The Differences Between the Traditional Artificial Intelligence Systems and Human Intelligence

Cambria, E., Poria, S., Bajpai, R., & Schuller, B. (2016, December). SenticNet 4: A semantic resource for sentiment analysis based on conceptual primitives. In Proceedings of COLING 2016, the 26th international conference on computational linguistics: Technical papers (pp. 2666-2677).

The authors of this article highlights the differences between the traditional Artificial Intelligence systems and human intelligence. The researchers argue that human intelligence surpasses AI in a situation where there is need to use common sense knowledge acquired from learning through experience. The article emphasizes that these are the essentials of making an informed decision. As a result, humans can have an easy time to adapt in a situation where AI fails terribly since it cannot apply situation-specific rules and generalization capabilities. Humans are also effective in operating in social situations since they can use common sense. The authors report that the previous versions of SenticNet were designed based on the collection of the above type of knowledge for sentiment analysis. This aspect made the tool to be limited because of its inability to generalize. SenticNet 4 is a newer and improved version where such limitations are resolved by leveraging on the automatic generation of conceptual primitives by dimensionality reduction and hierarchical clustering.  

‍