Issues in Temporal and Causal Inference

This paper discusses several key issues in temporal and causal inference in the context of AGI. The main conclusions are: (1) the representation of temporal information should take multiple forms; (2) classical conditioning can be carried out as temporal inference; (3) causal inference can be realized without a predefined causal relation. A central function of intelligence is prediction, the ability for a system to anticipate future situations according to past experience. It is often considered as a form of temporal inference or causal inference. This paper focuses on several key issues in this type of inference, by introducing the approach taken in NARS (Non-Axiomatic Reasoning System), and comparing it with other approaches. NARS is an AGI system designed according to the theory that intelligence is the ability for a system to adapt to the environment while working with insufficient knowledge and resources. The system takes the form of a general-purpose reasoning system, and carri...

• Type: Journal
• Year: 2015
• Where: AGI
• Authors: Pei Wang, Patrick Hammer
• Keywords: AGI 2015, Artificial Intelligence