فرایندهای یادگیری و بازیابی اطلاعات هوش مصنوعی را بیش از حافظه کاری پیش بینی می کند

Previous studies have shown that the ability to learn and the ability to retrieve relevant information from long-term memory are closely related to fluid intelligence. However, it remains unclear whether the effect of learning and retrieval processes on intelligence is unique or merely due to the variance shared with working memory. The current study attempted to achieve a relatively purified representation of learning and retrieval processes and to examine whether they predict fluid intelligence beyond working memory. A sample of 220 university students completed a rule-based learning task, the Posner task, two working memory tasks and two fluid intelligence scales. Fixed-links models were used to separate the core processes representing learning and retrieval from the auxiliary processes and to link them with fluid intelligence. Results showed that the learning and retrieval processes contributed significantly to fluid intelligence (r = 0.38 and − 0.35 respectively). More importantly, both learning and retrieval processes were still predictive of fluid intelligence when working memory was controlled for. These results suggest that the ability to learn abstract rules and the efficiency of retrieving information from long-term memory are two essential components underlying fluid intelligence in addition to working memory.