Filtering Data Based on Human-Inspired Forgetting

IEEE TRANSACTIONS ON CYBERNETICS (Impact Factor: 6.22). 01/2012; DOI: 10.1109/TSMCB.2011.2157142
Source: IEEE Xplore


Robots are frequently presented with vast arrays of diverse data. Unfortunately, perfect memory and recall provides a mixed blessing. While flawless recollection of episodic data allows increased reasoning, photographic memory can hinder a robot's ability to operate in real-time dynamic environments. Human-inspired forgetting methods may enable robotic systems to rid themselves of out-dated, irrelevant, and erroneous data. This paper presents the use of human-inspired forgetting to act as a filter, removing unnecessary, erroneous, and out-of-date information. The novel ActSimple forgetting algorithm has been developed specifically to provide effective forgetting capabilities to robotic systems. This paper presents the ActSimple algorithm and how it was optimized and tested in a WiFi signal strength estimation task. The results generated by real-world testing suggest that human-inspired forgetting is an effective means of improving the ability of mobile robots to move and operate within complex and dynamic environments.

1 Follower
8 Reads
  • Source
    • "Artificial intelligence and vision, which determine robots' understanding of the 3D world that surrounds them, also requires additional research (Costa et al., 2011). Finally, information processing, such as the ability to forget in order to purge useless information to avoid overloading sensors, must be improved to make these machines' performance consistent with their mission to operate in complex environments (Freedman and Adams, 2011). "
    [Show abstract] [Hide abstract]
    ABSTRACT: This report identifies nine emerging technological trends based on 21 technological foresight documents published by various specialized businesses and public agencies. These trends bring together technologies with the potential to initiate lasting transformation in the digital ecosystem, which we define as all of the infrastructure, software applications, content, and the social practices that determine how the ecosystem is used. The notion of an ecosystem allows us to examine in an integrated manner the interactions between the technical, economic, social, political and legal dimensions of this complex assemblage. These nine trends are as follows: 1) Cloud computing; 2) Big data; 3) The Internet of things; 4) Mobile Internet; 5) Brain-computer interfaces; 6) Near-field communication (NFC) payments; 7) Mobile robots; 8) Quantum computing; 9) Internet militarization/weaponization. The most frequently appearing implications include the increased number of opportunities for malicious attacks, the lack of consideration for security needs during the development of the technologies in question, even when these technologies are used to carry out financial transactions, the dilution of mechanisms for controlling system integrity because of the ever-increasing interconnection of machines, or the erosion of user privacy, including personal information that represents an irresistible source of added value to organizations.A few of the following themes that appear common to all nine trends are also mentioned in the conclusion: the interdependence of the technologies examined, which will require the implementation of integrated security policies to prevent a counterproductive fragmentation of resources; the expansion and diversification of the digital ecosystem, which will also require sophisticated coordination policies; the transformation of the notion of privacy; the convergence of the problems of cyber security with those of human security; the indispensable balance between having adequate cyber security and maintaining the economic and technical competitiveness that depends on a certain regulatory freedom; the risks of groups of individuals adopting self-defence practices in the event states fail to provide security; and finally positive contributions of the nine trends to cyber security.
    SSRN Electronic Journal 01/2012; DOI:10.2139/ssrn.2208548
  • [Show abstract] [Hide abstract]
    ABSTRACT: An algorithm is presented for autonomous motion development with unbounded waveform resolution. Rather than a single optimization in a very large space, memory is built to support incremental improvements; therefore, complexity is balanced by experience. Analogously, human development manages complexity by limiting it during initial learning stages. Motions are represented by cubic spline interpolation; therefore, the development technique applies broadly to function optimization. Adding a node to the splines allows all previous memory samples to transfer to the higher dimension space exactly. The memory-based model, which is a locally weighted regression (LWR), predicts the expected outcome for a motion and provides gradient information for optimizing the motion. Results are compared against bootstrapping a direct optimization (DO) on a mathematical problem. Additionally, the method has been implemented to learn voltage profiles with the lowest peak current for starting a motor. This method shows practical accuracy and scalability.
    Cybernetics, IEEE Transactions on 08/2013; 43(4):1178-1188. DOI:10.1109/TSMCB.2012.2226026 · 3.47 Impact Factor