Contents lists available at ScienceDirect Biologically Inspired Cognitive Architectures journal homepage: www.elsevier.com/locate/bica Research article Shifting and drifting attention while reading: A case study of nonlinear- dynamical attention allocation in the OpenCog cognitive architecture Misgana Bayetta Belachew a,b,c, ⁎ , Ben Goertzel a,b,c , Matthew Ikle’ d , David Hanson b,c a OpenCog Foundation, Hong Kong b SingularityNET, Hong Kong c Hanson Robotics, Hong Kong d Adams State University, United States ABSTRACT A simple experimental example of the general principle of “cognitive synergy” underlying the OpenCog AGI architecture is explored: An OpenCog system processing a series of articles that shifts from one topic (insects) to another (poisons), and using its nonlinear attention-allocation dynamics (based on the ECAN Economic Attention Networks framework) to spread attention back and forth between the nodes and links within OpenCog’s Atomspace knowledge store representing the words in the sentences, and other nodes and links containing related knowledge. With this setup, we study how the ECAN system shifts the attentional focus of the system based on changes in topic – in terms of both the speed of attention switching, and the contextual similarity of the content of attentional focus to the sentences being processed at a given point in time. This also provides an avenue for exploring the effects of particular design choices within the ECAN system. For instance, we find that in this particular example, if the parameters are set appropriately, ECAN indeed causes the system to assign particular importance to nodes and links related to the “insecticide” concept, when it is reading sentences about poisons in a situation where it has been primed by sentences about insects. This is an example of what we call “drifting” attention – the system’s attention moves to something suggested by its perceptions, even if not directly presented in them. Introduction One approach to creating AGI systems is the “integrative” strategy, involving combining multiple components embodying different struc- tures or algorithms, and relying on synergistic dynamics between components. One kind of integrative system involves various highly independent software components, each solving a specialized set of problems in a mostly standalone manner, with occasional commu- nication between each other in order to exchange problems and solu- tions. On the other end of the scale, are systems designed as tightly interconnected components that give rise to complex non-linear dyna- mical phenomena. Here, we are specifically focused on the latter ap- proach. We will discuss the particulars of one form of cognitive synergy – between natural language processing agents and nonlinear-dynamical attention allocation – within the context of one particular integrative AGI architecture, the OpenCog platform (Goertzel, 2009; Goertzel, Pennachin, & Geisweiller, 2013a; Goertzel, Pennachin, & Geisweiller, 2013b) (and using OpenCog according to a cognitive architecture now referred to as PrimeAGI, previously referred to as OpenCogPrime). The core cognitive processes involved here are: – Natural language processing that reads a sentence and creates a number of nodes and links (Atoms) representing syntactic and se- mantic relationships between the words in the sentence (represented as WordNodes in the OpenCog Atomspace hypergraph knowledge store) and their associated concepts and relationships. – Attention allocation that manages the spreading of STI (Short Term Importance) values among Atoms in the Atomspace. Atoms of initial interest are stimulated with STI, and then Atoms spread STI to other Atoms that they are linked to, according to the nonlinear-dynamical equations of the ECAN module. – The Attentional Focus, defined as the set of top-STI Atoms in the Atomspace, is updated via slightly different dynamics than the rest of the Atomspace; STI values of Atoms in the Attentional Focus are updated more frequently and consistently, and there is an optional process that builds HebbianLinks between pairs of Atoms in the Attentional Focus. The specific dynamics explored and demonstrated here, using these cognitive processes, is as follows: – The OpenCog NLP Pipeline processes (doing syntax parsing and then some light semantic interpretation) a series of articles – first insect related articles and then poison related articles. https://doi.org/10.1016/j.bica.2018.07.005 Received 5 July 2018; Accepted 6 July 2018 ⁎ Corresponding author at: OpenCog Foundation, Hong Kong. Biologically Inspired Cognitive Architectures xxx (xxxx) xxx–xxx 2212-683X/ © 2018 Elsevier B.V. All rights reserved. Please cite this article as: Belachew, M.B., Biologically Inspired Cognitive Architectures (2018), https://doi.org/10.1016/j.bica.2018.07.005