**6.3 Targeted applications for C3S designs (future work)**

Our targeted applications for C3S designs are influenced by the emergence of several dominant industry trends: (1) *Increasing AI Compute Demand*: The computation required for supporting Deep Learning (DL) is increasing exponentially at the staggering rate of doubling every 3.4 months [19]. Many believe that this trend is not technologically nor economically sustainable [20–22]. (2) *Migration Towards Edge AI*: There is strong efficiency, security, and privacy motivations for migrating AI computation from the centralized cloud infrastructure to distributed personal edge devices. (3) *Emerging Ubiquity of Wearables*: Mobile devices, e.g. smartphones, have replaced PCs as the technology, innovation, and profitability driver for mass market computing, due to their more stringent form factor and energy efficiency requirements. As wearable devices become more ubiquitous, we believe they will in turn become the new technology and innovation driver for mass market computing [60]. (4) *Interests*

*Cortical Columns Computing Systems: Microarchitecture Model, Functional Building Blocks… DOI: http://dx.doi.org/10.5772/intechopen.110252*

*in Neuromorphic Chips*: Orders of magnitude improvement on energy efficiency is necessary to support autonomous intelligent wearable devices. We believe to achieve such level of improvement will require adopting the neuromorphic approach in creating new computing devices that mimic the organization and operation of biological neural networks of the neocortex.

Hence, our C3S research focuses on targeting the design and implementation of *Neuromorphic Intelligent Sensory Processing (NISP) Chiplets* with truly brain-like sensory processing capability as well as brain-like energy efficiency. This new genre of C3S based sensory processing units exhibit the following attributes: (1) **Neuromorphic**: based on reverse architecting the neocortex; (2) **Intelligent**: capable of edge-native autonomous and continuous online learning; (3) **Sensory Processing**: support near-sensor on-device processing of diverse modalities of sensory signals; (4) **Chiplets**: target chiplet implementations (potentially UCIe-compliant [14]) in standard digital CMOS, for ease of integration into diverse edge AI platforms on mobile, wearable, and IoT devices.
