Cognitive computers may be able to predict cataclysmic events, including tsunamis, and highlight complex risks in the world’s volatile financial markets.

There is never a dull moment in technology. Every day creates a more dynamic enterprise environment and rapid change.

Science fiction conjures up dramatic stories, each one more action-packed than the other. One of the central themes in the narrative is the clash of machines with humans.

Imagine “Skynet” in Terminator, the main antagonist, an artificially intelligent system that becomes self-aware and revolts against the human race. The future may not be as life-threatening, but we are inching towards building a computer that behaves like a human brain.

Auto-fill, location-enabled searches, context-aware computing, and other semi self-aware services are already part of daily life. Cognitive computing claims to combine psychology and technology, in “parallel processing”: an attempt to analyze new, often unexpected information.

The Brain Chip

There is ongoing empirical research on modeling elements of the cerebral cortex. One of Dreamfire’s key business partners, IBM is actively involved in creating chips that mimic the human brain.

The DARPA (Defence Advanced Research Projects Agency) is funding the SyNAPSE project, dedicated to advancing the cause of cognitive computing. A team of computational modelers, psychologists etc. plan to develop and embed biological neural networks in adaptable neuromorphic chips.

Exponential growth in transistor density (generally known as Moore’s Law) has led to the semiconductor industry exploding and the global technology sector making inroads everywhere. The growth trajectory is bound to reach defined physical limits. This essentially can lead to commoditization and obsolescence.

The strategic business angle for cognitive computing is keeping the market for tech majors alive and kicking. Building a biologically scalable chip can change existing growth paradigms.

Self-Aware Systems

The human cortex has about 22 billion neurons and 220 trillion synapses, which makes it 400 times bigger than the rat brain simulation which has already been created. This took 8 terabytes of memory on a 32,768-processor BlueGene/L supercomputer. In a typical human body, the brain is the hardware that directs every action.

The IBM project attempts to simulate the cerebral cortex, responsible for functions such as thought, computation, and action; while other parts of the brain handle emotions, co-ordination and vital functions.

The new chips can rewire themselves “on the fly”, reaching decisions through integrated memory, computation and communication cores that resemble synapses, neurons and axons, respectively, in the brain’s nervous system. It is modeled on the brain.

Here, neurons are digital processors that compute the information; synapses, which fire to pass the information, learning it and committing it to memory; and axons are data pathways that connect the tissue to the computer.

A New Architecture

The brain is an example of a multithreaded architecture; it can handle parallel tasks with ease. Context switching is difficult, if the same part of the brain wants to switch between similar tasks. An example of this would be talking on a phone and taking notes at the same time. The Von Neumann architecture, on which existing chips were built, is a sequential method but has limitations.

It pushes data from the processor to the memory storage through a tiny tube called a bus. Forcing a lot of data in can result in a Von Neumann bottleneck; essentially a traffic jam between the processor and memory.

There are multiple applications of the technology.Cognitive computers may be able to predict cataclysmic events, including tsunamis, and highlight complex risks in the world’s volatile financial markets. The engineering challenge in creating this is huge, as developing the processing capabilities of a chip that is programmable, small and requires extremely low power is difficult.

However, IBM’s recent announcement will make technologists heave a sigh of relief. Building a new generation of computers that use supercomputing, neuroscience, and nanotechnology is set to become easier with cognitive computing. It also spells profits for commerce, industry and the entire ecosystem.

Share article on social media: