Is Neuro-Symbolic AI Meeting its Promise in Natural Language Processing? A Structured Review

Abstract

Advocates for Neuro-Symbolic AI (NeSy) assert that combining deep learning with symbolic reasoning will lead to stronger AI than either paradigm on its own. As successful as deep learning has been, it is generally accepted that even our best deep learning systems are not very good at abstract reasoning. And since reasoning is inextricably linked to language, it makes intuitive sense that Natural Language Processing (NLP), would be a particularly well-suited candidate for NeSy. We conduct a structured review of studies implementing NeSy for NLP, challenges and future directions, and aim to answer the question of whether NeSy is indeed meeting its promises: reasoning, out-of-distribution generalization, interpretability, learning and reasoning from small data, and transferability to new domains. We examine the impact of knowledge representation, such as rules and semantic networks, language structure and relational structure, and whether implicit or explicit reasoning contributes to higher promise scores. We find that knowledge encoded in relational structures and explicit reasoning tend to lead to more NeSy goals being satisfied. We also advocate for a more methodical approach to the application of theories of reasoning, which we hope can reduce some of the friction between the symbolic and sub-symbolic schools of AI.

Sentiment analysis in SemEval: a review of sentiment identification approaches

“…Sentiment analysis focuses on various tasks including sentiment classification, aspect extraction, and topic modeling. It has evolved from a simple classification task to a problem Int J Elec & Comp Eng ISSN: 2088-8708 ❒ 3323 of automatic opinion detection and opinion-related attributes identification [3] profiting from advancements in neural networks and other emerging research trends such as pre-trained language models (PLM) [4] and neuro-symbolic artificial intelligence (AI) [5]. Currently, sentiment analysis systems are widely used in the social media context and have many applications in other domains such as healthcare, social sciences, market research, and political science [6].…”

Semantic-Aware Dynamic Retrospective-Prospective Reasoning for Event-Level Video Question Answering

“…The model uses SRL information to learn an explicit semantic connection between the text-based questions and visual information in videos. Additionally, we carry out a multi-step reasoning mechanism over video frames to avoid adapting to spurious correlation and shortcuts by explicitly learning the reasoning process itself (Yi et al, 2018;Picco et al, 2021;Hamilton et al, 2022;Zhu, 2022).…”