Table 1 - uploaded by Loo-Nin Teow
Content may be subject to copyright.
Source publication
Memory enables past experiences to be remembered and acquired as useful knowledge to support decision making, especially when perception and computational resources are limited. This paper presents a neuropsychological-inspired dual memory model for agents, consisting of an episodic memory that records the agent's experience in real time and a sema...
Context in source publication
Context 1
... direct semantic memory only learn the event whenever the NPC shot hits THEN ROCKET_LAUNCHER effectiveness 0.932 weapon range categorization: extremely near:0-299; very near:300-599;near:600-899;medium near:900-1199; not so near:1200-1499;midrange:1500-1799;not so far:1800-2099; medium far:2100-2399;far:2400-2699;very far:2700-2999; exremely far:3000 or more the opponent. Table 1 illustrates sample learnt rules of weapon effective- ness translated into symbolic forms. Each rule corresponds to a category node in F2 layer of the semantic memory. ...
Similar publications
Any experienced event may be encoded and retained in detail as part of our episodic memory, and may also refer and contribute to our generalized knowledge stored in semantic memory. The beginnings of this declarative memory formation are only poorly understood. Even less is known about the interrelation between episodic and semantic memory during t...
Citations
... Essentially, in the acquisition of information system, forgetting is the natural and inevitable cost of storage. Alternatively, memory consolidation, as a memory management mechanism, strengthens the memory traces and connections and leads to the formation of memory for words and automatic processing (Subagdja et al., 2012;Bakker et al., 2015). These brain processes encompass modifications in cell signaling during the retrieval of the accumulated information across the consolidation period (Davis and Zhong, 2017). ...
Multisensory input is an aid to language comprehension; however, it remains to be seen to what extent various combinations of senses may affect the P200 component and attention-related cognitive processing associated with L2 sentence comprehension along with the N400 as a later component. To this aim, we provided some multisensory input (enriched with data from three (i.e., exvolvement) and five senses (i.e., involvement)) for a list of unfamiliar words to 18 subjects. Subsequently, the words were embedded in an acceptability judgment task with 360 pragmatically correct and incorrect sentences. The task, along with the ERP recording, was conducted after a 1-week consolidation period to track any possible behavioral and electrophysiological distinctions in the retrieval of information with various sense combinations. According to the behavioral results, we found that the combination of five senses leads to more accurate and quicker responses. Based on the electrophysiological results, the combination of five senses induced a larger P200 amplitude compared to the three-sense combination. The implication is that as the sensory weight of the input increases, vocabulary retrieval is facilitated and more attention is directed to the overall comprehension of L2 sentences which leads to more accurate and quicker responses. This finding was not, however, reflected in the neural activity of the N400 component.
... Fusion ART has been used in the realization of a hierarchical planner , as well as of different types of long term memory models: episodic (Leconte, Ferland, & Michaud, 2014, 2016Nasir, Kim, & Kim, 2019;Nasir, Yoo, Kim, & Kim, 2018;Park & Kim, 2016;Park, Yoo, & Kim, 2015;Park, Yoo, Kim, & Kim, 2018;Subagdja & Tan, 2015;Subagdja, Wang, Tan, Tan, & Teow, 2012;Wang, Subagdja, Tan, & Starzyk, 2010, 2012aWang, Tan, & Teow, 2017), semantic (Nasir et al., 2019;Nasir et al., 2018;Wang et al., 2012Wang et al., , 2017 and procedural (Wang et al., , 2017) among a number of other applications. See also Tan, Subagdja, Wang, and Meng (2019) in this issue. ...
This survey samples from the ever-growing family of adaptive resonance theory (ART) neural network models used to perform the three primary machine learning modalities, namely, unsupervised, supervised and reinforcement learning. It comprises a representative list from classic to contemporary ART models, thereby painting a general picture of the architectures developed by researchers over the past 30 years. The learning dynamics of these ART models are briefly described, and their distinctive characteristics such as code representation, long-term memory, and corresponding geometric interpretation are discussed. Useful engineering properties of ART (speed, configurability, explainability, parallelization and hardware implementation) are examined along with current challenges. Finally, a compilation of online software libraries is provided. It is expected that this overview will be helpful to new and seasoned ART researchers.
Learning and memory are two intertwined cognitive functions of the human brain. This paper shows how a family of biologically-inspired self-organizing neural networks, known as fusion Adaptive Resonance Theory (fusion ART), may provide a viable approach to realizing the learning and memory functions. Fusion ART extends the single-channel Adaptive Resonance Theory (ART) model to learn multimodal pattern associative mappings. As a natural extension of ART, various forms of fusion ART have been developed for a myriad of learning paradigms, ranging from unsupervised learning to supervised learning, semi-supervised learning, multimodal learning, reinforcement learning, and sequence learning. In addition, fusion ART models may be used for representing various types of memories, notably episodic memory, semantic memory and procedural memory. In accordance with the notion of embodied intelligence, such neural models thus provide a computational account of how an autonomous agent may learn and adapt in a real-world environment. The efficacy of fusion ART in learning and memory shall be discussed through various examples and illustrative case studies.
Thanks to the spread of digital photography and available devices, taking photographs has become effortless and tolerated nearly everywhere. This makes people easily ending up with hundreds or thousands of photos, for example, when returning from a holiday trip or taking part in ceremonies, concerts, and other events. Furthermore, photos are also taken of more mundane motives, such as food and aspects of everyday life, further increasing the number of photos to be dealt with. The decreased prices of storage devices make dumping the whole set of photos common and affordable. However, this practice frequently makes the stored collections a kind of dark archives, which are rarely accessed and enjoyed again in the future. The big size of the collections makes revisiting them time demanding. This suggests to identify, with the support of automated methods, the sets of most important photos within the whole collections and to invest some preservation effort for keeping them accessible over time. Evaluating the importance of photos to their owners is a complex process, which is often driven by personal attachment, memories behind the content, and personal tastes that are difficult to capture automatically. Therefore, to better understand the selection process for photo preservation and future revisiting, the first part of this chapter presents a user study on a photo selection task where participants selected subsets of most important pictures from their own collections. In the second part of this chapter, we present methods to automatically select important photos from personal collections, in light of the insights emerged from the user study. We model a notion of photo importance driven by user expectations, which represents what photos users perceive as important and would have selected. We present an expectation-oriented method for photo selection, where information at both photo and collection levels is considered to predict the importance of photos.
Building a cyber-physical system consisting of many reactive elements that continually interact with each other may require considerable efforts to ensure that it behaves as desired. The scale of the issues to tackle is much expanded when it comprises human-factors to analyze and change (e.g persuasion). In this paper, a development methodology is proposed for building the kind of system that developer interacts and communicates directly with the components in the domain. The developer interacts with them in simulated or runtime context in order to instruct and shape the whole system. As a technique of Agent-Based Modeling, the entire system is considered to consist of interacting (semi-) autonomous agents that can reason, adapt, and learn at runtime. This approach of development requires a particular agent architecture that can reason and learn about human activity and social conditions comprising spatial, temporal, and hierarchical structure of information. In this paper, the teachable approach is exemplified using a simulated persuasive multi-agent system for elderly care in ageing-in-place domain.
In a knowledge based service system like IT services, the requirements of skills to service customer requests keep changing with time. The service workers are expected to learn the required skills very quickly and become productive. Due to high attrition rate and demand, service workers are given basic class room training and then rest of the training is carried out on-job. When a service worker learns multiple skills simultaneously, learning slows down due to factors like forgetting and interference. At the same time, the organization needs to meet service level agreements (SLA). We have developed a model for on-job training which extends the business process for IT service delivery. The key idea is to model learning, forgetting and interference in service time estimation to get realistic service times. Accurate estimation of service time taken by a service worker to resolve the service tickets helps in resource allocation and planning decisions for achieving the desired objectives of upskilling and SLA success. The simulation of execution of the augmented business process provides insights into what kind of planning and dispatch policies should be practiced for achieving the desired goals of multi-skill learning and SLA success.
Episodic Memory (EM) abilities are important for many types of intelligent virtual agents (IVAs). However, the few IVA–EM systems implemented to date utilize indexed logs of events as the underlying memory representation, which makes it hard to model some crucial facets of human memory, including hierarchical organization of episodes, reconstructive memory retrieval, and encoding of episodes with respect to previously learnt schemata. Here, we present a new general framework for EM modeling, DyBaNeM, which capitalizes on bayesian representation and, consequently, enables modeling these (and other) features easily. By means of a proof-of-concept implementation, we demonstrate that our approach to EM modeling is promising, at least for domains of moderate complexity.
