With its countless breakthroughs, Machine Learning is one of the most active research area in science. Many companies use this technology intensively, the results are mostly good — in some cases, excellent — and keep getting more accurate over time. Let's see some situations where these algorithms shine.
Using machine learning (and deep leaning) techniques — more specifically, in the fields of pattern recognition and anomaly detection — you can train a model to detect any object in any type of data, including a threat in a program or a malware/ransomware.
It's becoming popular to detect malware via behavioral analysis methods instead of signature based ones. Which mean that algorithm can learn that a software is probably a threat if it does X, Y, and Z, instead of learning that it's probably a threat because it contain some attributes (signatures) known to be malicious.
A malware detection algorithm can also tracks a computer's softwares activity for long periods of time and alert you when they do something unusual, in order to to protect you from malware that hide themselves as "benign" softwares.
Many big companies now use A.I. in their process to detect malware, you may want to take a look at Cyberbit.
Major search engines like Google or Bing use machine learning in many components of their algorithms, so lets do a list :
- Many steps of search ranking
- Spelling suggestion or correction
- Related searches
- Generating sitelinks
- Page classification (is it a blog ? forum ? ...)
- Spam detection
- Sentiment analysis
Yes, it's probably used in almost every part of many search engines.
You likely already used Amazon, Netflix or Youtube at least once, didn't you ? Well, they all use a machine learning algorithm to generate suggestion. Products, series, or videos, it doesn't really matter, they do the same thing : providing content based on previous actions of the user to maximise the occurrence of an event — buying a product or watching a video in these cases. The morality of these software can be controversial, but one thing is sure : it works amazingly well, who never spent too much time on Youtube ?
I'm sure you already heard of it a thousand times, how autonomous vehicles will impact our lives, with projects like Waymo, Drive.ai, or Uber, so I would rather like to make a brief overview (for non-tech).
Each sensor is (in most of the cases) connected to its own specific algorithm, which gives its output to the system that makes the final decisions. Many (almost all) of theses softwares rely on machine learning, so with a enough labeled data — like pictures with bounding boxes around other vehicles, but it's not sufficient — the algorithms can detect and analyse its surrounding, and make decisions it "thinks" is best. So it's kind of like a human, but with more sensors, quicker reflexes, and less mistakes.
Many fear autonomous vehicles because they think that it's less safe than a human driver, but that's wrong. Actually, removing the human makes traveling way more safer ! The algorithm never gets tired, falls asleep or pays no attention, and will always keep track of what happens in the surrounding, at a 360 degrees angle.
13 minor fender-benders in more than 1.8 million miles of autonomous and manual driving—and still, not once was the self-driving car the cause of the accident. — Jacquelyn Miller, Google spokeswoman
Of course, there are way more applications of machine learning — some are more interesting or creative — but this will be for another blog post.