The main objective of this paper is to propose a quantum-inspired approach to extracting the maximum possible semantic value from sensor data, even when the quality and level of detail of that data varies widely. The approach has broader cognitive implications as well.
The approach is based on a pragmatic view of naturally occurring quantum systems, referred to in this paper as Wave Packet Network (WPN) theory. This pragmatic view of quantum mechanics interprets large classical systems as multi-scale networks of wave packets. Exchanges of information between pairs of WPN packets cause both packets to restructure ("collapse") in ways that restrict the sets of possible futures associated with each wave packet.
The premise of this paper is that well-designed emulations of WPN using modern simulation hardware could lead to a general model for exploiting sensor data. Available sensor data would "collapse" a set of simulated wave packets that represent the best current interpretation of the external world. It is postulated that WPN emulation will enable sensor processing that is more top-down, more linear in time, and more compatible with control systems due to its use of object-like packets to represent the world.
