and embedded devices, built on 3 pillars: small size, cross-platform capabilities, and optimized performance.
Technical jargon aside, the arrival of TensorFlow Lite will open new opportunities to improve on-device image recognition, conversational models that provide one-touch replies to incoming chat messages, and enhanced camera + voice interaction models, among others.
Even though TensorFlow already supported mobile deployment of machine learning applications through the TensorFlow Mobile API, Google is pushing TensorFlow Lite as an “evolution” of TensorFlow Mobile, which means that the latter will most likely be phased out in the near future.
One of the biggest challenges in Artificial Intelligence is developing a system that is capable of holding a decent conversation. Facebook has lead the charge in this area with its implementation of chatbots, and took a step further this year with the release of the ParlAI framework.
In short, ParlAl offers a simpler way to build conversational AI systems, plus the ability to combine different approaches to improve machine dialogue, in order to build smarter chatbots that won’t get so easily confused by an unexpected question.
The ultimate goal of ParlAI, which is built on Python, is to help develop a complete question-answering system that can recognize and respond appropriately to the nuisances of language. Although this is a highly ambitious goal, researchers expect this platform to make great progress in the next 5 years.
Accelerating the development of self-driven vehicles became a priority for Silicon Valley’s top players this year. Waymo, the autonomous vehicle branch of Alphabet (Google’s parent company) has been testing a fleet of driverless cars in Arizona since October, with the goal of offering a commercial ride-hail service integrated with platforms like Uber and Lyft.
In the same field, Uber recently announced a $1 billion purchase of self-driving cars from Volvo, which expands on the $300 million partnership that these companies struck last year. Also, the lesser-known company Otto, which outfits trucks with the equipment needed to drive themselves for long-haul delivery, was acquired by Uber for a reported $680 million last year, which has given them access to 500 engineers at Uber that have focused on improving self-driving technology this year.
Big money is definitely pouring into the autonomous vehicle field, which is expected to become commercially viable and widespread within the next 5 to 10 years.
Now for the downside…
Not everything has been roses in tech in 2017. There’s definitely a darker side of technology that has had a negative impact this year.
2017 was the year of “fake news”. Pressured by intense scrutiny from Congress, Facebook was forced to accept that fake posts from Russia had reached the news feed of 126 million people, which generated intense debate about the dangerous power and influence of social media networks in the U.S. presidential election. This is far from the only criticism that platforms like Facebook and Youtube have received this year, as they have also been condemned for their inefficiency in policing content and avoiding the proliferation of violent videos and hate speech from extremist groups.
Interestingly, just this week it was reported that a team of college students from Yale developed a Chrome extension called Open Mind, with the intention of countering fake and biased news. When active, the plugin will display a warning when the user enters a site that has been flagged for spreading fake news. In addition, the plugin will be able to identify political slant and bias, and will recommend alternative stories from the other side of the spectrum.
Lastly, the tech debate has started to shed light on the relentless pace of automation, which entails the potential loss of millions of jobs. Just considering automated vehicles, a report by the White House estimates that self-driving technology could threaten up to 3.1 million jobs in the U.S. alone.
At the end, it seems wise that on the new year the decision makers within the tech ecosystem take pause and reflect on the famous words of Dr. Ian Malcom: "Your scientists were so preoccupied with whether or not they could, that they didn't stop to think if they should."