SEMANTIC and episodic memories belong to the category of declarative memories, where the past, consciously experienced events and knowledge are stored and can be recalled or declared [1]. Thus, they are critical components of any cognitive system. The two memories differ in their organization and use. Semantic memory (SM) integrates sensory experiences and is responsible for the creation and recognition of concepts. It is grounded, and as such, integrates unconscious sensory inputs to recognizable objects and ideas. Episodic memory (EM) uses the concepts represented by the SM to record the personal history of events located in time and space. While SM is developed incrementally over time, EM is developed immediately upon consciously recognized events. The two memories are interdependent since episodes can only be consciously experienced after the SM recognizes them, while past episodes help to develop new knowledge stored in the SM, help to build associations between various objects and actions, and help to anticipate future events [2]. Until now, there has been no study of artificial neural networks implementing integrated architectures of semantic and episodic memories. This paper does it. Tulving [3] in his seminal paper defined the original concept of EM and contrasted it with SM. Both memories are essential for the retrieval of past experiences, learning, planning, and anticipation [4], [5]. Tulving [6] reiterated the distinction between EM and SM by taking into account required operations and awareness involved in retrieving information stored in each type of memory. In addition to episodic events containing what, where, and when information, EM accumulates one’s own past existence. This allows one to mentally revisit the past experiences while being aware that the recollection is related to earlier time [7]. In contrast, SM involves classification and recognition of objects without referring to where or when the event regarding them occurred. It refers to an individual’s knowledge and perception of its environment at the present time. SM provides a cognitive interpretation of the perceived objects and events, and is critical to having a conscious experience. EM plays an important role in support of cognitive functions such as: 1) representation of events in the spatio-temporal domain; 2) formation of concepts in SM; and 3) supervision of tasks in the implementation of goals [8]. Research on the hippocampus (the part of the brain where episodic memories are stored) indicates its importance for providing the context of the observed events. It plays a critical role in reinforcement learning, representing relationships between stimuli [9]. Using an SM, we can model some of the associative processes observed in biological nervous systems. An SM can bind together neuronal representations of trained data, so the relations between data could be expressed through connections and changing the sensitivity of neurons that specialize in recognizing various classes of objects. The major role of SM is to link objects that often occur in the same or similar context, and enable them to be recalled associatively [41].