Description of the Research Plan Title: A novel situated social-personalized learning approach Keywords: e-learning, personalized learning, knowledge space construction, situatedknowledge representation, learning path analysis, cognitive learning process, ande-learning assistant agent Objectives: This research program is to develop an intelligent E-LearningAssistant (ELA) agent for personalized learning. A novel situated social-personalizedlearning approach is proposed in this proposal. This research plan covers threeaspects: The first one is to develop a situated knowledge representation andorganization (SKRO) model. The second is to research nonlinear personallearning path or behavior model. The third is to research social-personalizedlearning mechanism based on machine learning algorithms. The details of thesethree aspects are as follows: 1. Research on situated knowledge representation and organization model Knowledge representationand organization method is the fundamental element for building an e-learningsystem. Here, enlightened by the belief of cognitive learning and pedagogy, i.e.“knowledge is situated and learning is also situated”. A situated knowledgerepresentation and organization (SKRO) model is proposed for building asituated e-learning environment. The SKRO model not only can construct acontextual relationship of knowledge, but also can build a map of differentkinds of heterogeneous knowledge (e.g. text, audio, video, animation, imagesand others). The SKRO model is also a modeling foundation of personal learningprocess and how they learn. 2. Research on nonlinear personal learning path or behavior model With the SKRO model, the concept ofconstructive memory is proposed for storing learning content about thelearners’ situations, goals, and learned knowledge. More important, all theselearned knowledge will be recorded according to a time series and a specificsequence of learning activities. An individual learning path (actually shouldbe a map, not a path) is modeled here to record his learning process andcognitive process. At last, the personal learning space (PLS) is constructedwith personal learned knowledge and nonlinear learning path/behavior. 3. The social-personalized learning mechanism based on machine learningalgorithms The thirdresearch highlight is that we hold a belief that one’s learning process andcognitive process may be helpful to others. Therefore, a novelsocial-personalized learning mechanism based on machine learning algorithms isproposed for personalized learning. The social-personalized learning mechanismmeans that intelligent ELA agent can sense social learning processes (i.e. eachlearner’s sequence of learning activities in any order) to analyze therelevance of knowledge. After that, it constructs a situated knowledgerelevance model through employing statistic-based machine learning algorithms. Sucha model can tell one learner what others’ learning process are and how theylearn through analyzing personal and social learning path. At the same time, thesocial-personalized learning mechanism can also tell the learner what knowledgeI should learn now according to his situation and goals. The framework of intelligent E-Learning Assistant(ELA) agent is as following: