51 Hits in 1.6 sec

NMODE --- Neuro-MODule Evolution [article]

Keyan Ghazi-Zahedi
2017 arXiv   pre-print
Modularisation, repetition, and symmetry are structural features shared by almost all biological neural networks. These features are very unlikely to be found by the means of structural evolution of artificial neural networks. This paper introduces NMODE, which is specifically designed to operate on neuro-modules. NMODE addresses a second problem in the context of evolutionary robotics, which is incremental evolution of complex behaviours for complex machines, by offering a way to interface
more » ... o-modules. The scenario in mind is a complex walking machine, for which a locomotion module is evolved first, that is then extended by other modules in later stages. We show that NMODE is able to evolve a locomotion behaviour for a standard six-legged walking machine in approximately 10 generations and show how it can be used for incremental evolution of a complex walking machine. The entire source code used in this paper is publicly available through GitHub.
arXiv:1701.05121v1 fatcat:wfrhcootg5dlfp6s6h3l6dxf34

Information Theoretically Aided Reinforcement Learning for Embodied Agents [article]

Guido Montufar, Keyan Ghazi-Zahedi, Nihat Ay
2016 arXiv   pre-print
Svenja is simulated in YARS (Zahedi et al., 2008) , which uses the physics engine bullet.  ...  Zahedi et al. (2013) studied the predictive information as a supplement to an extrinsic task related reward.  ... 
arXiv:1605.09735v1 fatcat:r6nqqgrdmvhxlpak7odn5vwnoa

Reinforcement Learning of Artificial Microswimmers [article]

Santiago Muiños-Landin, Keyan Ghazi-Zahedi, Frank Cichos
2018 arXiv   pre-print
The behavior of living systems is based on the experience they gained through their interactions with the environment [1]. This experience is stored in the complex biochemical networks of cells and organisms to provide a relationship between a sensed situation and what to do in this situation [2-4]. An implementation of such processes in artificial systems has been achieved through different machine learning algorithms [5, 6]. However, for microscopic systems such as artificial microswimmers
more » ... ch mimic propulsion as one of the basic functionalities of living systems [7, 8] such adaptive behavior and learning processes have not been implemented so far. Here we introduce machine learning algorithms to the motion of artificial microswimmers with a hybrid approach. We employ self-thermophoretic artificial microswimmers in a real world environment [9, 10] which are controlled by a real-time microscopy system to introduce reinforcement learning [11-13]. We demonstrate the solution of a standard problem of reinforcement learning - the navigation in a grid world. Due to the size of the microswimmer, noise introduced by Brownian motion if found to contribute considerably to both the learning process and the actions within a learned behavior. We extend the learning process to multiple swimmers and sharing of information. Our work represents a first step towards the integration of learning strategies into microsystems and provides a platform for the study of the emergence of adaptive and collective behavior.
arXiv:1803.06425v2 fatcat:2hkuxjopr5gedklkqkd5ldi6da

Morphological Computation: Synergy of Body and Brain

Keyan Ghazi-Zahedi, Carlotta Langer, Nihat Ay
2017 Entropy  
There are numerous examples that show how the exploitation of the body's physical properties can lift the burden of the brain. Examples include grasping, swimming, locomotion, and motion detection. The term Morphological Computation was originally coined to describe processes in the body that would otherwise have to be conducted by the brain. In this paper, we argue for a synergistic perspective, and by that we mean that Morphological Computation is a process which requires a close interaction
more » ... f body and brain. Based on a model of the sensorimotor loop, we study a new measure of synergistic information and show that it is more reliable in cases in which there is no synergistic information, compared to previous results. Furthermore, we discuss an algorithm that allows the calculation of the measure in non-trivial (non-binary) systems.
doi:10.3390/e19090456 fatcat:uwt6cjqxffcglf2cbcwjdxa5ja

Quantifying Morphological Computation

Keyan Zahedi, Nihat Ay
2013 Entropy  
The field of embodied intelligence emphasises the importance of the morphology and environment with respect to the behaviour of a cognitive system. The contribution of the morphology to the behaviour, commonly known as morphological computation, is well-recognised in this community. We believe that the field would benefit from a formalisation of this concept as we would like to ask how much the morphology and the environment contribute to an embodied agent's behaviour, or how an embodied agent
more » ... an maximise the exploitation of its morphology within its environment. In this work we derive two concepts of measuring morphological computation, and we discuss their relation to the Information Bottleneck Method. The first concepts asks how much the world contributes to the overall behaviour and the second concept asks how much the agent's action contributes to a behaviour. Various measures are derived from the concepts and validated in two experiments which highlight their strengths and weaknesses.
doi:10.3390/e15051887 fatcat:i4t5cb7dkrczvfsnyyrwfyvtca

SO(2)-Networks as Neural Oscillators [chapter]

Frank Pasemann, Manfred Hild, Keyan Zahedi
2003 Lecture Notes in Computer Science  
Using discrete-time dynamics of a two neuron network with recurrent connectivity it is shown that for specific parameter configurations the output signals of neurons can be of almost sinusoidal shape. These networks live near the Sacker-Neimark bifurcation set, and are termed SO(2)-networks, because their weight matrices correspond to rotations in the plane. The discretized sinus-shaped waveform is due to the existence of quasi-periodic attractors. It is shown that the frequency of the
more » ... rs can be controlled by only one parameter. Signals from the neurons have a phase shift of π/2 and may be useful for various kinds of applications; for instance controlling the gait of legged robots.
doi:10.1007/3-540-44868-3_19 fatcat:ggllv24fgbciveg5xtjitv4jby

Quantifying Morphological Computation based on an Information Decomposition of the Sensorimotor Loop [article]

Keyan Ghazi-Zahedi, Johannes Rauh
2015 arXiv   pre-print
As in (Zahedi and Ay, 2013) , the following two assumptions are made without loss of generality.  ...  We also know from previous experiments (see Zahedi and Ay, 2013) , that the conditional mutual information I(W : W |A) drops to zero for increasing µ.  ... 
arXiv:1503.05113v1 fatcat:5eavdzd4lzevhcnefhcmwyogby

On the Causal Structure of the Sensorimotor Loop [chapter]

Nihat Ay, Keyan Zahedi
2014 Guided Self-Organization: Inception  
This article deals with the causal structure of an agent's sensorimotor loop. Of particular interest are causal effects that can be identified from an agent-centric perspective based on in situ observations. Within this identification, the world model of the agent plays a central role. Furthermore, various kinds of information flows through the sensorimotor loop are considered, including causal as well as associational ones. Transfer entropy and predictive information are discussed in more
more » ... l. The maximization of the latter leads to coordinated behavior within an experimental case study. Corresponding results on the relation to morphological computation are presented. Index Terms -Embodiment, Sensorimotor Loop, Morphological Computation, Causal Information Flows. the next state of X as well as the next state of Y only depend on the current state of Y .
doi:10.1007/978-3-642-53734-9_9 fatcat:wsfqj3hctraexb52iur4fuzv5u

Editorial: Recent Trends in Morphological Computation

Keyan Ghazi-Zahedi, John Rieffel, Syn Schmitt, Helmut Hauser
2021 Frontiers in Robotics and AI  
., 2014; Zahedi and Ay, 2013; Ghazi-Zahedi et al., 2016) .  ... 
doi:10.3389/frobt.2021.708206 pmid:34136536 pmcid:PMC8202073 fatcat:l5kwc5shdnctvossnfulgel7rq

Representing Robot-Environment Interactions by Dynamical Features of Neuro-controllers [chapter]

Martin Hülse, Keyan Zahedi, Frank Pasemann
2003 Lecture Notes in Computer Science  
This article presents a method, which enables an autonomous mobile robot to create an internal representation of the external world. The elements of this internal representation are the dynamical features of a neuro-controller and their time regime during the interaction of the robot with its environment. As an examples of this method the behavior of a Khepera robot is studied, which is controlled by a recurrent neural network. This controller has been evolved to solve an obstacle avoidance
more » ... . Analytical investigations show that this recurrent controller has four behavior relevant attractors, which can be directly related to the following environmental categories: free space, obstacle left/right, and deadlock situation. Temporal sequences of those attractors, which occur during a run of the robot are used to characterize the robot-environment interaction. To represent the temporal sequences a technique, called macro-action maps, is applied. Experiments indicate that macro-action maps allow to built up more complex environmental categories and enable an autonomous mobile robot to solve navigation tasks.
doi:10.1007/978-3-540-45002-3_13 fatcat:tzh3dkiu45avpkgi4pia2uk2bm

Causal Effects for Prediction and Deliberative Decision Making of Embodied Systems [chapter]

Nihat Ay, Keyan Zahedi
2013 Advances in Cognitive Neurodynamics (III)  
This article deals with the causal structure of an agent's sensori-motor loop and its relation to deliberative decision making. Of particular interest are causal effects that can be identified from an agent-centric perspective based on in situ observations. Within this identification, an optimal world model of the agent plays a central role. Its optimality is characterized in terms of prediction quality.
doi:10.1007/978-94-007-4792-0_67 fatcat:rms4fnxr4jgpbdoxqmxkldj4ke

Quantifying Morphological Computation based on an Information Decomposition of the Sensorimotor Loop

Keyan Ghazi-Zahedi, Johannes Rauh
2015 07/20/2015-07/24/2015  
As in (Zahedi and Ay, 2013) , the following two assumptions are made without loss of generality.  ...  We also know from previous experiments (see Zahedi and Ay, 2013) , that the conditional mutual information I(W : W |A) drops to zero for increasing µ.  ... 
doi:10.7551/978-0-262-33027-5-ch017 dblp:conf/ecal/Ghazi-ZahediR15 fatcat:lnsms4i7kvhvleezl6v6hfw4ay

Geometry and Determinism of Optimal Stationary Control in Partially Observable Markov Decision Processes [article]

Guido Montufar, Keyan Ghazi-Zahedi, Nihat Ay
2016 arXiv   pre-print
It is well known that for any finite state Markov decision process (MDP) there is a memoryless deterministic policy that maximizes the expected reward. For partially observable Markov decision processes (POMDPs), optimal memoryless policies are generally stochastic. We study the expected reward optimization problem over the set of memoryless stochastic policies. We formulate this as a constrained linear optimization problem and develop a corresponding geometric framework. We show that any POMDP
more » ... has an optimal memoryless policy of limited stochasticity, which allows us to reduce the dimensionality of the search space. Experiments demonstrate that this approach enables better and faster convergence of the policy gradient on the evaluated systems.
arXiv:1503.07206v2 fatcat:7e7s74p5mrbkhm7rjstfcaig6q

On the records [article]

Andrew Berdahl, Uttam Bhat, Vanessa Ferdinand, Joshua Garland, Keyan Ghazi-Zahedi, Justin Grana, Joshua A. Grochow, Elizabeth Hobson, Yoav Kallus, Christopher P. Kempes, Artemy Kolchinsky, Daniel B. Larremore, Eric Libby, Eleanor A. Power (+1 others)
2017 arXiv   pre-print
World record setting has long attracted public interest and scientific investigation. Extremal records summarize the limits of the space explored by a process, and the historical progression of a record sheds light on the underlying dynamics of the process. Existing analyses of prediction, statistical properties, and ultimate limits of record progressions have focused on particular domains. However, a broad perspective on how record progressions vary across different spheres of activity needs
more » ... rther development. Here we employ cross-cutting metrics to compare records across a variety of domains, including sports, games, biological evolution, and technological development. We find that these domains exhibit characteristic statistical signatures in terms of rates of improvement, "burstiness" of record-breaking time series, and the acceleration of the record breaking process. Specifically, sports and games exhibit the slowest rate of improvement and a wide range of rates of "burstiness." Technology improves at a much faster rate and, unlike other domains, tends to show acceleration in records. Many biological and technological processes are characterized by constant rates of improvement, showing less burstiness than sports and games. It is important to understand how these statistical properties of record progression emerge from the underlying dynamics. Towards this end, we conduct a detailed analysis of a particular record-setting event: elite marathon running. In this domain, we find that studying record-setting data alone can obscure many of the structural properties of the underlying process. The marathon study also illustrates how some of the standard statistical assumptions underlying record progression models may be inappropriate or commonly violated in real-world datasets.
arXiv:1705.04353v2 fatcat:wwg2fdh5dzcs7eoy4njvajlgua

Linear combination of one-step predictive information with an external reward in an episodic policy gradient setting: a critical analysis [article]

Keyan Zahedi and Georg Martius and Nihat Ay
2013 arXiv   pre-print
The simulator YARS [Zahedi et al., 2008] was used for all experiments conducted in this section. Different values for the PGPE parameters were evaluated.  ...  Campos et al., 2010 , von Twickel et al., 2011 , Markelić and Zahedi, 2007 , thereby using the same neuron model as in the cart-pole experiment (see above).  ... 
arXiv:1309.6989v1 fatcat:7hotimj2qzhi7dqkoxejdx6cwa
« Previous Showing results 1 — 15 out of 51 results