TY - GEN
T1 - Extensions to Generalized Annotated Logic and an Equivalent Neural Architecture
AU - Shakarian, Paulo
AU - Simari, Gerardo I.
N1 - Funding Information:
P.S. is supported by internal funding from the ASU Fulton Schools of Engineering. G.S. is supported by Universidad Na-cional del Sur (UNS) under grant PGI 24/ZN34 and Agencia Nacional de Promoción Científica y Tecnológica under grant PICT-2018-0475(PRH-2014-0007).
Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - While deep neural networks have led to major advances in image recognition, language translation, data mining, and game playing, there are well-known limits to the paradigm such as lack of explainability, difficulty of incorporating prior knowledge, and modularity. Neuro symbolic hybrid systems have recently emerged as a straightforward way to extend deep neural networks by incorporating ideas from symbolic reasoning such as computational logic. In this paper, we propose a list desirable criteria for neuro symbolic systems and examine how some of the existing approaches address these criteria. We then propose an extension to generalized annotated logic that allows for the creation of an equivalent neural architecture comprising an alternate neuro symbolic hybrid. However, unlike previous approaches that rely on continuous optimization for the training process, our framework is designed as a binarized neural network that uses discrete optimization. We provide proofs of correctness and discuss several of the challenges that must be overcome to realize this framework in an implemented system.
AB - While deep neural networks have led to major advances in image recognition, language translation, data mining, and game playing, there are well-known limits to the paradigm such as lack of explainability, difficulty of incorporating prior knowledge, and modularity. Neuro symbolic hybrid systems have recently emerged as a straightforward way to extend deep neural networks by incorporating ideas from symbolic reasoning such as computational logic. In this paper, we propose a list desirable criteria for neuro symbolic systems and examine how some of the existing approaches address these criteria. We then propose an extension to generalized annotated logic that allows for the creation of an equivalent neural architecture comprising an alternate neuro symbolic hybrid. However, unlike previous approaches that rely on continuous optimization for the training process, our framework is designed as a binarized neural network that uses discrete optimization. We provide proofs of correctness and discuss several of the challenges that must be overcome to realize this framework in an implemented system.
KW - Logic programming
KW - Machine learning
KW - Neural networks
UR - http://www.scopus.com/inward/record.url?scp=85143414223&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85143414223&partnerID=8YFLogxK
U2 - 10.1109/TransAI54797.2022.00017
DO - 10.1109/TransAI54797.2022.00017
M3 - Conference contribution
AN - SCOPUS:85143414223
T3 - Proceedings - 2022 4th International Conference on Transdisciplinary AI, TransAI 2022
SP - 63
EP - 70
BT - Proceedings - 2022 4th International Conference on Transdisciplinary AI, TransAI 2022
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 4th International Conference on Transdisciplinary AI, TransAI 2022
Y2 - 20 September 2022 through 22 September 2022
ER -