Nessy3l: a Neurosymbolic System with 3 Levels
نویسنده
چکیده
This paper presents an hybrid neurosymbolic system composed of three levels, contrary to the majority of other hybrid systems which only consists of two levels, one symbolic and one neural. By using more levels we gain simple interfaces , more modularity, a better task sharing between the components of the hybrid system. Our supplementary level is based on rule translation into a neural network, and we present a versatile neural network which is convenient. The three levels architecture is currently applied to the problem of obstacle avoidance in robotics.
منابع مشابه
Generating and Classifying Recall Images by Neurosymbolic Computation
The neurosymbolic hybrid system ARCS, which extends a classifier for certain kinds of visually presented objects, generates recall images it is then capable of classifying. The modules performing classification are exploited for imagery, too. In particular, each weightless neural discriminator has been modified so as to generate a non-crisp example of the class of simple visual features it was ...
متن کاملNeurosymbolic Integration: Cognitive Grounds and Computational Strategies
The ultimate|if implicit|goal of artiicial intelligence (AI) research is to model the full range of human cognitive capabilities. Symbolic AI and connectionism, the major AI paradigms, have each tried|and failed|to attain this goal. In the meantime, the idea has gained ground that this goal might still be within reach if we could harness the respective strengths of these two paradigms in integr...
متن کاملNeuroSymbolic Processing: Non-Monotonic Operators and Their FPGA Implementation
Recalling the insights of McCulloch and Pitts [4], von Neumann [9] and Minsky [5] we introduce some non monotonic operators implemented by means of new electronic devices (FPGA) for a NeuroSymbolic Processor (NSP). Some nets performing logical and arithmetic functions are also presented.
متن کاملA Unified Neurosymbolic Model of the Mutual Influence of Memory, Context and Prediction of Time Ordered Sequential Events During the Audition of Tonal Music
Wc describe a hybrid system to model context formulation and resulting expectations created by a listener while attending to tonal music. The model is hybrid in that we use modular sub-networks to simulate the distinct yet mutually influential schemas involved in constructing expectations for sequential events and the temporal cyclical grid that creates metrical support for these expectations. ...
متن کاملInteractive Causation: a Neurosymbolic Agent
Interactive Causation Agent (ICA), a novel unsupervised adaptive neurosymbolic model based on a theory of causality (Interactive Causation), is defined, offering an interactive computational approach to problems in the algorithmic computing of interaction. Interactive Causation provides a compatible framework of nonalgorithmic computation and environment, permitting tractable computation of tem...
متن کامل