Looking back at 2015 and looking forward to 2016, what progress has been made in the field of machine learning?

[[160860]]

A few decades ago, artificial intelligence was a relatively new topic among engineers and developers. But in recent years, machine learning has emerged as an ideal product of big data, and its emergence has injected new vitality into fields such as artificial intelligence.

Here is a brief overview of the impact of machine learning on our lives in 2015 and what we can expect from it in the coming year.

Market changes in 2015

For the uninitiated, machine learning is a sub-discipline of artificial intelligence that involves computer algorithms being able to learn and improve their performance over time by taking in new information without human guidance.

For example, Tesla's self-driving cars used machine learning, which caused a sensation in 2015. But in fact, machine learning can do much more than just transportation.

This year, Google, Amazon, Apple, Facebook, and Microsoft have all started developing machine learning. When Google reorganized into Alphabet, it made it clear that the company would continue to move into machine learning. Apple, Google's longtime rival, wants to hire experts in machine learning to better predict customer needs. Amazon has also launched its own machine learning platform. This summer, Microsoft also unveiled the "mystery" of its Cortana analytics suite.

Although some of the above companies are not pioneers in the field of machine learning, everyone’s attention to machine learning has reflected the important position of machine learning in our daily lives in the future.

Facebook has even launched the artificial intelligence assistant "M", claiming that the virtual assistant "M" created for the communication application Messenger can complete some complex tasks and is very practical.

However, behind the virtual assistant "M" completing tasks, this also hides the most critical point of machine learning: the "human in the loop" mode. This design mode refers to the process and decision that the machine is not sure to make, which is completed by humans. Machine learning does not automate everything. They use algorithms to handle issues such as emails and calendars with clear intentions, and pass more complex messages and requests to humans.

Looking ahead to 2016, a critical year for machine learning

Computers have always been used to improve the efficiency and ability of humans to do things. This is most experienced by front-end users, such as the automatic completion of input and spelling check functions on computers. However, Facebook has completely subverted this model and wants to use humans to improve the work of computers.

Based on the premise of the "human in the loop" model, machine learning will make a qualitative leap in computing and achieve outstanding achievements in many fields in 2016. The qualitative leap of machine learning is mainly reflected in three areas: natural language processing, personalization and security.

Speak the same language

In 2016, natural language processing (also known as NLP) will become more popular with the promotion of virtual assistants “M”.

We expect Facebook to continue to promote the "M", and we also expect its competitors to add the "human" element to their technical models using machine learning.

As developers continue to close the gap between human and electronic language processing, personal assistants can perform more tasks, and the pace of machine learning will accelerate. In addition to personal assistants, NLP will also play an important role in spoken conversations: digital operators will be more interactive and responsive.

In everyday life, NLP can also enable computers to analyze plain text in real time, such as plain text in emails, Word documents, or slides. This will bring many new features, such as automatic fact detection and automatic electronic citations and footnotes. Even things that were thought to be something only humans could do (designing presentations and selecting pictures) can be done automatically by computers.

NLP can also suggest attachments for emails, insert the right charts in board reports, and transform an ordinary list of important items into a high-tech visual feast. In addition, translation services will be greatly improved. Soon, you can have online conversations with people who speak different languages ​​without a translator. The translation is very accurate and there will be almost no misunderstandings.

Get personal

As machine learning understands what you say, it also learns about you, including an in-depth assessment of your personality traits. The more we use our devices, the more information machine learning has. Naturally, it can understand us better.

On the user side, the most direct difference this brings is more distinctive personalized search results. The more we know about users, the fewer unnecessary or even irrelevant ads users are forced to see.

In 2016, the digital world will become your own, personalized world. Your computer will assess your best friends, your favorite foods, and even your mood.

Enhanced security

The more machine learning learns about a user, the better the security will be by improving user authentication.

Behavioral characteristics have proven to be very effective: machine learning technology can create a unique profile of the user, which can analyze how the user operates on the bank's website. If there are any abnormalities, machine learning can also alert the bank in time to prevent fraud.

Machine learning has applications beyond banking. If someone uses your credit card and personal information in a way that doesn’t fit your personality, the credit card company or other security services will be alerted.

This feature can also be applied to device security. For example, if the way your computer and mobile phone run programs or connect to devices does not match your file information, you will receive an alert. Such a system can resist the emergence of malware, Trojans and prevent hackers from invading.

In 2015, we’ve only scratched the surface of the potential of machine learning technology. Today, the most technologically savvy experts are working to create devices that can learn and adapt to changes in the world without the need for a human.

Machine learning has become an increasingly important part of our daily lives. Although some large companies have taken the lead in the field of machine learning, we are more likely to see the emergence of small, nimble startups in 2016 to lead the new direction of machine learning.