tags: [知识追踪, 素材]
categories: [知识追踪, 素材]


素材整理,每天进步一点点!

注:英语多使用被动句, 长短句

常见缩写

etc. 全称et cetera,译为:等等(“and so forth; and the others; and other things; and the rest; and so on”)。常用列物品项时
et al. 全称et alia,译为:等人,等等(”and others; and co-workers”),常用在列文献作者时
e.g., 全称exempli gratia,译为:例如(for example 或such as或for instance),如……
i.e., 全称id test,译为:那就是说(that is) 或 换句话说(in other words),即……

替换词

利用:employ、use、utilize、leverage、exploit
呈现/表明:show、present、demonstrate、suggest
表示:represent、indicate、denote
执行:perform、conduct
因此:thus、therefore、hence、consequently、as a result
假设:assume、hypothesize
此外:additionly、further、moreover、in addition、besides
刻画、表征: characterize、profile
考虑:takes into account、
解决:tackle、solve、address、alleviate
处理、应对:cope with、deal with
为此:to this end
有益于:beneficial、conducive
建模/模拟:model、simulate
称为:referred as
领域:domain,area,field
然而:However、Nevertheless
具体来说:Specifically、Detailedly
提取:extract、draw、abstract、distill、derive
属于:pertain to
下文:hereinafter
由于:owing to、due to
值得注意的的是:Notably

短语

poses a demanding challenge, to meet the challenge
benchmark datasets
computer-aided education (CAE) system
fundamental task
it is generally appreciated that ……
As introduced/mentioned above
is referred as
As far as we know/To the best of our knowledge
surpass
step further
As it can be seen
we will detail these sth in what follows.
it is not feasible to do sth
holds the view that
For the ease of explanation
demonstrated superior performance / exhibit excellent interpretability

Experiments

Datasets

To evaluate our model, the experiments are conducted on three widely-used datasets in KT and the detailed statistics are shown in Table 1.

介绍public/benchmark datasets:assist09、assist12、Ednet etc. 页面底部插入脚注,介绍引用的数据集的网址

Introduction

As the education landscape shifts toward distance learning, the online learning systems advance in complexity and capacity.

heterogeneous types of activities

评估是教育的中心任务,因为它涉及到元认知、跟踪技能轨迹、推荐内容、调整辅导策略]和评分。随着在线教育平台的出现,利用用户交互历史数据构建评估模型的需求越来越大。跟踪用户技能的一种方法是知识跟踪(KT),它是根据学生在智能辅导系统(ITS)中的课程交互对他们的知识进行建模的任务。

To evaluate KT’s performance, it is asked to predict the correctness of students’ future exercises with the retrieved knowledge states as Figure 1. Traditionally, KT was regarded as a sequential behavior mining task [12, 25], and therefore various methods established models with the theory of bayesian probability (BKT [4]) and psycho-statistics (IRT [8]), providing excellent interpretability and good performance.

The remainder of this paper was structured/organized as follows. The related work on KT and other data mining areas was introduced in Section 2. The problem define and comprehensive implementation of the proposed CPKT was presented in Section 3. The Experimental setup of CPKT was formally stated in Section 4. Experiments and results analysis was demonstrated in Section 5. Finally, Section 6 concluded this paper and showed our outlook for future work / looked ahead to our future work.

RELATED WORK

知识追踪定义:知识追踪的目的是捕捉学生的知识状态和知识状态转移模式,并进一步用于学生成绩预测、智能课程设计、学生任务结构的解释和发现等任务

Baselines

介绍要对比的模型Baselines

Implementation Details

Other hyperparameters are chosen by grid search.

Ablation Studies

References

参考文献在30个左右