Big Blue Is Still the Big Dog of Patents
IBM said today that it was granted 6,478 patents in 2012, which is also a record. The company has about 8,000 researchers and inventors working in 46 states in the U.S. and 35 countries around the world. The full tally of patents over 20 years amounts to nearly 67,000.
The next nine companies in the Top 10 list of patent recipients are:
- Samsung, which received 5,081 patents
- Canon: 3,174
- Sony: 3,032
- Panasonic: 2,769
- Microsoft: 2,613
- Toshiba: 2,447
- Hon Hai : 2,013
- General Electric: 1,652
- LG Electronics: 1,624
So what does IBM do that other companies don’t? I had a quick conversation with Katherine Frase, IBM’s VP for Industry Solutions and Emerging Business. “The process of getting to so many patents means that inside the company there’s a mindset that’s geared toward writing down what you do when do something that’s original and that has business value. There’s a tangible focus on writing things down. And you’ll write down five to 10 times the number of things that actually pass muster toward getting a patent. But that process keeps the notion of innovation at the front of your mind, not at the back of the mind,” she says. “It shows up in patents, and that’s an indicator, but the cultural assumption that innovation isn’t an accident but is made up of lots of little things that you remembered to capture along the way is more important than the patents themselves.”
So, what sorts of things did IBM receive patents for in 2012? Here’s a sampling:
U.S. Patent #8,275,803: System and method for providing answers to questions. Remember Watson? The talking supercomputer that cleaned humanity’s clock on the TV game show “Jeopardy,” and then followed it up by going to medical school and becoming a big-shot doctor working on treating cancer? This would be the patent on how Watson takes in questions expressed in natural language and returns an answer.
U.S. Patent #8,250,010: Electronic learning synapse with spike-timing dependent plasticity using unipolar memory-switching elements. If Watson weren’t enough for you at mimicking and improving upon humanity, IBM is working on something even more complex: Systems of Neuromorphic Adaptive Plastic Scalable Electronics, or SyNAPSE. It’s a project focused on cognitive computing aimed at emulating the workings of the human brain.
U.S. Patent #8,185,480: System and method for optimizing pattern recognition of non-Gaussian parameters. I don’t have the slightest idea what a non-Gaussian parameter is, so I’m not going to even try to explain this one, beyond saying that it has to do with recognizing patterns in data as the volume of information grows. One example IBM gives is traffic data: If you’re measuring traffic patterns, every day you get more data, and thus the patterns change and evolve, or existing ones become more pronounced and predictable.
There are 6,475 more or these patents from 2012 and, no, I won’t even try to list any more. Here’s a short video that IBM produced on the subject: