IBM launches programming environment for cognitive computing devices

Said to be radically different to the way current devices are programmed, IBM's approach is intended for use on distributed, highly interconnected, large scale cognitive computing architectures. "Architectures and programs are closely intertwined and a new architecture necessitates a new programming paradigm," said Dr Dharmendra Modha, principal investigator and senior manager of IBM Research. "We are working to create a Fortran for synaptic computing chips. While complementing today's computers, this will bring forth a fundamentally new technological capability in terms of programming and applying emerging learning systems." When synaptic chips were created, each core had to be programmed individually. To help build cognitive applications, the team has created a programming ecosystem based on corelets. The ecosystem comprises a simulator, neuron model, programming model, library and laboratory. In advance of chips, developers can begin to write code by using a simulator. This is a multithreaded, massively parallel and highly scalable functional software simulator of a cognitive computing architecture, comprising a network of neurosynaptic cores. Neural models, meanwhile, are simple, digital, highly parameterised units that form a fundamental information processing unit of brain like computation. A network of neurons may be able to sense, remember and act upon a variety of stimuli. The programming model is based on 'corelets'. Each corelet is a network of neurosynaptic cores that performs a given function. Corelets can be combined to produce larger, more complex units. IBM has developed a library with more than 150 corelets. IBM's long term goal is to build a system with 10billion neurons and 100trillion synapses which consumes 1kW and occupies less than 2litre.