5 Ways You’re Already Using Deep Learning
To put it simply, deep learning is a type of machine learning algorithm that mimics the neurons and synapses found in the human brain. It allows for the modeling of complex data patterns in multilayered networks that continuously self-improve. Deep learning was first developed in the 1950s during psychologist Frank Rosenblatt’s quest to create a mechanical brain. Rosenblatt’s machine, the Perceptron, delighted many as it accurately identified shapes without human assistance. However, the public fanfare soon died down as critics pointed out the system’s limits. While deep learning has only recently resurfaced to the public’s interest, technology has enabled deep learning to evolve significantly over the past 60 years.
In 2012, deep learning made front page news through a New York Times article that claimed it would enable machines to “perform human activities like seeing, listening, and thinking.” It’s clear that we’re shifting away from the classic approach to programming, which consists of coding machines with a set of rules. Coupled with the big data revolution, deep learning is certain to increase the accuracy of machines in many industries, from manufacturing to health care. And when it comes to people’s excitement about deep learning, “putting your money where your mouth is” couldn’t be truer. In 2015 alone, there were about 300 VC deals totaling $2.8 billion into artificial intelligence companies.
But for the average American, how much does deep learning impact our everyday lives? Apparently, a lot! Here are five ways that you’re already tapping into deep learning-fueled technology:
1. Language translation
Google’s language translation program uses deep learning to analyze various “soft boundaries and ambiguities” of language for more accurate and grammatically correct sentences.
2. Anti-fraud technology
Raise your hand if you’ve been a victim of identity theft! If so, you know that it’s not such a fun experience. Fortunately, electronic payment systems are now using deep learning to strengthen their ability to pinpoint suspicious transactions quickly.
3. Digital assistants
Most digital assistants, including the “beloved” Siri, use voice recognition based on deep learning. Apple purchased a natural language tech company called VocalIQ in 2015, with plans to make customers’ conversations with Siri more nuanced and natural. We identify IVA® most closely with this type of technology; in fact, IVA® already uses natural language and communicates with customers in the way that Apple is striving to achieve.
4. Photo tagging
Deep learning enables Facebook’s ability to suggest tags on photos based on people’s faces. Their facial recognition project, Deepface, has a 97.25% accuracy in telling whether two photos have the same face.
5. Speech-to-text
Have you ever dictated text using speech recognition on your phone? If so, you’re using transcription technology powered by Deep Neural Networks (DNN)! Language modeling facilitates the recognition of individual sounds, and the matching of these sounds to existing words.
In the customer service industry, we’ve tapped into the power of deep learning to anticipate customers’ needs based on past communications. For example, if a certain type of customer or phrase is linked to checking order status, a deep learning software module will shift its top menu so that the order status option is presented first to certain customer profiles. In effect, customer service technologies will eventually process customers’ inputs much like a human – and probably better – leading to a more streamlined and specific process for each customer need. In the long run, our machines will train themselves to take a proactive approach to customer service, and prevent customers from being confused or frustrated with self-service in the first place.
Of course, deep learning is nowhere near 100% accuracy. Once in a while, Facebook suggests tagging you as your mom, or Siri fails to comprehend even your best enunciation. And like we mentioned above, this is why an increasing number of companies are committing to furthering the integration of artificial intelligence into their services. In fact, the recently released Artificial Intelligence Global Forecast report forecasts that the artificial intelligence and machine learning market will grow more than 50% by 2020, to a staggering $5.05 billion. There’s no limit to the exciting changes ahead, but for now, sit back and enjoy the deep learning renaissance.
Sources:
https://www.technologyreview.com/s/513696/deep-learning/
https://www.technologyreview.com/s/544606/can-this-man-make-aimore-human/
http://www.fastcompany.com/3026423/why-google-is-investing-in-deep-learning
http://venturebeat.com/2015/10/02/apple-acquires-deep-learning-startup-vocaliq-to-make-siri-smarter/
http://googleresearch.blogspot.com/2015/08/the-neural-networks-behind-google-voice.html
http://www.huffingtonpost.com/pradeep-aradhya/artificial-intelligence-d_b_9698204.html
