Women have better recall when it comes to remembering specifics, according to new research. Females apparently have the edge when it comes to remembering features of a conversation or where missing objects might be because they fare better with episodic memory. Episodic memory is the ability to recall autobiographical events such as what happened last week or whether the cat was fed this morning. As one of the most sensitive memory systems it can be impacted by lack of sleep, depression or aging. The research also indicates women are better at remembering faces and recalling sensory memories such as smells.
Reinforcement learning (RL) algorithms have made huge progress in recent years by leveraging the power of deep neural networks (DNN). Despite the success, deep RL algorithms are known to be sample inefficient, often requiring many rounds of interaction with the environments to obtain satisfactory performance. Recently, episodic memory based RL has attracted attention due to its ability to latch on good actions quickly. In this paper, we present a simple yet effective biologically inspired RL algorithm called Episodic Memory Deep Q-Networks (EMDQN), which leverages episodic memory to supervise an agent during training. Experiments show that our proposed method can lead to better sample efficiency and is more likely to find good policies. It only requires 1/5 of the interactions of DQN to achieve many state-of-the-art performances on Atari games, significantly outperforming regular DQN and other episodic memory based RL algorithms.
As countermeasure for preventing dementia of aging population, coimagination method has been developed. The coimagination method helps participants in utilizing brain cognitive functions of maintaining recent episodic memorization, retention and recall by the process of conversations. Hence, the risk of older adults in getting into mild cognitive impairment (MCI), which is a previous stage of dementia caused by disuse of brain cognitive functions, will decline. However, we observed situations of some older adults that recent episodic memory functions were not activated as expected. Such situations are older adults who talk about knowledge rather than episodic memories or older adults who talk about past experiences rather than recent experiences. Therefore, a novel coimagination program named coimagination method with expedition was developed to solve these situations. By adding expedition in a sightseeing area before the coimagination method, older adults have the opportunity to find topic of conversations through expedition. During conversation supported by the coimagination method, older adults are expected to recall their episodic memories in expedition and talk about it. The purpose of this research is to verify the effect of the coimagination method with expedition in older adults, by comparing mental time of older adults in the coimagination methods with and without expedition. Firstly, we estimate the mental time of older adults by analyzing their utterances during conversations supported by both coimagination methods. The past, present and future mental times of participants are enumerated in percentage. Secondly, we study the mental time travelling of participants during conversations. Finally, we study the transition points of mental time to find tendency of participants to talk about recent experiences. In this research, the analytical results validate the effectiveness of helping older adults to talk about recent episodic memories during conversation supported by the coimagination method with expedition compared to the coimagination method.
There is not much evaluation technique of coimagination method, which is one of the group conversation techniques have been proposed for the purpose of cognitive function training. As one of the indicator of usefulness of cognitive function training, episodic memory is usable. Therefore we have proposed an analytical method for measuring the utilization of episodic memory in coimaginaiton method. Thereafter, We conducted the experiment of group conversation base on walking around in order to give the common experience to the participants, and analyzed the results by the proposed method. In consequence, it is revealed the occurrence of past episodic memory. Furthermore, it indicates individual difference of episodic memory utilization quantitatively in terms of memory taxonomy.
We propose an episodic memory-based approach to the problem of pattern capture and recognition. We show how a generic episodic memory module can be enhanced with an incremental retrieval algorithm that can deal with the kind of data available for this application. We evaluate this approach on a goal schema recognition task on a complex and noisy dataset. The memory module was able to achieve the same level of performance as statistical approaches and doing so in a scalable manner.