Characteristics of Forming Episodic Associations Between Words (2004)
The goal of the present study was to explore factors that influence how new episodic associations between words are formed and to characterize the dynamic properties of this process, Of particular interest was the impact of semantic knowledge on such associations. To achieve this goal, I combined psychological experimental methods and computational techniques in a multidisciplinary research approach. First, I tested human participants in a series of psychological experiments and revealed several characteristics of how new associations between words are learned. Second, I developed a computational model based on a neural network constrained by the empirical results, and previously published knowledge in the field. This model was validated by a computer simulation that replicated the human performance. The model was then used to derive testable new predictions about human associative learning. Third, some of these predictions were tested in additional psychological experiments. [...] Overall, the results of this study enrich our knowledge on the process of forming associations between words. Although this process is episodic by nature, it interacts with the semantic system. These theoretical and empirical findings suggest that although the semantic and associative networks are based on different principles, they are highly intertwined in human memory and interact during learning. On the basis of the strong support that it received from human experimentation, SEMANT can provide a theoretical framework for many additional studies aimed at understanding associate learning in particular and human memory and cognition in general.
PhD Thesis, The Hebrew University of Jerusalem, Israel, 2004.

Yaron Silberman Ph.D. Alumni yarons [at] alice nc huji ac il