Perceptual computing: the next big thing

Perceptual computing is a new technique in the already crowded technology area. Perceptual computing may reinvent the way we interact with our computing devices, making it more natural, intuitive and immersive. The computing devices will be able to perceive user actions through new capabilities including close range hand gestures, finger articulation, speech recognition, face tracking, augmented reality experiences, and more.

Perceptual computing is an efficient alternative to  the mouse and keyboard and controls computers with gesture, eye motions, and voice.


Let’s look at the architecture of the Perceptual computer to understand how perceptual computing is implemented.

Perceptual Computer consists of three components: encoder, computing with words (CWW) engine and decoder. Perceptions such as words are provided as an input in order to activate the perceptual computer and  the words are presented as perceptual computer output along with the data, thereby interacting with the user using vocabulary.

The encoder transforms the words into fuzzy sets and leads to a codebook with their fuzzy set models. The output of the encoder activates a computing with words (CWW) engine whose output is other fuzzy sets which are then mapped by the decoder into a recommendation with supporting data. The recommendation may be in the form of a word, group of similar words, rank or class.

The perceptual computer establishes vocabulary for an application and collects data in certain intervals from various sources for all the words in the vocabulary. The perceptual computer maps the collected data into words resulting in the codebook for the application and completes the functioning of the encoder in the perceptual computer. The CWW engine maps the fuzzy sets into one or more fuzzy sets. The CWW engine may consist of If-Then rules and Linguistic Weighted Averages. If there is a CWW engine for the application, then perceptual computer uses its available mathematics to compute its output. Otherwise, the perceptual computer develops mathematics for the new kind of the CWW engine. The new CWW engine should be constrained such that its output resembles the words in the codebook for the application.

Further, the perceptual computer maps the fuzzy set output from the CWW engine into a recommendation at the output of the decoder. If the recommendation is a word, rank or class, then the perceptual computer uses existing mathematical logic for mapping. Otherwise, the perceptual computer develops a new mathematical logic for the new kind of decoder.

Intel is betting big on perceptual computing and has announced devices which work on the technique. Though perceptual computing is still in development stage, it presents innovative way of using your imagination and shows us the vision for the future of computing.


Author, Chandrasekhar is Manager, Patent and Trademark Agent at IP Astra


Semiconductor Industry: Major Challenges

What makes the electronics industry tick? How do we bridge the gap between the electronic goods manufactured in India and the electronic goods consumed in India? What role does IP play in the semiconductor industry?

Dr Ronald Black, president and CEO of Rambus Incorporated provided valuable insights on these topics at a talk hosted by IESA at ITC gardenia, Bangalore on the 2nd of July 2013. Roughly 65% of the electronic goods in India are imported.

Out of the 35% electronic goods manufactured in India,  Only 10% account for high value manufacturing. Bridging the gap between the manufactured goods and consumed goods is the key to self-reliance with regard to electronics for India. Technology and design trends on an economic standpoint, implications of IP, changes in start-up culture, implications for Indian design centres, were some of the things that were discussed in the talk.

We have seen the semiconductor industry taking great leaps since 1972. The transistor sizes have shrunk into diminutive proportions. On another side of the coin, the semiconductor fabs have grown to be so incredibly expensive that save for a few gigantic corporations like Intel, Samsung, TSMC and Global Founders, it is virtually impossible for a company to provide the funding required for a fab. A major problem that the semiconductor companies face today is that the costs of masks that are used to produce integrated circuits have skyrocketed. What used to cost a few hundred thousand dollars a few decades ago have become tens of millions of dollars today. In such a condition it is extremely critical for the produced integrated circuits to be right at the first time. mistakes in designing today costs several times more today.

A single mistake in designing could push the company into bankruptcy. Traditionally, dramatic cost reduction has been observed in case of transistors as new technologies emerge. It has been predicted that the technology that the 28 nm node shall last longer than the other introduced nodes and hence change in IP shall occur less frequently. A 2012 study shows that the average number of Intellectual property blocks used in a system on chip (SoC) shall rise to 150 by 2016. There has been an increase in the IP developed by others being used in SoCs, moreover, the IP that has been used earlier are used by the companies are being increasingly reused. The same 2012 study mentioned earlier predicts that the 75% of the IP in a SoC shall be IP that had been used earlier.

For India, development of SoC design capability is critical, but so is semiconductor IP creation, or partnerships to ensure that one has access to the highest quality semiconductor IP. However, having the lowest cost design team, one that delivers poor quality because of the IP used, would be a problem – best in class and first time right designs are not just important, they will be a matter of survival. China’s state-run capitalism has greatly benefited their country. They have outstanding infrastructure and local strategic investment, and though semiconductor design has been slow, and software even slower, they are making significant progress. India, if bold enough can take a true global leadership role, leveraging its strong software and growing semi designing firms. Key to success will be to develop a strong local demand for technology, not just to perform low-cost outsourcing.

Author, Leo Paul Johnson, is a patent engineer working at IP ASTRA