A Neural Dynamic Network Drives an Intentional Agent That Autonomously Learns Beliefs in Continuous Time

Tekülve, Jan; Schöner, Gregor

doi:10.1109/tcds.2020.3013768

Cited by 3 publications

(2 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Regular synaptic learning rules are a natural part of the DFT framework (Sandamirskaya 2014 ) and can be used to understand how the patterned connectivity arises that gives nodes their sensory-motor meaning (Sandamirskaya and Schöner 2010 ; Tekülve and Schöner 2020 ). When successful, such accounts explain how learning unfolds autonomously as a neural dynamic architecture generates mental and behavioral states.…”

Section: Discussionmentioning

confidence: 99%

Neural dynamic foundations of a theory of higher cognition: the case of grounding nested phrases

Sabinasz,

Richter,

Schöner

2023

Cogn Neurodyn

View full text Add to dashboard Cite

Because cognitive competences emerge in evolution and development from the sensory-motor domain, we seek a neural process account for higher cognition in which all representations are necessarily grounded in perception and action. The challenge is to understand how hallmarks of higher cognition, productivity, systematicity, and compositionality, may emerge from such a bottom-up approach. To address this challenge, we present key ideas from Dynamic Field Theory which postulates that neural populations are organized by recurrent connectivity to create stable localist representations. Dynamic instabilities enable the autonomous generation of sequences of mental states. The capacity to apply neural circuitry across broad sets of inputs that emulates the function call postulated in symbolic computation emerges through coordinate transforms implemented in neural gain fields. We show how binding localist neural representations through a shared index dimension enables conceptual structure, in which the interdependence among components of a representation is flexibly expressed. We demonstrate these principles in a neural dynamic architecture that represents and perceptually grounds nested relational and action phrases. Sequences of neural processing steps are generated autonomously to attentionally select the referenced objects and events in a manner that is sensitive to their interdependencies. This solves the problem of 2 and the massive binding problem in expressions such as “the small tree that is to the left of the lake which is to the left of the large tree”. We extend earlier work by incorporating new types of grammatical constructions and a larger vocabulary. We discuss the DFT framework relative to other neural process accounts of higher cognition and assess the scope and challenges of such neural theories.

show abstract

Section: Discussionmentioning

confidence: 99%

Neural dynamic foundations of a theory of higher cognition: the case of grounding nested phrases

Sabinasz,

Richter,

Schöner

2023

Cogn Neurodyn

View full text Add to dashboard Cite

show abstract

“…The neural machinery for sequentially organizing cognitive operations may likewise be supported by innate substrate and shaped by cognitive development. Given such systems, learning the connectivity pattern for individual concepts and relations from experience may become a well‐defined theoretical problem (see Tekülve & Schöner, 2020, for an account in a related context). The broader issues of autonomous learning are relevant to all aspects of the architecture, however: How are neural dynamic architectures tuned to ensure that meaningful sequences of cognitive operations emerge?…”

Section: Discussionmentioning

confidence: 99%

A Neural Dynamic Model of the Perceptual Grounding of Spatial and Movement Relations

2021

View full text Add to dashboard Cite

How does the human brain link relational concepts to perceptual experience? For example, a speaker may say "the cup to the left of the computer" to direct the listener's attention to one of two cups on a desk. We provide a neural dynamic account for both perceptual grounding, in which relational concepts enable the attentional selection of objects in the visual array, and for the generation of descriptions of the visual array using relational concepts. In the model, activation in neural populations evolves dynamically under the influence of both inputs and strong interaction as formalized in dynamic field theory. Relational concepts are modeled as patterns of connectivity to perceptual representations. These generalize across the visual array through active coordinate transforms that center the representation of target objects in potential reference objects. How the model perceptually grounds or generates relational descriptions is probed in 104 simulations that systematically vary the spatial and movement relations employed, the number of feature dimensions used, and the number of matching and nonmatching objects. We explain how sequences of decisions emerge from the time-and state-continuous neural dynamics, how relational hypotheses are generated and either accepted or rejected, followed by the selection of new objects or the generation of new relational hypotheses. Its neural realism distinguishes the model from information processing accounts, its capacity to autonomously generate sequences of processing steps distinguishes it from deep neural network accounts. The model points toward a neural dynamic theory of higher cognition.

show abstract

Neural Dynamic Principles for an Intentional Embodied Agent

Tekülve,

Schöner

2024

Cognitive Science

View full text Add to dashboard Cite

How situated embodied agents may achieve goals using knowledge is the classical question of natural and artificial intelligence. How organisms achieve this with their nervous systems is a central challenge for a neural theory of embodied cognition. To structure this challenge, we borrow terms from Searle's analysis of intentionality in its two directions of fit and six psychological modes (perception, memory, belief, intention‐in‐action, prior intention, desire). We postulate that intentional states are instantiated by neural activation patterns that are stabilized by neural interaction. Dynamic instabilities provide the neural mechanism for initiating and terminating intentional states and are critical to organizing sequences of intentional states. Beliefs represented by networks of concept nodes are autonomously learned and activated in response to desired outcomes. The neural dynamic principles of an intentional agent are demonstrated in a toy scenario in which a robotic agent explores an environment and paints objects in desired colors based on learned color transformation rules.

show abstract

A Neural Dynamic Network Drives an Intentional Agent That Autonomously Learns Beliefs in Continuous Time

Cited by 3 publications

References 32 publications

Neural dynamic foundations of a theory of higher cognition: the case of grounding nested phrases

Neural dynamic foundations of a theory of higher cognition: the case of grounding nested phrases

A Neural Dynamic Model of the Perceptual Grounding of Spatial and Movement Relations

Neural Dynamic Principles for an Intentional Embodied Agent

Contact Info

Product

Resources

About