Organization of the information flow in the perception-action loop of evolved agents

Klyubin, A.S.; Polani, Daniel; Nehaniv, Chrystopher L.

doi:10.1109/eh.2004.1310828

Cited by 52 publications

(65 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It has been shown (Klyubin et al, 2004;Prokopenko et al, 2006;Klyubin et al, 2007) that using this idea as a fitness function can lead to the evolution of robust and adaptive agents, namely maximizing the mutual information between sensors and environment.…”

Section: Law Of Information Organizationmentioning

confidence: 99%

The World as Evolving Information

Gershenson

2012

Unifying Themes in Complex Systems VII

View full text Add to dashboard Cite

This paper discusses the benefits of describing the world as information, especially in the study of the evolution of life and cognition. Traditional studies encounter problems because it is difficult to describe life and cognition in terms of matter and energy, since their laws are valid only at the physical scale. However, if matter and energy, as well as life and cognition, are described in terms of information, evolution can be described consistently as information becoming more complex.The paper presents five tentative laws of information, valid at multiple scales, which are generalizations of Darwinian, cybernetic, thermodynamic, and complexity principles. These are further used to discuss the notions of life and cognition and their evolution.

show abstract

Section: Law Of Information Organizationmentioning

confidence: 99%

The World as Evolving Information

Gershenson

2012

Unifying Themes in Complex Systems VII

View full text Add to dashboard Cite

show abstract

“…In these studies it could be demonstrated that maximization of empowerment can be used for control tasks (such as pole balancing) as well as for an evolution of the sensorimotor system or even to construct contexts which can assign semantic "meaning" to the robot's actions (Klyubin et al 2005(Klyubin et al , 2008. Similar to the work of Klyubin et al (2004Klyubin et al ( , 2005Klyubin et al ( , 2007Klyubin et al ( , 2008 in the study of Prokopenko et al (2006) the authors used two measures called generalized correlation entropy and generalized excess entropy to alter the locomotion of a simulated modular robotic system (snake-like robot) by an evolution process. The mentioned studies differ from our approach, since in these works information measures had been used to drive a sensorimotor adaptation on a relatively large time scales (simulating evolution by using genetic algorithms) whereas in our approach we use information measures to investigate the behavior of closed-loop system during on-line learning on relatively short time scales.…”

Section: Information Flow In Adaptive Closed-loop Systemsmentioning

confidence: 51%

“…In their approach a perception-action loop is interpreted in terms of a communication channel-like model. They show that maximization of information flow can evolve into a meaningful sensorimotor structure (Klyubin et al 2004(Klyubin et al , 2007. In Klyubin et al (2005Klyubin et al ( , 2008) the authors present a universal agent-centric measure, called "empowerment", which is defined as the information-theoretic capacity of an agent's actuation channel (the maximum mutual information for the channel over all possible distributions of the transmitted signal).…”

Section: Information Flow In Adaptive Closed-loop Systemsmentioning

confidence: 99%

See 1 more Smart Citation

Behavioral analysis of differential hebbian learning in closed-loop systems

et al. 2010

View full text Add to dashboard Cite

Understanding closed loop behavioral systems is a non-trivial problem, especially when they change during learning. Descriptions of closed loop systems in terms of information theory date back to the 1950s, however, there have been only a few attempts which take into account learning, mostly measuring information of inputs. In this study we analyze a specific type of closed loop system by looking at the input as well as the output space. For this, we investigate simulated agents that perform differential Hebbian learning (STDP). In the first part we show that analytical solutions can be found for the temporal development of such systems for relatively simple cases. In the second part of this study we try to answer the following question: How can we predict which system from a given class would be the best for a particular scenario? This question is addressed using energy, input/output ratio and entropy measures and investigating Tomas Kulvicius and Christoph Kolodziejski have contributed equally to this work. their development during learning. This way we can show that within well-specified scenarios there are indeed agents which are optimal with respect to their structure and adaptive properties. T. Kulvicius (B)·

show abstract

“…An example of a selection pressure is the acquisition of information from the environment: there is some evidence that pushing the information flow to the information-theoretic limit (i.e., maximization of information transfer in perception-action loops) can give rise to intricate behaviour, induce a necessary structure in the system, and ultimately be responsible for adaptively reshaping the system [13,14]. Other important selection pressures applicable to distributed systems include stability of self-organizing hierarchies [19]; efficiency of multi-cellular communication topologies [20]; efficiency of locomotion and distributed actuation [21,23].…”

Section: Background and Motivationmentioning

confidence: 99%

Optimizing Potential Information Transfer with Self-referential Memory

Prokopenko

Polani

Wang

2006

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper investigates an information-theoretic design principle, intended to support an evolution of a memory structure fitting a specific selection pressure: potential information transfer through the structure. The proposed criteria measure how much does associativity in memory add to the information transfer in terms of precision, recall and effectiveness. Maximization of the latter results in holographic memory structures that can be interpreted in self-referential terms. The study introduces an analogy between self-replication and memory retrieval, with DNA as a partiallyassociative memory containing relevant information. DNA decoding by a complicated protein machinery ("cues" or "keys") may corresponds to an associative recall: i.e., a replicated offspring is an associatively-recalled prototype. The proposed informationtheoretic criteria intend to formalize the notion of information transfer involved in self-replication, and enable bio-inspired design of more effective memory structures.

show abstract

Organization of the information flow in the perception-action loop of evolved agents

Abstract: Sensor evolution in nature aims at improving the acquisition of information from the environment and is in

Cited by 52 publications

References 6 publications

The World as Evolving Information

The World as Evolving Information

Behavioral analysis of differential hebbian learning in closed-loop systems

Optimizing Potential Information Transfer with Self-referential Memory

Contact Info

Product

Resources

About