Dynamic graph representation learning has become essential for analyzing evolving networks in domains such as social network analysis, recommendation systems, and traffic analysis. However, existing continuous-time methods face three key challenges: (1) some methods depend solely on node-specific memory without effectively incorporating information from neighboring nodes, resulting in embedding staleness; (2) most fail to explicitly capture correlations between node neighborhoods, limiting contextual awareness; and (3) many fail to fully capture the structural dynamics of evolving graphs, especially in absence of rich link attributes. To address these limitations, we introduce TAWRMAC—a novel framework that integrates Temporal Anonymous Walks with Restart, Memory Augmentation, and Neighbor Co-occurrence embedding. TAWRMAC enhances embedding stability through a memory-augmented GNN with fixed-time encoding and improves contextual representation by explicitly capturing neighbor correlations. Additionally, its Temporal Anonymous Walks with Restart mechanism distinguishes between nodes exhibiting repetitive interactions and those forming new connections beyond their immediate neighborhood. This approach captures structural dynamics better and supports strong inductive learning. Extensive experiments on multiple benchmark datasets demonstrate that TAWRMAC consistently outperforms state-of-the-art methods in dynamic link prediction and node classification under both transductive and inductive settings across three different negative sampling strategies. By providing stable, generalizable, and context-aware embeddings, TAWRMAC advances the state of the art in continuous-time dynamic graph learning.

Large language models are increasingly used to generate programming practice, yet their cognitive demands for CS1 learners remain underexplored. We investigate student-perceived cognitive load on GPT-generated Python exercises using a custom intelligent tutoring system that captures fine-grained keystrokes and offers in-exercise AI feedback. Nineteen undergraduates completed nine exercises spanning loops, recursion, and classes and objects, yielding 171 submissions and 83,377 keystrokes. We triangulate self-reports with keystroke dynamics, code quality metrics, and knowledge component (KC) mastery. Perceived load largely tracked labeled difficulty and was inversely related to satisfaction. Keystroke features were the strongest correlates of load, with more typing bursts and longer inter-keystroke latencies accompanying higher demand; among code metrics, the SLoC-LLoC differential showed the highest association with demand, aligning higher structural complexity with greater mental effort. Lower mastery of core knowledge components—especially loops plus string and list operations—coincided with higher load. Students rated instant AI feedback helpful for debugging and best practices, though at times redundant or inconsistent with instructions. To estimate load automatically, we trained models on observed traces. An Extra Trees regressor predicted the primary load measure Demanding with RMSE 0.81, and a three-level classifier achieved 86 % accuracy with weighted F1 0.85 on held-out students. These results suggest actionable design levers for LLM-based tutors: calibrate task complexity, scaffold high-strain concepts, and deliver concise, context-aligned feedback while using keystroke- and code-based signals for real-time load monitoring. The code and data are available: https://github.com/DSAatUSU/LLMCognitiveLoadProgramming.

The rapid rise of online social networks underscores the need to understand the heterogeneous strengths of online relationships. Yet, efforts to assess tie strength (TS) are hindered by the lack of ground-truth labels, differing research perspectives, and limited model performance in real-world settings. To address this gap, we introduce BTS, a comprehensive Benchmark for Tie Strength prediction, aiming to establish a standardized foundation for evaluating and advancing TS prediction methodologies. Specifically, our contributions are: TS Pseudo-Label Techniques -- we categorize TS into seven standardized pseudo-labeling techniques based on prior literature; TS Dataset Collection -- we present a representative collection of three social networks and perform data analysis by investigating the class distributions and correlations across the generated pseudo-labels; TS Pseudo-Label Evaluation Framework -- we propose a standardized framework to evaluate the pseudo-label quality from the perspective of tie resilience; Benchmarking -- we evaluate existing tie strength prediction model performance using the BTS dataset collection, exploring the effects of different experiment settings, models, and evaluation criteria on the results. Furthermore, we derive key insights to enhance existing methods and shed light on promising directions for future research in this domain. The BTS dataset collection, along with the curation codes and experimental scripts, is all available at: https://github.com/XueqiC/Awesome-Tie-Strength-Prediction.

Graphs are essential for modeling complex relationships, analyzing networks, and offering versatile representations that capture diverse data structures. Graph Neural Networks (GNNs) excel in processing graph-structured data by leveraging the relational information encoded in graph topology. However, not all types of data possess an explicit graph structure, particularly tabular data, which is ubiquitous in the real world. To enable the use of GNNs for tabular data, it is necessary to convert tabular data into graph-structured data. Existing methods for this conversion often lack a generic, straightforward approach with unrestrictive assumptions that can directly apply GNNs to tabular data for downstream tasks. In this paper, we introduce RF-GNN, a novel method that enhances traditional machine learning approaches by transforming tabular data into graph structures and leveraging GNNs. Our approach calculates the similarity between pairs of samples based on Random Forest (RF) proximities, which measure how often the same pair appears in the same terminal nodes of a tree in a Random Forest. This enables the creation of an adjacency matrix for instances of tabular data, allowing the application of GNNs. Extensive experiments on 36 different datasets demonstrate that RF-GNN consistently outperforms traditional machine learning models and recent methods in terms of weighted F1-score. We conduct additional experiments to evaluate the effectiveness of RF-GNN components and settings. The code is available in https://github.com/DSAatUSU/RF-GNN.

This study thoroughly examines various Generative Pretrained Transformer (GPT) methodologies in sentiment analysis, specifically in the context of Task 4 on the SemEval 2017 dataset. Three primary strategies are employed: 1) prompt engineering using the advanced GPT-3.5 Turbo, 2) fine-tuning GPT models, and 3) an inventive approach to embedding classification. The research yields detailed comparative insights among these strategies and individual GPT models, revealing their unique strengths and potential limitations. Additionally, the study compares these GPT-based methodologies with other current, high-performing models previously used with the same dataset. The results illustrate the significant superiority of the GPT approaches in terms of predictive performance, with more than 22% in the F1-score compared to the state-of-the-art. Further, the paper sheds light on common challenges in sentiment analysis tasks, such as understanding context and detecting sarcasm. It underscores the enhanced capabilities of the GPT models to handle these complexities effectively. These findings highlight the promising potential of GPT models in sentiment analysis, setting the stage for future research in this field. The code can be found at https://github.com/DSAatUSU/SentimentGPT

Online coding platforms (OCPs) often offer a limited selection of exercises, which can restrict the scope of Computer Science (CS) education. This study investigates the capabilities of Large Language Models (LLMs), particularly GPT-4 Turbo, in broadening this scope by autonomously generating Python programming exercises. These exercises are tailored to the CS1 curriculum-an introductory course in computer science. Utilizing curriculum-driven prompt engineering, we developed a dataset of 11,700 exercises, characterized by a variety of cate-gories, types, and difficulty levels. These exercises are distributed across 78 unique topics, which were derived from the CS1 course catalogs of leading universities and supplemented with online educational resources. To evaluate the effectiveness of GPT-4 Turbo in generating CSI Python programming exercises, we conducted a user study involving both students and instruc-tors. The study focused on several metrics: exercise quality, curriculum relevance, understandability, appropriate difficulty level, and the generation of useful hints. Our findings indicate that GPT-4 Turbo can produce high-quality, educationally effective programming exercises at scale, provided that the prompts are systematically crafted. Based on insights from the user study, adjustments to prompt design are recommended to optimize exercise generation. Our research concludes that GPT-4 Turbo can be seamlessly integrated into AI-driven OCPs, offering a scalable, cost and time-effective method to enhance CS edu-cation. This is achieved through targeted prompt engineering and thorough data preprocessing to mitigate inconsistencies. The code is available online: https://github.com/DSAatUSU/GPT_CS1400_Exercise_Generation

Keystroke data in programming reveals intricate patterns that reflect the behavior of programmers. These patterns hold promise for predicting grades and other applications, providing insights into the skills of both proficient and less proficient programmers. Analyzing these patterns can yield tailored feedback for students who need support, enabling effective interventions. Our study utilizes a keystroke dataset from the CS1 (Introduction to Computer Science) course at Utah State University. We developed novel features by combining elements like key presses, timestamps, source locations, and programming terminology, drawing on prior research, our insights, and an analysis of programming behavior. An ensemble-based feature selection method identifies key features, which are then used in hyperparameter optimization and grade prediction with six classification and three regression algorithms. We categorized grades into three levels: Low, Average, and High. Despite challenges such as class imbalance, plagiarism, limited data per assignment, and the ceiling effect, we attained a notable weighted F1 score of 78%. We also introduce an ensemble classification strategy, merging Isolation Forest outlier detection with a refined Random Forest classifier, achieving 80% accuracy on our test set. Additionally, we provide a detailed interpretation of our features, supported by results and a case study of our dataset. This research aims to enhance computer science education at the undergraduate level, focusing on improving its overall quality. Code and data are available https://github.com/DSAatUSU/Student-Coding-Behavior.git.

Online social networks are the breeding grounds for user connections, fostering information exchange, communication, content sharing, and community building. However, the dissolution of these digital relationships, often a less-explored facet, complements the studies of tie formation and maintenance. A comprehensive grasp of these connections, encompassing their inception, unraveling, and the potential foresight of disconnections, offers invaluable insights into network dynamics and the progression of interpersonal bonds. Yet, the investigation of broken ties faces a substantial challenge: the paucity of longitudinal and detailed data. To bridge this gap, this paper curates an expansive dataset, spanning over 120,000 Twitter users tracked across 15 weeks with weekly snapshots. Armed with this dataset, we embark on an extensive exploration of Twitter links, delving into five distinct categories within the Twitter social graph. These categories encompass structural features like centrality, content-related aspects, including post polarity, user profile attributes like verified status, egocentric network elements such as reciprocity, and dense user representations typified by node2vec. Subsequently, we conduct a thorough analysis of these diverse features to unveil meaningful patterns.

In recent years, Massive Open Online Courses (MOOCs) have gained significant traction as a rapidly growing phenomenon in online learning. Unlike traditional classrooms, MOOCs offer a unique opportunity to cater to a diverse audience from different backgrounds and geographical locations. Renowned universities and MOOC-specific providers, such as Coursera, offer MOOC courses on various subjects. Automated assessment tasks like grade and early dropout predictions are necessary due to the high enrollment and limited direct interaction between teachers and learners. However, current automated assessment approaches overlook the structural links between different entities involved in the downstream tasks, such as the students and courses. Our hypothesis suggests that these structural relationships, manifested through an interaction graph, contain valuable information that can enhance the performance of the task at hand. To validate this, we construct a unique knowledge graph for a large MOOC dataset, which will be publicly available to the research community. Furthermore, we utilize graph embedding techniques to extract latent structural information encoded in the interactions between entities in the dataset. These techniques do not require ground truth labels and can be utilized for various tasks. Finally, by combining entity-specific features, behavioral features, and extracted structural features, we enhance the performance of predictive machine learning models in student assignment grade prediction. Our experiments demonstrate that structural features can significantly improve the predictive performance of downstream assessment tasks. The code and data are available in https://github.com/DSAatUSU/MOOPer_grade_prediction

Online social media platforms have become a crucial part of teachers’ professional life. They utilize these platforms for various purposes, e.g., creating professional communities, curating pedagogical resources, and interacting with their students. Hence, recent years have witnessed many studies attempting to inform us about opportunities and challenges of social media usage by PK-12 teachers. Nevertheless, how male and female teachers leverage online social media, particularly P interest, is still being determined. Given the importance of online social media in shaping teaching careers and that teachers’ gender arguably affects educational achievements and learning environments, it is essential to delineate teachers’ online (professional) behavior on social media while considering their genders. To this end, first, we build a large dataset of teachers on P interest with fine-grained information about their online activities. T hen, we perform a thorough data analysis of the online behavior of male and female teachers. In particular, we shed light on similarities and differences in the resource curation process of both gender groups, e.g., the topics and sources of their resources. The novel findings of this paper serve as a valuable reference for many entities concerned with gender-aware teaching practices, such as educational scholars, policymakers, and state-level and national-level government agencies.

The central idea of individual fairness is based on an auspicious yet intuitive assertion: similar individuals should be treated similarly. Nevertheless, the fulfillment of individual fairness is hindered by three major obstacles. First, one needs to determine individuals who should receive similar treatment. Second, seamlessly formulating the notion of individual fairness in an ML learning process is another challenge. Third, effectively evaluating the notion of similar treatment in probabilistic classifiers is another challenge. To overcome these challenges, we propose a novel framework called FairMatch. Our proposed framework offers a new approach to pairing similar and dissimilar individuals using a causal analysis method called propensity score matching. Moreover, we formulate individual fairness as a representation learning problem where we incorporate similar and dissimilar pairs in a triplet-based loss function. Eventually, we devise a novel metric to evaluate individual fairness that captures the notion of similar treatment in probabilistic classifiers in a better way. Experimental results on four real-world datasets verify the superiority of FairMatch to existing solutions where we demonstrate it can deliver fairer decisions without scarifying the predictive performance.

With the emergence of online social media platforms, there has been a surge of teachers/educators turning to these platforms for professional purposes, e.g., supplementing their students’ educational needs. Consequently, teachers in social media have been the subject of many educational studies. Despite the progress in this line of research, one of the major obstacles is the limited number of teachers being investigated. Current studies usually suffice to at most a few hundreds of surveyed teachers while there are thousands of other teachers online. To better understand teachers in online social media and enable modern machine learning approaches to process teacher-related data, we need to identify more teachers. Thus, this paper proposes a framework to automatically identify teachers on Pinterest– an image-based social media platform popular among teachers. We formulate the teacher identification problem as a positive unlabeled learning task where positive samples are a small set of surveyed teachers, and unlabeled samples are their connected users on Pinterest. We perform extensive experiments on a real dataset of teachers on Pinterest and show the effectiveness of our framework. We believe the proposed framework can potentially improve the quality of many research endeavors concerned with studying teachers in social media.

Network alignment, in general, seeks to discover the hidden underlying correspondence between nodes across two (or more) networks when given their network structure. However, most existing network alignment methods have added assumptions of additional constraints to guide the alignment, such as having a set of seed node-node correspondences across the networks or the existence of side-information. Instead, we seek to develop a general unsupervised network alignment algorithm that makes no additional assumptions. Recently, network embedding has proven effective in many network analysis tasks, but embeddings of different networks are not aligned. Thus, we present our Deep Adversarial Network Alignment (DANA) framework that first uses deep adversarial learning to discover complex mappings for aligning the embedding distributions of the two networks. Then, using our learned mapping functions, DANA performs an efficient nearest neighbor node alignment. Furthermore, we present an unsupervised heuristic to perform model selection for DANA. We perform experiments on real world datasets to show the effectiveness of our framework for first aligning the graph embedding distributions and then discovering node alignments that outperform existing methods.

One crucial aspect that yet remains fairly unknown while can inform us about the behavior of deep neural networks is their decision boundaries. Trust can be improved once we understand how and why deep models carve out a particular form of decision boundary and thus make particular decisions. Robustness against adversarial examples is directly related to the decision boundary as adversarial examples are basically 'missed out' by the decision boundary between two classes. Investigating the decision boundary of deep neural networks, nevertheless, faces tremendous challenges. First, how we can generate instances near the decision boundary that are similar to real samples? Second, how we can leverage near decision boundary instances to characterize the behaviour of deep neural networks? Motivated to solve these challenges, we focus on investigating the decision boundary of deep neural network classifiers. In particular, we propose a novel approach to generate instances near decision boundary of pre-trained DNNs and then leverage these instances to characterize the behaviour of deep models.

In recent times we have seen a trend of having the ideologies of the two dominant political parties in the U.S. growing further and further apart. Simultaneously we have entered the age of big data raising enormous interest in computational approaches to solve problems in many domains such as political elections. However, an overlooked problem lies in predicting what happens once our elected officials take office, more specifically, predicting the congressional votes, which are perhaps the most influential decisions being made in the U.S. This, nevertheless, is far from a trivial task, since the congressional system is highly complex and heavily influenced by both ideological and social factors. Thus, dedicated efforts are required to first effectively identify and represent these factors, then furthermore capture the interactions between them. To this end, we proposed a robust end-to-end framework Multi-Factor Congressional Vote Prediction (MFCVP) that defines and encodes features from indicative ideological factors while also extracting novel social features. This allows for a principled expressive representation of the complex system, which ultimately leads to MFCVP making accurate vote predictions. Experimental results on a dataset from the U.S. House of Representatives shows the superiority of MFCVP to several representatives approaches when predicting votes for individual representatives and also the overall outcome of the bill voted on. Finally, we perform a factor analysis to understand the effectiveness and interplay between the different factors.

Social media, e.g. Twitter, has become a widely used medium for the exchange of information, but it has also become a valuable tool for hackers to spread misinformation through compromised accounts. Hence, detecting compromised accounts is a necessary step toward a safe and secure social media environment. Nevertheless, detecting compromised accounts faces several challenges. First, social media activities of users are temporally correlated which plays an important role in compromised account detection. Second, data associated with social media accounts is inherently sparse. Finally, social contagions where multiple accounts become compromised, take advantage of the user connectivity to propagate their attack. Thus how to represent each user's network features for compromised account detection is an additional challenge. To address these challenges, we propose an End-to-End Compromised Account Detection framework (E2ECAD). E2ECAD effectively captures temporal correlations via an LSTM (Long Short-Term Memory) network. Further, it addresses the sparsity problem by defining and employing a user context representation. Meanwhile, informative network-related features are modeled efficiently. To verify the working of the framework, we construct a real-world dataset of compromised accounts on Twitter and conduct extensive experiments. The results of experiments show that E2ECAD outperforms the state of the art compromised account detection algorithms.

Social media has become a valuable tool for hackers to disseminate misleading content through compromised accounts. Detecting compromised accounts, however, is challenging due to the noisy nature of social media posts and the difficulty in acquiring sufficient labeled data that can effectively capture a wide variety of compromised tweets from different types of hackers (spammers, vandals, cybercriminals, revenge hackers, etc). To address these challenges, this proposal presents a multiview learning framework that employs nonlinear autoencoders to learn the feature embedding from multiple views, such as the tweets' content, source, location, and timing information and then projects the embedded features into a common lower-rank feature representation. Suspicious user accounts are detected based on their reconstruction errors in the shared subspace. Our empirical results show the superiority of CADET compared to several existing representative approaches when applied to a realworld Twitter dataset.

There are various real-world applications such as video ads, airport screenings, courtroom trials, and job interviews where deception detection can play a crucial role. Hence, there are immense demands on deception detection in videos. However, videos are inherently complex; moreover, they lack detective labels in many real-world applications, which poses tremendous challenges to traditional deception detection methods. In this paper, we study the problem of deception detection in videos. In particular, we provide a principled way to capture rich information into a coherent model and propose an end-to-end framework DEV to detect DEceptive Videos automatically, which is robust to the small number of training data. Experimental results on real-world videos demonstrate the effectiveness of the proposed framework and further experiments are conducted to understand important factors of deception detection in videos.