Knowledge Tracing: A Survey（个人笔记）

“1 INTRODUCTION”

“technology-enhanced teaching has also posed new challenges, one of which is how to effectively track the learning progress of a student through their online interaction with teaching materials – a challenge known as the Knowledge Tracing (KT) problem [5, 6, 112].”技术上得到增强的教学也面临了新挑战，其中之一是如何利用一位学生对教育材料的网上交互有效地追踪其学习过程——这一问题被称为知识追踪问题

“In a nutshell, KT aims to observe, represent, and quantify a student’s knowledge state” 简而言之，KT旨在观察、表征和量化学生的知识状态

“The most notable differences are the comprehensive coverage of KT models; an important discussion on the differences in representation learning techniques related to key aspects of KT (such as knowledge state, forgetting behaviors, and knowledge components); an in-depth coverage of the KT datasets used by the relevant literature highlighting their similarities and issues; and, an extensive report of the results obtained by different KT models which allows future and accurate comparisons between them.” 最明显的区别是KT模型的全面覆盖；关于KT (如知识状态、遗忘行为、知识成分等)关键方面相关的表征学习技术差异的重要讨论；对相关文献使用的KT数据集进行了深入的覆盖，突出了它们的相似之处和问题；并且，通过不同的KT模型获得的结果的广泛报告，允许他们之间的未来和准确的比较。

“2 CATEGORIZATION OF KNOWLEDGE TRACING MODELS”

“2.1 Traditional Knowledge Tracing Models”

“2.1.1 Bayesian Knowledge Tracing”

“2.1.2 Factor Analysis Models”

“The key idea is to estimate student performance by learning a function, usually a logistic function, based on various factors in a population of students who solve a set of problems.” 其核心思想是通过学习一个函数，通常是一个logistic函数，根据学生群体中解决一组问题的各种因素来估计学生的表现。

• “Item Response Theory (IRT)”

“IRT is built upon the following assumptions: (a) The probability that a student correctly answers an item can be formulated as an item response function based on the parameters of the student and the item; (b) The item response function monotonically increases with respect to the ability θi of a student i; (c) For a student with the ability θi , items are considered conditionally independent.” IRT建立在以下假设之上：( a )学生正确回答一个项目的概率可以根据学生和项目的参数形成项目响应函数；( b )项目反应函数关于学生i的能力θ i单调递增；( c )对于能力为θ i的学生，项目被认为是有条件独立的。

• “Additive Factor Model (AFM)”

“The Additive Factor Model [17], originated from Learning Factors Analysis (LFA) [16], is a logistic regression model under four assumptions: (1) the prior knowledge of students may vary; (2) students learn at the same rate; (3) some skills are more likely to be known than others; and, (4) some skills are easier to be learnt than others. In this model, a difficulty parameter βk and a learning rate parameter γk are assigned for each skill k, respectively.” 加性因素模型( Additive Factor Model ) [ 17 ]起源于学习因素分析( Learning Factor Analysis，LFA ) [ 16 ]，是在四个假设下的逻辑回归模型：( 1 )学生的先验知识可能不同；( 2 )学生以同样的速度学习；( 3 )某些技能比其他技能更容易被人熟知；而且，( 4 )有些技能比其他技能更容易学习。在这个模型中，为每个技能k分别分配一个难度参数β k和一个学习率参数γ k。

• “Performance Factor Analysis (PFA)”

“The key idea of PFA is to discard the parameter θi used in the previous models and instead count for success and unsuccessful attempts separately. PFA has three parameters for each skill, including: (1) βk is for the difficulty of a skill k; (2) γS k is for the effect of learning a skill k after successful attempts; and, (3) γF k is for the effect of learning a skill k after unsuccessful attempts.” (Abdelrahman 等, 2022, p. 10) PFA的核心思想是抛弃以往模型中使用的参数θ i，而是分别计算成功和失败的尝试。PFA对每个技能有3个参数，包括：( 1 ) β k为技能k的难度；( 2 ) γS k为成功尝试后学习一项技能的效果k；并且，( 3 ) γF k是对尝试不成功后学习技能k的影响。

• “Knowledge Tracing Machine (KTM)”

“2.1.3 Discussion”

“Both Bayesian knowledge tracing and factor analysis models have strengths and weaknesses” 贝叶斯知识追踪和因子分析模型各有优缺点

1. Model Parameters

   “For factor analysis models, most of them have been developed with similar or more flexible model parameters than BKT models.” 对于因子分析模型，大多数已经开发出与BKT模型相似或更灵活的模型参数。

2. Model inference

   “BKT models typically assume a first-order Markov chain when making inference based on a sequence of past questionanswering history,” BKT模型在根据过去的问答历史序列进行推理时，通常假设一个一阶马尔可夫链。

   “factor analysis models usually do not explicitly make inferences on knowledge states of a student” (Abdelrahman 等, 2022, p. 12) 因子分析模型通常不会明确地对学生的知识状态做出推断

3. Temporal analysis

   “BKT models essentially deal with a sequence prediction problem based on the history of student learning. In contrast, factor analysis models do not consider the order of questions in which a student’s answers are observed.” (Abdelrahman 等, 2022, p. 12) BKT模型本质上处理的是基于学生学习历史的序列预测问题。相比之下，因子分析模型没有考虑学生回答问题的顺序。

“2.2 Deep Learning Knowledge Tracing Models”

“2.2.1 Sequence Modeling for Knowledge Tracing”

“A knowledge tracing task is typically modeled as a sequence prediction problem from a machine learning perspective” 从机器学习的角度，知识追踪任务通常被建模为序列预测问题

“Deep Knowledge Tracing (DKT) [89] pioneered the use of deep learning for knowledge tracing. It employs Recurrent Neural Network (RNN) [63] and a Long Short Term Memory (LSTM) [46] to predicate the probability of correctly answering a question at each time step.” 深度知识追踪( Deep Knowledge Tracing，DKT ) [ 89 ]开创了利用深度学习进行知识追踪的先河。它使用循环神经网络( Recurrent Neural Network，RNN ) [ 63 ]和长短期记忆网络( Long Short Term Memory，LSTM ) [ 46 ]来预测每个时间步正确回答问题的概率。

“DKT has several limitations”

1. “it assumes only one hidden KC (i.e., a skill) in a student’s knowledge state ht .” 它假设在学生的知识状态ht中只有一个隐藏的KC (也就是说,一种技能)。

2. “it cannot model the relationships among multiple KCs.” 它无法对多个KCs之间的关系进行建模。

3. “it assumes that all questions are equally likely related to each other, which may not hold in many scenarios as some questions may be more relevant to each other than the remaining of the sequence” 它假设所有的问题都有相同的可能性彼此相关，这在许多场景中可能不成立，因为有些问题可能比序列的其余部分更相关

“2.2.2 Memory-Augmented Knowledge Tracing Models”

“To trace complex KCs learned by students, several works have extended DKT by augmenting an external memory structure, inspired by memory-augmented neural networks” 为了追踪学生学习到的复杂KCs，一些工作在记忆增强神经网络的启发下，通过增加外部记忆结构来扩展DKT

“2.2.3 Attentive Knowledge Tracing Models”

“Following the Transformer architecture [110], several works [21, 36, 79, 80, 99] have attempted to incorporate an attention mechanism into KT models.” 继Transformer架构[ 110 ]之后，一些工作[ 21、36、79、80、99]尝试将注意力机制融入到KT模型中。

“2.2.4 Graph-Based Knowledge Tracing Models”

“To capture such relational structures for better addressing the KT problem, a recent trend is to explore the power of graph representation learning techniques such as Graph Neural Network (GNN).” 为了捕获这种关系结构以更好地处理KT问题，最近的一个趋势是探索图表示学习技术的威力，如图神经网络( Graph Neural Network，GNN )。

“2.2.5 Text-Aware Knowledge Tracing Models”

“Text-aware KT models are motivated by leveraging the textual features of questions to enhance the performance in tackling the KT tasks.” 文本感知的KT模型通过利用问题的文本特征来提高处理KT任务的性能。

“2.2.6 Forgetting-Aware Knowledge Tracing Models”

“This is because the knowledge mastery level of a student tends to decline with an exponential rate over time since the last practice of the relevant questions.” 这是因为学生的知识掌握水平从最后一次练习相关题目开始，随着时间的推移，往往会以指数级的速度下降。

“2.2.7 Discussion”

“Below, we discuss several key aspects that are crucial for being considered in their designs.” 下面，我们讨论了在他们的设计中需要考虑的几个关键因素。

• Knowledge State

“One fundamental assumption underlying each deep learning KT model is whether a knowledge state is considered over a single KC or multiple KCs. Accordingly, modeling a student’s knowledge states based on the mastery level of KCs by the student is an important task in designing the deep learning KT models.” 每个深度学习KT模型的一个基本假设是，一个知识状态是考虑在单个KC上还是多个KC上。因此，根据学生对KCs的掌握程度对学生的知识状态进行建模是设计深度学习KT模型的重要任务。

• KC Dependency

“In a KT task, each question is assumed to associate with a single KC or multiple KCs, which is often provided as a prior knowledge such as Q-matrix. One main challenge faced by deep learning KT models is to discover the dependencies among different KCs” 在KT任务中，每个问题被假设与单个KC或多个KC相关联，这通常作为先验知识提供，如Q矩阵。深度学习KT模型面临的一个主要挑战是发现不同KCs之间的依赖关系

• Feature Augmentation

“On one hand, augmenting additional features can usually lead to more accurate prediction on student learning performance; on the other hand, the augmentation of such additional features depends on their availability in databases” 一方面，增加额外的特征通常可以更准确地预测学生的学习表现；另一方面，这些额外特征的增强依赖于它们在数据库中的可用性

“3 KNOWLEDGE TRACING DATASETS”

“4 KNOWLEDGE TRACING APPLICATIONS”

“4.1 Recommender Systems”

“the aim of such online education systems is twofold” 这样的在线教育系统的目的是双重的

• “to estimate the knowledge state of a student using a KT model” 使用KT模型估计学生的知识状态

• “to recommend learning materials conditioned on the knowledge state using a recommendation model” 使用推荐模型推荐基于知识状态的学习材料

“4.2 Learning Provision and Quality Assurance”

“Another potential application area is to provision a learning curriculum (e.g., an ordered list of topics to study) for a specific subject based on the knowledge state of a student.” 另一个潜在的应用领域是根据学生的知识状态为特定学科提供学习课程(例如,要研究的主题的有序列表)。

“4.3 Interactive Learning”

“Interactive education aims at making the learning process more exciting and engaging through delivering knowledge components into a gaming shell.” 交互式教育旨在通过将知识组件交付到游戏外壳中，使学习过程更加令人兴奋和吸引人。

“4.4 Learning to Teach”

“Virtual students, such as intelligent agents which adopt a reinforcement learning setup or machine learning models, can be treated in a similar way as real-life students who are in need to learn a set of skills in different machine learning tasks.” 虚拟学生，例如采用强化学习设置或机器学习模型的智能代理，可以以类似像现实生活中的要在不同的机器学习任务中学习一组技能的学生一样被对待。

“5 KNOWLEDGE TRACING FUTURE RESEARCH DIRECTIONS”

• “Multimodal and informative representation learning & datasets.”多模态和信息丰富的表示学习和数据集。

“some data in the description of a question such as images and mathematical equations that can lead to more informative embedding representations, are overlooked, either by the proposed models or by the available datasets” 问题描述中的一些数据，如图像和数学方程，被所提出的模型或可用的数据集所忽略，这些数据可以做到更多的信息嵌入表示，

“the representations of multimodal and domain-specific data such as mathematical equations and code snippets remain mostly unexplored in the literature resulting in low-informative representation learning for KT models”多模态和特定领域数据的表示，比如说数学等式和代码片段，在大多数文献中都是未经探索的，这导致了KT模型只能进行低信息表征学习。

• “Self-supervised learning in knowledge tracing.” 知识追踪中的自监督学习。

“Supervised Learning still has a major drawback: the need for large, high-quality labeled data for training” 监督学习仍然有一个主要的缺点：需要大量的、高质量的标注数据进行训练

“Such pre-trained models can thus be transferred to a downstream task to train in a supervised learning manner with a limited amount of labeled data.” 因此，这些预训练的模型可以转移到下游任务中，以监督学习的方式用有限的标记数据进行训练。

• “Interactive knowledge tracing” 交互式知识跟踪

“interactive methods, driven by question answering response behavior, to better understand the dynamics of students’ knowledge states are still unexplored.” 由问答反应行为驱动的交互式方法尚未被探索，而这一方法旨在更好地理解学生知识状态的动态变化。

“another potential future work is to  develop optimized question sampling policies to enhance the performance of KT models in, but not limited to, cold start scenarios.” 另一个潜在的未来工作是开发优化问题采样策略，以增强KT模型在不限于冷启动场景中的性能。

• “eXplainable Artificial Intelligence (XAI)” 可解释人工智能( XAI )

“Potential research avenues includes  the development of techniques and methods to understand and explain the prediction process in KT models” 潜在的研究途径包括开发技术和方法来理解和解释KT模型中的预测过程