Data Mining Practical Machine Learning Tools and Techniques

The Architecture Of Cognition

Networks evolve continuously over time with the addition, deletion, and changing of links and nodes. Although many networks contain this type of temporal information, the majority of research in network representation learning has focused on static snapshots of the graph and has largely ignored the temporal dynamics of the network. In this work, we describe a general framework for incorporating temporal information into network embedding methods. The framework gives rise to methods for learning time-respecting embeddings from continuous-time dynamic networks. Overall, the experiments demonstrate the effectiveness of the proposed framework and dynamic network embedding approach as it achieves an average gain of 11.9% across all methods and graphs. The results indicate that modeling temporal dependencies in graphs is important for learning appropriate and meaningful network representations.

Continuous-Time Dynamic Network Embeddings

This survey is an updated and improved version of the previous one published in 2013 in this journal with the title “data mining in education”. It reviews in a comprehensible and very general way how Educational Data Mining and Learning Analytics have been applied over educational data. In the last decade, this research area has evolved enormously and a wide range of related terms are now used in the bibliography such as Academic Analytics, Institutional Analytics, Teaching Analytics, Data‐Driven Education, Data‐Driven Decision‐Making in Education, Big Data in Education, and Educational Data Science. This paper provides the current state of the art by reviewing the main publications, the key milestones, the knowledge discovery cycle, the main educational environments, the specific tools, the free available datasets, the most used methods, the main objectives, and the future trends in this research area.

Educational data mining and learning analytics: An updated survey

Deep learning approaches to anomaly detection have recently improved the state of the art in detection performance on complex datasets such as large collections of images or text. These results have sparked a renewed interest in the anomaly detection problem and led to the introduction of a great variety of new methods. With the emergence of numerous such methods, including approaches based on generative models, one-class classification, and reconstruction, there is a growing need to bring methods of this field into a systematic and unified perspective. In this review we aim to identify the common underlying principles as well as the assumptions that are often made implicitly by various methods. In particular, we draw connections between classic 'shallow' and novel deep approaches and show how this relation might cross-fertilize or extend both directions. We further provide an empirical assessment of major existing methods that is enriched by the use of recent explainability techniques, and present specific worked-through examples together with practical advice. Finally, we outline critical open challenges and identify specific paths for future research in anomaly detection.

A Unifying Review of Deep and Shallow Anomaly Detection

Anomaly detection is an important problem with multiple applications, and thus has been studied for decades in various research domains. In the past decade there has been a growing interest in anomaly detection in data represented as networks, or graphs, largely because of their robust expressiveness and their natural ability to represent complex relationships. Originally, techniques focused on anomaly detection in static graphs, which do not change and are capable of representing only a single snapshot of data. As real-world networks are constantly changing, there has been a shift in focus to dynamic graphs, which evolve over time.

Anomaly detection in dynamic networks: a survey

Reinforcement Learning (RL) is one of the best machine learning approaches for decision making in interactive environments. RL focuses on inducing effective decision making policies with the goal of maximizing the agent's cumulative reward. In this study, we investigated the impact of both immediate and delayed reward functions on RL-induced policies and empirically evaluated the effectiveness of induced policies within an Intelligent Tutoring System called Deep Thought. Moreover, we divided students into Fast and Slow learners based on their incoming competence as measured by their average response time on the initial tutorial level. Our results show that there was a significant interaction effect between the induced policies and the students' incoming competence. More specifically, Fast learners are less sensitive to learning environments in that they can learn equally well regardless of the pedagogical strategies employed by the tutor, but Slow learners benefit significantly more from effective pedagogical strategies than from ineffective ones. In fact, with effective pedagogical strategies the slow learners learned as much as their faster peers, but with ineffective pedagogical strategies the former learned significantly less than the latter.

Reinforcement Learning: the Sooner the Better, or the Later the Better?

Bayesian Knowledge Tracing (BKT) is one of the most widely adopted student-modeling methods. It uses performance (incorrect,correct) to infer student knowledge state (unlearned, learned). However, performance can be noisy and thus we explored another type of observations -- student response time. Furthermore, we proposed Intervention Bayesian Knowledge Tracing (Intervention-BKT) which can incorporate multiple types of instructional interventions into the conventional BKT model. Our results show that for next-step performance predictions, Intervention-BKT is more effective than BKT; whereas to predict students' post-test scores, including student response time would yield better result than using performance alone.

Incorporating Student Response Time and Tutor Instructional Interventions into Student Modeling

We explored a series of feature selection methods for modelbased Reinforcement Learning (RL). More specifically, we explored four common correlation metrics and based on them, we proposed the fifth one named Weighed Information Gain (WIG). While much existing correlation-based feature selection methods mostly explored high correlation by default, we explored two options: High vs. Low. The former selects the next feature that has the highest correlation measure with existing selected ones while the latter selects the one with the lowest correlations. The 10 correlation-based methods were compared against previous feature selection methods for model-based RL across several datasets collected from two vastly different intelligent tutoring systems. Our results showed that the 10 correlation-based methods significantly outperform all other methods across all datasets. Among the five correlation metrics, WIG performed best. Surprisingly, for each of correlation metrics, the low option significantly outperform its high correlation peer and thus it suggests that low correlation-based feature selection methods are more effective for model-based RL than high ones.

/pdf/aim-low-correlation-based-feature-selection-for-model-based-4cx1jsqgpl.pdf

Aim Low: Correlation-Based Feature Selection for Model-Based Reinforcement Learning.

Constrained action-based decision-making is one of the most challenging decision-making problems. It refers to a scenario where an agent takes action in an environment not only to maximize the expected cumulative reward but where it is subject to certain action-based constraints; for example, an upper limit on the total number of certain actions being carried out. In this work, we construct a general data-driven framework called Constrained Action-based Partially Observable Markov Decision Process (CAPOMDP) to induce effective pedagogical policies. Specifically, we induce two types of policies: CAPOMDPLG using learning gain as reward with the goal of improving students' learning performance, and CAPOMDPTime using time as reward for reducing students' time on task. The effectiveness of CAPOMDPLG is compared against a random yet reasonable policy and the effectiveness of CAPOMDPTime is compared against both a Deep Reinforcement Learning induced policy and a random policy. Empirical results show that there is an Aptitude-Treatment Interaction effect: students are split into High vs. Low based on their incoming competence; while no significant difference is found among the High incoming competence groups, for the Low groups, students following CAPOMDPTime indeed spent significantly less time than those using the two baseline policies and students following CAPOMDPLG significantly outperform their peers on both learning gain and learning efficiency.

https://dl.acm.org/doi/pdf/10.1145/3209219.3209232

Shitian Shen

Papers

Anomaly detection in dynamic networks: a survey

Reinforcement Learning: the Sooner the Better, or the Later the Better?

Incorporating Student Response Time and Tutor Instructional Interventions into Student Modeling

Aim Low: Correlation-Based Feature Selection for Model-Based Reinforcement Learning.

Improving Learning & Reducing Time: A Constrained Action-Based Reinforcement Learning Approach