A Theory of Cheap Control in Embodied Systems.
Ontology highlight
ABSTRACT:
SUBMITTER: Montufar G
PROVIDER: S-EPMC4556690 | biostudies-literature | 2015 Sep
REPOSITORIES: biostudies-literature
Publications
A Theory of Cheap Control in Embodied Systems.
Montúfar Guido G Ghazi-Zahedi Keyan K Ay Nihat N
PLoS computational biology 20150901 9
We present a framework for designing cheap control architectures of embodied agents. Our derivation is guided by the classical problem of universal approximation, whereby we explore the possibility of exploiting the agent's embodiment for a new and more efficient universal approximation of behaviors generated by sensorimotor control. This embodied universal approximation is compared with the classical non-embodied universal approximation. To exemplify our approach, we present a detailed quantita ...[more]
Similar Datasets
Project description:Embodied artificial cognitive systems, such as autonomous robots or intelligent observers, connect cognitive processes to sensory and effector systems in real time. Prime candidates for such embodied intelligence are neurally inspired architectures. While components such as forward neural networks are well established, designing pervasively autonomous neural architectures remains a challenge. This includes the problem of tuning the parameters of such architectures so that they deliver specified functionality under variable environmental conditions and retain these functions as the architectures are expanded. The scaling and autonomy problems are solved, in part, by dynamic field theory (DFT), a theoretical framework for the neural grounding of sensorimotor and cognitive processes. In this paper, we address how to efficiently build DFT architectures that control embodied agents and how to tune their parameters so that the desired cognitive functions emerge while such agents are situated in real environments. In DFT architectures, dynamic neural fields or nodes are assigned dynamic regimes, that is, attractor states and their instabilities, from which cognitive function emerges. Tuning thus amounts to determining values of the dynamic parameters for which the components of a DFT architecture are in the specified dynamic regime under the appropriate environmental conditions. The process of tuning is facilitated by the software framework cedar, which provides a graphical interface to build and execute DFT architectures. It enables to change dynamic parameters online and visualize the activation states of any component while the agent is receiving sensory inputs in real time. Using a simple example, we take the reader through the workflow of conceiving of DFT architectures, implementing them on embodied agents, tuning their parameters, and assessing performance while the system is coupled to real sensory inputs.
| S-EPMC5089998 | biostudies-literature
Project description:Do we have to presuppose a self to account for human self-consciousness? If so, how should we characterize the self? These questions are discussed in the context of two alternatives, i.e., the no-self position held by Metzinger (2003, 2009) and the claim that the only self we have to presuppose is a narrative self (Dennett, 1992; Schechtman, 2007; Hardcastle, 2008) which is primarily an abstract entity. In contrast to these theories, I argue that we have to presuppose an embodied self, although this is not a metaphysical substance, nor an entity for which stable necessary and jointly sufficient conditions can be given. Self-consciousness results from an integration of an embodied, basic affective flow with an intentional object (the self as agent or as center of imagination or thought), where this integration remains anchored in an embodied self. This embodied self is a flexible and variable entity, which we can account for only with a pattern theory of the self (in line with Gallagher, 2013). Furthermore, I outline how this pattern theory of the self fits into the predictive coding framework, which also answers the open question whether self-representation is prior to world-representation or the other way around. The principal organization of a mechanism of building up a self-model is such that both types of representations are always activated and developed in parallel. Modeling oneself is a process which is always activated when one interacts with the world - much as a shadow is present when a person walks in the sun.
| S-EPMC6265368 | biostudies-other
Project description:Understanding the real-time dynamical mechanisms of neural systems remains a significant issue, preventing the development of efficient neural technology and user trust. This is because the mechanisms, involving various neural spatial-temporal ingredients [i.e., neural structure (NS), neural dynamics (ND), neural plasticity (NP), and neural memory (NM)], are too complex to interpret and analyze altogether. While advanced tools have been developed using explainable artificial intelligence (XAI), node-link diagram, topography map, and other visualization techniques, they still fail to monitor and visualize all of these neural ingredients online. Accordingly, we propose here for the first time "NeuroVis," real-time neural spatial-temporal information measurement and visualization, as a method/tool to measure temporal neural activities and their propagation throughout the network. By using this neural information along with the connection strength and plasticity, NeuroVis can visualize the NS, ND, NM, and NP via i) spatial 2D position and connection, ii) temporal color gradient, iii) connection thickness, and iv) temporal luminous intensity and change of connection thickness, respectively. This study presents three use cases of NeuroVis to evaluate its performance: i) function approximation using a modular neural network with recurrent and feedforward topologies together with supervised learning, ii) robot locomotion control and learning using the same modular network with reinforcement learning, and iii) robot locomotion control and adaptation using another larger-scale adaptive modular neural network. The use cases demonstrate how NeuroVis tracks and analyzes all neural ingredients of various (embodied) neural systems in real-time under the robot operating system (ROS) framework. To this end, it will offer the opportunity to better understand embodied dynamic neural information processes, boost efficient neural technology development, and enhance user trust.
| S-EPMC8751631 | biostudies-literature
Project description:The efficacy of evolutionary or reinforcement learning algorithms for continuous control optimization can be enhanced by including an additional neural network dedicated to features extraction trained through self-supervision. In this paper we introduce a method that permits to continue the training of the features extracting network during the training of the control network. We demonstrate that the parallel training of the two networks is crucial in the case of agents that operate on the basis of egocentric observations and that the extraction of features provides an advantage also in problems that do not benefit from dimensionality reduction. Finally, we compare different feature extracting methods and we show that sequence-to-sequence learning outperforms the alternative methods considered in previous studies.
| S-EPMC8049238 | biostudies-literature
Project description:Skyrmion is an emergent particle consisting of many spins in magnets, and has many nontrivial features such as (i) nano-scale size, (ii) topological stability, (iii) gyrodynamics, and (iv) highly efficient spin transfer torque, which make skyrmions the promising candidate for the magnetic devices. Earlier works were focusing on the bulk or thin film of Dzyaloshinskii-Moriya (DM) magnets, while recent advances are focusing on the skyrmions induced by the interfaces. Therefore, the superstructures naturally leads to the interacting skyrmions on different interfaces, which has unique dynamics compared with those on the same interface. Here we theoretically study the two skyrmions on bilayer systems employing micromagnetic simulations as well as the analysis based on Thiele equation, revealing the reaction between them such as the collision and bound state formation. The dynamics depends sensitively on the sign of DM interactions, i.e., helicities, and skyrmion numbers of two skyrmions, which can be well described by Thiele equation. Furthermore, we have found the colossal spin-transfer-torque effect of bound skyrmion pair on antiferromagnetically coupled bilayer systems.
| S-EPMC5309827 | biostudies-other
Project description:Outbreaks of the corallivorous crown-of-thorns seastar Acanthaster planci (COTS) represent one of the greatest disturbances to coral reef ecosystems in the Indo-Pacific, affecting not only coral reefs but also the coastal communities which rely on their resources. While injection approaches are increasingly used in an attempt to control COTS densities, most of them display severe drawbacks including logistical challenges, high residual environmental impacts or low cost-effectiveness. We tested a new alternative control method based upon acidic injections of cheap, 100% natural products. We investigated the lethal doses, intra- and inter-specific disease transmission and immune responses of COTS when injected with fresh lime juice (extracted from local Citrus arantifolia) and white spirit vinegar. High COTS mortality was achieved with small volumes: 10-20 ml per seastar induced death in 89%/97% of injected specimens after an average 34.3 h/29.8 h for lime juice and vinegar respectively. Highest efficiency was reached for both solutions with double shots of (2 × 10 ml) in two different areas on the body: 100% mortality occurred within 12-24 h, which is similar or faster compared with other current injection methods. Multiple immune measures suggested that death was very likely caused by pH stress from the acidic solutions rather than a bacterial infection. Contagion to either conspecifics or a variety of other reef species was not observed, even at COTS densities 15 times higher than the highest naturally reported. 10 to 20 l lime juice/vinegar could kill up to a thousand COTS at a cost of less than 0.05 USD per specimen; no permits or special handling procedures are required. We conclude that injections of lime juice and vinegar offer great advantages when compared to current best practises and constitute a cheap and natural option for all reefs affected by COTS.
| S-EPMC4565713 | biostudies-literature
Project description:The past several years have witnessed an increased presence of control theoretic concepts in synthetic biology. This review presents an organized summary of how these control design concepts have been applied to tackle a variety of problems faced when building synthetic biomolecular circuits in living cells. In particular, we describe success stories that demonstrate how simple or more elaborate control design methods can be used to make the behaviour of synthetic genetic circuits within a single cell or across a cell population more reliable, predictable and robust to perturbations. The description especially highlights technical challenges that uniquely arise from the need to implement control designs within a new hardware setting, along with implemented or proposed solutions. Some engineering solutions employing complex feedback control schemes are also described, which, however, still require a deeper theoretical analysis of stability, performance and robustness properties. Overall, this paper should help synthetic biologists become familiar with feedback control concepts as they can be used in their application area. At the same time, it should provide some domain knowledge to control theorists who wish to enter the rising and exciting field of synthetic biology.
| S-EPMC4971224 | biostudies-literature
Project description:This paper examines the contributions of dynamic systems theory to the field of cognitive development, focusing on modeling using dynamic neural fields. A brief overview highlights the contributions of dynamic systems theory and the central concepts of dynamic field theory (DFT). We then probe empirical predictions and findings generated by DFT around two examples-the DFT of infant perseverative reaching that explains the Piagetian A-not-B error, and the DFT of spatial memory that explain changes in spatial cognition in early development. A systematic review of the literature around these examples reveals that computational modeling is having an impact on empirical research in cognitive development; however, this impact does not extend to neural and clinical research. Moreover, there is a tendency for researchers to interpret models narrowly, anchoring them to specific tasks. We conclude on an optimistic note, encouraging both theoreticians and experimentalists to work toward a more theory-driven future.
| S-EPMC4454421 | biostudies-literature
Project description:Quantum embedding theory aims to provide an efficient solution to obtain accurate electronic energies for systems too large for full-scale, high-level quantum calculations. It adopts a hierarchical approach that divides the total system into a small embedded region and a larger environment, using different levels of theory to describe each part. Previously, we developed a density-based quantum embedding theory called density functional embedding theory (DFET), which achieved considerable success in metals and semiconductors. In this work, we extend DFET into a density-matrix-based nonlocal form, enabling DFET to study the stronger quantum couplings between covalently bonded subsystems. We name this theory density-matrix functional embedding theory (DMFET), and we demonstrate its performance in several test examples that resemble various real applications in both chemistry and biochemistry. DMFET gives excellent results in all cases tested thus far, including predicting isomerization energies, proton transfer energies, and highest occupied molecular orbital-lowest unoccupied molecular orbital gaps for local chromophores. Here, we show that DMFET systematically improves the quality of the results compared with the widely used state-of-the-art methods, such as the simple capped cluster model or the widely used ONIOM method.
| S-EPMC5754786 | biostudies-literature
Project description:By convention, slices in a tomosynthesis reconstruction are created on planes parallel to the detector. It has not yet been demonstrated that slices can be generated along oblique directions through the same volume, analogous to multiplanar reconstructions in computed tomography (CT). The purpose of this work is to give a proof-of-principle justification for oblique reconstructions in tomosynthesis, which acquires projection images over a smaller angular range than CT.To investigate the visibility of individual frequencies in an oblique reconstruction, a theoretical framework is developed in which the reconstruction of a sinusoidal input is calculated. The test frequency is pitched at an angle in a 2D parallel-beam acquisition geometry. Reconstructions are evaluated along the pitch of the object. The modulation transfer function (MTF) is calculated from the relative signal at various test frequencies. The MTF determines whether modulation is within detectable limits in oblique reconstructions. In the previous linear systems (LS) model [B. Zhao and W. Zhao, "Three-dimensional linear system analysis for breast tomosynthesis," Med. Phys. 35(12), 5219-5232 (2008)], the MTF was calculated only in reconstructed slices parallel to the detector. This paper generalizes the MTF calculation to reconstructed slices at all possible pitches. Unlike the previous LS model, this paper also analyzes the effect of object thickness on the MTF. A second test object that is considered is a rod whose long axis is pitched similar to the sinusoidal input. The rod is used to assess whether the length of an object can be correctly estimated in oblique reconstructions.To simulate the conventional display of the reconstruction, slices are first created along a 0° pitch. This direction is perpendicular to the rays of the central projection. The authors show that the input frequency of a pitched sinusoidal object cannot be determined using these slices. By changing the pitch of the slice to match the object, it is shown that the input frequency is properly resolved. To prove that modulation is preserved in pitched slices, the MTF is also calculated. Modulation is within detectable limits over a broad range of pitches if the object is thin, but is detectable over a narrower range of pitches if the object is thick. Turning next to the second test object, it is shown that the length of a pitched rod can be correctly determined in oblique reconstructions. Concordant with the behavior of the MTF, the length estimate is accurate over a broad range of pitches if the object is thin, but is correct over a narrower range of pitches if the object is thick.This work justifies the feasibility of oblique reconstructions in tomosynthesis. It is demonstrated that pitched test objects are most easily visualized with oblique reconstructions instead of conventional reconstructions. In order to achieve high image quality over a broad range of pitches, the object must be thin. By analyzing the effect of reconstruction pitch and object thickness on image quality, this paper generalizes the previous LS model for tomosynthesis.
| S-EPMC3808438 | biostudies-other