Back in 2008, Microsoft (NASDAQ:MSFT) filed patent application 2008/134,132, described as a way of "developing software components based on brain lateralization." Back then, the vague patent application may not have seemed like a very useful one. Fast-forward a few years, however, and the human brain mingling with modern technology makes more sense.
Breaking down the language barrier
Microsoft's chief research officer, Rick Rashid, spoke at a presentation in China in November 2012. In the latter part of his speech, the audience heard fluent Mandarin after it was translated from Rashid's English. The speech-to-speech software he used can translate in real time with an estimated error rate of roughly one word in eight, which is about 30% more accurate than previous translators. The video can be seen here.
The translation software was developed using Deep Neural Networks, which are like an artificial neural network that models thousands of interconnected "neurons." Rashid wrote in a blog that it was "patterned after human brain behavior," allowing translations that can even retain a speaker's accent and intonation. At the Chinese presentation, he summed it up with, "We may not have to wait until the 22nd century for a usable equivalent of Star Trek's universal translator."
The technology is still a work in progress, but its sudden success rate is a direct result of abandoning the traditional pattern matching approach to speech translation, and instead patterning it after human brain behavior. Microsoft and Rashid hope "that in a few years we will have systems that can completely break down language barriers."
International Business Machines' big brain
Scientists at International Business Machines (NYSE:IBM) revealed a new computing architecture in August. The new software ecosystem was inspired by the brain -- its size, function, and even minimal use of power -- and has the potential to spawn an entirely new generation of machines that function just like human brains. Dubbed "TrueNorth," IBM's project builds on previous projects surrounding cognitive computing. For example, the company released chips that use "neurosynaptic cores" back in 2011, which manage information similarly to the way neurons function in the brain.
Most modern computer systems have separate units for storing information and processing it sequentially, but like the neurons and synapses in our brains, TrueNorth stores and processes information in a parallel way. According to the MIT Technology Review, researchers want to eventually use the technology in everything from smartphones to cars, and even develop systems as powerful as human vision. IBM currently has plans in place with a partner in Zurich, iniLabs, to develop a visual sensor utilizing TrueNorth.
Qualcomm's brain chips
A brain-on-a-chip may sound like science fiction now, but Qualcomm (NASDAQ:QCOM) has already started on new chips that may turn fiction into fact. This month, CTO Matt Grob told an audience at a conference that by next year, Qualcomm would be taking on partners to start building chips that mimic the neural structures and processing methods of the brain. These chips will enable applications for not only artificial vision sensors, but also things like robot controllers and even brain implants.
Qualcomm has also been working with Brain Corp to develop hardware and algorithms that mimic the human brain's processes. It also built prototype robots with neuro-inspired chips, but research is still in the early stages. Grob explained:
"What is new now is the ability to drop down large numbers of these structures on silicon. The tools we can create are very sophisticated. The promise of this is a kind of machine that can learn, and be programmed without software -- be programmed the way you teach your kid."
The bottom line
Tech companies are starting to build neurons instead of simulate them, looking to create supercomputers that function like our biological brains. While this may seem far-fetched, it's quickly becoming reality as research continues and technology improves.
According to Thomas Serre, a computational neuroscience researcher at MIT, neurons are actually slower than computers. He explained that, "the reason why computers seem much slower is that they are serial machines, while our brains run in parallel." Parallel computing will likely expand technological innovation significantly going forward.
Microsoft, IBM, and Qualcomm have figured this out, and are investing in brain-like technology for the future. Shares of Microsoft currently trade at around 13.5 times earnings, IBM at 13 times earnings, and Qualcomm at around 18 times earnings. All three companies pay income-generating dividends above 2% and are reasonably valued at today's levels.
Joseph Harry owns shares of International Business Machines and Microsoft. The Motley Fool owns shares of International Business Machines, Microsoft, and Qualcomm. Try any of our Foolish newsletter services free for 30 days. We Fools may not all hold the same opinions, but we all believe that considering a diverse range of insights makes us better investors. The Motley Fool has a disclosure policy.