Reasoning-Learning Systems Based on Non-Axiomatic Reasoning System Theory

Patrick Hammer
Proceedings of the Third International Workshop on Self-Supervised Learning, PMLR 192:89-107, 2022.

Abstract

In this paper a strong motivation for real-time reasoning-learning systems based on Non-Axiomatic Reasoning System (NARS) Theory as an approach to build intelligent systems with agency is given. This contains the requirement to work under the Assumption of Insufficient Knowledge and Resources which demands open-ended adaptation while obeying to strict computational resource restrictions to allow for real-time response. We show how this aligns with the phenomenon of intelligence as found in nature, allowing for systems which can both react instantly, and plan ahead deliberately dependent on implicitly outcome-dependent time pressures. In this context a specific implementation design is considered, OpenNARS for Applications (ONA), and how its learning and reasoning abilities lead to data-efficient adaptation in novel circumstances in various domains, whereby we compare with a reinforcement learning method, Q-Learning, in Space Invaders, Pong and a grid robot environment. We will see that both techniques perform comparably well for reactive tasks in Markovian environments, while the uncertainty reasoner performs better when the Markov property is violated, with the additional property that it can plan ahead to exploit task compositionality, also taking explicit background knowledge into account.

Cite this Paper

BibTeX
@InProceedings{pmlr-v192-hammer22a, title = {Reasoning-Learning Systems Based on {Non-Axiomatic Reasoning System} Theory}, author = {Hammer, Patrick}, booktitle = {Proceedings of the Third International Workshop on Self-Supervised Learning}, pages = {89--107}, year = {2022}, editor = {Thórisson, Kristinn R.}, volume = {192}, series = {Proceedings of Machine Learning Research}, month = {28--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v192/hammer22a/hammer22a.pdf}, url = {https://proceedings.mlr.press/v192/hammer22a.html}, abstract = {In this paper a strong motivation for real-time reasoning-learning systems based on Non-Axiomatic Reasoning System (NARS) Theory as an approach to build intelligent systems with agency is given. This contains the requirement to work under the Assumption of Insufficient Knowledge and Resources which demands open-ended adaptation while obeying to strict computational resource restrictions to allow for real-time response. We show how this aligns with the phenomenon of intelligence as found in nature, allowing for systems which can both react instantly, and plan ahead deliberately dependent on implicitly outcome-dependent time pressures. In this context a specific implementation design is considered, OpenNARS for Applications (ONA), and how its learning and reasoning abilities lead to data-efficient adaptation in novel circumstances in various domains, whereby we compare with a reinforcement learning method, Q-Learning, in Space Invaders, Pong and a grid robot environment. We will see that both techniques perform comparably well for reactive tasks in Markovian environments, while the uncertainty reasoner performs better when the Markov property is violated, with the additional property that it can plan ahead to exploit task compositionality, also taking explicit background knowledge into account.} }
Endnote
%0 Conference Paper %T Reasoning-Learning Systems Based on Non-Axiomatic Reasoning System Theory %A Patrick Hammer %B Proceedings of the Third International Workshop on Self-Supervised Learning %C Proceedings of Machine Learning Research %D 2022 %E Kristinn R. Thórisson %F pmlr-v192-hammer22a %I PMLR %P 89--107 %U https://proceedings.mlr.press/v192/hammer22a.html %V 192 %X In this paper a strong motivation for real-time reasoning-learning systems based on Non-Axiomatic Reasoning System (NARS) Theory as an approach to build intelligent systems with agency is given. This contains the requirement to work under the Assumption of Insufficient Knowledge and Resources which demands open-ended adaptation while obeying to strict computational resource restrictions to allow for real-time response. We show how this aligns with the phenomenon of intelligence as found in nature, allowing for systems which can both react instantly, and plan ahead deliberately dependent on implicitly outcome-dependent time pressures. In this context a specific implementation design is considered, OpenNARS for Applications (ONA), and how its learning and reasoning abilities lead to data-efficient adaptation in novel circumstances in various domains, whereby we compare with a reinforcement learning method, Q-Learning, in Space Invaders, Pong and a grid robot environment. We will see that both techniques perform comparably well for reactive tasks in Markovian environments, while the uncertainty reasoner performs better when the Markov property is violated, with the additional property that it can plan ahead to exploit task compositionality, also taking explicit background knowledge into account.
APA
Hammer, P.. (2022). Reasoning-Learning Systems Based on Non-Axiomatic Reasoning System Theory. Proceedings of the Third International Workshop on Self-Supervised Learning, in Proceedings of Machine Learning Research 192:89-107 Available from https://proceedings.mlr.press/v192/hammer22a.html.

Related Material