Open-ended tasks, such as coding problems that are common in computer science education, provide detailed insights into student knowledge.However, training large language models (LLMs) to simulate and predict possible student errors in their responses to these problems can be challenging: they often suffer from mode collapse and fail to fully capture the diversity in syntax, style, and solution approach in student responses.In this work, we present KASER (Knowledge-Aligned Student Error Simulator), a novel approach that aligns errors with student knowledge.We propose a training method based on reinforcement learning using a hybrid reward that reflects three aspects of student code prediction: i) code similarity to the ground-truth, ii) error matching, and iii) code prediction diversity.On two realworld datasets, we perform two levels of evaluation and show that: At the per-student-problem pair level, our method outperforms baselines on code and error prediction; at the per-problem level, our method outperforms baselines on error coverage and simulated code diversity.

Advances in large language models (LLMs) enable many new innovations in education.However, evaluating the effectiveness of new technology requires real students, which is timeconsuming and hard to scale up.Therefore, many recent works on LLM-powered tutoring solutions have used simulated students for both training and evaluation, often via simple prompting.Surprisingly, little work has been done to ensure or even measure the quality of simulated students.In this work, we formally define the student simulation task, propose a set of evaluation metrics that span linguistic, behavioral, and cognitive aspects, and benchmark a wide range of student simulation methods on these metrics.We experiment on a realworld math tutoring dialogue dataset, where both automated and human evaluation results show that prompting strategies for student simulation perform poorly; supervised fine-tuning and preference optimization yield much better but still limited performance, motivating future work on this challenging task. 1

The widespread availability of large language models (LLMs), such as ChatGPT, has significantly impacted education, raising both opportunities and challenges. Students can frequently interact with LLM-powered, interactive learning tools, but their usage patterns need to be observed and understood. We introduce StudyChat, a publicly available dataset capturing real-world student interactions with an LLM-powered tutoring chatbot in a semester-long, university-level artificial intelligence (AI) course. We deploy a web application that replicates ChatGPT’s core functionalities, and use it to log student interactions with the LLM while working on programming assignments. We collect 16,851 interactions, which we annotate using a dialogue act labeling schema inspired by observed interaction patterns and prior research. We analyze these interactions, highlight usage trends, and analyze how specific student behavior correlates with their course outcome. We find that students who prompt LLMs for conceptual understanding and coding help tend to perform better on assignments and exams. Moreover, students who use LLMs to write reports and circumvent assignment learning objectives have lower outcomes on exams than others. StudyChat serves as a shared resource to facilitate further research on the evolving role of LLMs in education.

Vocabulary acquisition poses a significant challenge for second-language (L2) learners, especially when learning typologically distant languages such as English and Korean, where phonological and structural mismatches complicate vocabulary learning.Recently, large language models (LLMs) have been used to generate keyword mnemonics by leveraging similar keywords from a learner's first language (L1) to aid in acquiring L2 vocabulary.However, most methods still rely on direct IPA-based phonetic matching or employ LLMs without phonological guidance.In this paper, we present PHONI-TALE, a novel cross-lingual mnemonic generation system that performs IPA-based phonological adaptation and syllable-aware alignment to retrieve L1 keyword sequence and uses LLMs to generate verbal cues.We evaluate PHONI-TALE through automated metrics and a shortterm recall test with human participants, comparing its output to human-written and prior automated mnemonics.Our findings show that PHONITALE consistently outperforms previous automated approaches and achieves quality comparable to human-written mnemonics.

Learning Japanese vocabulary is a challenge for learners from Roman alphabet backgrounds due to script differences.Japanese combines syllabaries like hiragana with kanji, which are logographic characters of Chinese origin.Kanji are also complicated due to their complexity and volume.Keyword mnemonics are a common strategy to aid memorization, often using the compositional structure of kanji to form vivid associations.Despite recent efforts to use large language models (LLMs) to assist learners, existing methods for LLM-based keyword mnemonic generation function as a black box, offering limited interpretability.We propose a generative framework that explicitly models the mnemonic construction process as driven by a set of common rules, and learn them using a novel Expectation-Maximization-type algorithm.Trained on learner-authored mnemonics from an online platform, our method learns latent structures and compositional rules, enabling interpretable and systematic mnemonics generation.Experiments show that our method performs well in the cold-start setting for new learners while providing insight into the mechanisms behind effective mnemonic creation.

Item (question) difficulties play a crucial role in educational assessments, enabling accurate and efficient assessment of student abilities and personalization to maximize learning outcomes.Traditionally, estimating item difficulties can be costly, requiring real students to respond to items, followed by fitting an item response theory (IRT) model to get difficulty estimates.This approach cannot be applied to the coldstart setting for previously unseen items either.In this work, we present SMART (Simulated Students Aligned with IRT), a novel method for aligning simulated students with instructed ability, which can then be used in simulations to predict the difficulty of open-ended items.We achieve this alignment using direct preference optimization (DPO), where we form preference pairs based on how likely responses are under a ground-truth IRT model.We perform a simulation by generating thousands of responses, evaluating them with a large language model (LLM)-based scoring model, and fit the resulting data to an IRT model to obtain item difficulty estimates.Through extensive experiments on two real-world student response datasets, we show that SMART outperforms other item difficulty prediction methods by leveraging its improved ability alignment.

Open-ended coding tasks, which ask students to construct programs according to certain specifications, are common in computer science education. Student modeling can be challenging since their open-ended nature means that student code can be diverse. Traditional knowledge tracing (KT) models that only analyze response correctness may not fully capture nuances in student knowledge from student code. In this paper, we introduce Test case-Informed Knowledge Tracing for Open-ended Coding (TIKTOC), a framework to simultaneously analyze and predict both open-ended student code and whether the code passes each test case. We augment the existing CodeWorkout dataset with the test cases used for a subset of the open-ended coding questions, and propose a multitask learning KT method to simultaneously analyze and predict 1) whether a student’s code submission passes each test case and 2) the student’s open-ended code, using a large language model as the backbone. We quantitatively show that these methods outperform existing KT methods for coding that only use the overall score a code submission receives. We also qualitatively demonstrate how test case information, combined with open-ended code, helps us gain fine-grained insights into student knowledge.

Automated teaching assistants and chatbots have significant potential to reduce the workload of human instructors, especially for logistics-related question answering, which is important to students yet repetitive for instructors.However, due to privacy concerns, there is a lack of publicly available datasets.We introduce SYLLABUSQA 1 , an open-source dataset with 63 real course syllabi covering 36 majors, containing 5, 078 open-ended course logisticsrelated question-answer pairs that are diverse in both question types and answer formats.Since many logistics-related questions contain critical information like the date of an exam, it is important to evaluate the factuality of answers.We benchmark several strong baselines on this task, from large language model prompting to retrieval-augmented generation.We introduce Fact-QA, an LLM-based (GPT-4) evaluation metric to evaluate the factuality of predicted answers.We find that despite performing close to humans on traditional metrics of textual similarity, there remains a significant gap between automated approaches and humans in terms of fact precision.1

High-quality distractors are crucial to both the assessment and pedagogical value of multiplechoice questions (MCQs), where manually crafting ones that anticipate knowledge deficiencies or misconceptions among real students is difficult.Meanwhile, automated distractor generation, even with the help of large language models (LLMs), remains challenging for subjects like math.It is crucial to not only identify plausible distractors but also understand the error behind them.In this paper, we introduce DiVERT (Distractor Generation with Variational Errors Represented as Text), a novel variational approach that learns an interpretable representation of errors behind distractors in math MCQs.Through experiments on a real-world math MCQ dataset with 1, 434 questions used by hundreds of thousands of students, we show that DiVERT, despite using a base open-source LLM with 7B parameters, outperforms state-of-the-art approaches using GPT-4o on downstream distractor generation.We also conduct a human evaluation with math educators and find that DiVERT leads to error labels that are of comparable quality to humanauthored ones.

Feature acquisition in predictive modeling is an important task in many practical applications. For example, in patient health prediction, we do not fully observe their personal features and need to dynamically select features to acquire. Our goal is to acquire a small subset of features that maximize prediction performance. Recently, some works reformulated feature acquisition as a Markov decision process and applied reinforcement learning (RL) algorithms, where the reward reflects both prediction performance and feature acquisition cost. However, RL algorithms only use zeroth-order information on the reward, which leads to slow empirical convergence, especially when there are many actions (number of features) to consider. For predictive modeling, it is possible to use first-order information on the reward, i.e., gradients, since we are often given an already collected dataset. Therefore, we propose differentiable feature acquisition (DiFA), which uses a differentiable representation of the feature selection policy to enable gradients to flow from the prediction loss to the policy parameters. We conduct extensive experiments on various real-world datasets and show that DiFA significantly outperforms existing feature acquisition methods when the number of features is large.

Mathematical language in scientific communications and educational scenarios is important yet relatively understudied compared to natural languages.Recent works on mathematical language focus either on representing stand-alone mathematical expressions, especially in their natural tree format, or mathematical reasoning in pre-trained natural language models.Existing works on jointly modeling and generating natural and mathematical languages simply treat mathematical expressions as text, without accounting for the rigid structural properties of mathematical expressions.In this paper, we propose a series of modifications to existing language models to jointly represent and generate text and math: representing mathematical expressions as sequences of node tokens in their operator tree format, using math symbol and tree position embeddings to preserve the semantic and structural properties of mathematical expressions, and using a constrained decoding method to generate mathematically valid expressions.We ground our modifications in GPT-2, resulting in a model MathGPT, and demonstrate that it outperforms baselines on mathematical expression generation tasks.

Solutions to math word problems (MWPs) with step-by-step explanations are valuable, especially in education, to help students better comprehend problem-solving strategies.Most existing approaches only focus on obtaining the final correct answer.A few recent approaches leverage intermediate solution steps to improve final answer correctness but often cannot generate coherent steps with a clear solution strategy.Contrary to existing work, we focus on improving the correctness and coherence of the intermediate solutions steps.We propose a step-by-step planning approach for intermediate solution generation, which strategically plans the generation of the next solution step based on the MWP and the previous solution steps.Our approach first plans the next step by predicting the necessary math operation needed to proceed, given history steps, then generates the next step, token-by-token, by prompting a language model with the predicted math operation.Experiments on the GSM8K dataset demonstrate that our approach improves the accuracy and interpretability of the solution on both automatic metrics and human evaluation.

In sequential recommender system applications, it is important to develop models that can capture users' evolving interest over time to successfully recommend future items that they are likely to interact with. For users with long histories, typical models based on recurrent neural networks tend to forget important items in the distant past. Recent works have shown that storing a small sketch of past items can improve sequential recommendation tasks. However, these works all rely on static sketching policies, i.e., heuristics to select items to keep in the sketch, which are not necessarily optimal and cannot improve over time with more training data. In this paper, we propose a differentiable policy for sketching (DiPS), a framework that learns a data-driven sketching policy in an end-to-end manner together with the recommender system model to explicitly maximize recommendation quality in the future. We also propose an approximate estimator of the gradient for optimizing the sketching algorithm parameters that is computationally efficient. We verify the effectiveness of DiPS on real-world datasets under various practical settings and show that it requires up to 50% fewer sketch items to reach the same predictive quality than existing sketching policies.

Traditional learning-based approaches to student modeling generalize poorly to underrepresented student groups due to biases in data availability. In this paper, we propose a methodology for predicting student performance from their online learning activities that optimizes inference accuracy over different demographic groups such as race and gender. Building upon recent foundations in federated learning, in our approach, personalized models for individual student subgroups are derived from a global model aggregated across all student models via meta-gradient updates that account for subgroup heterogeneity. To learn better representations of student activity, we augment our approach with a self-supervised behavioral pretraining methodology that leverages multiple modalities of student behavior (e.g., visits to lecture videos and participation on forums), and include a neural network attention mechanism in the model aggregation stage. Through experiments on three real-world datasets from online courses, we demonstrate that our approach obtains substantial improvements over existing student modeling baselines in predicting student learning outcomes for all subgroups. Visual analysis of the resulting student embeddings confirm that our personalization methodology indeed identifies different activity patterns within different subgroups, consistent with its stronger inference ability compared with the baselines.

In education applications, knowledge tracing refers to the problem of estimating students' time-varying concept/skill mastery level from their past responses to questions and predicting their future performance.One key limitation of most existing knowledge tracing methods is that they treat student responses to questions as binary-valued, i.e., whether they are correct or incorrect.Response correctness analysis/prediction ignores important information on student knowledge contained in the exact content of the responses, especially for open-ended questions.In this paper, we conduct the first exploration into open-ended knowledge tracing (OKT) by studying the new task of predicting students' exact open-ended responses to questions.Our work is grounded in the domain of computer science education with programming questions.We develop an initial solution to the OKT problem, a student knowledge-guided code generation approach, that combines program synthesis methods using language models with student knowledge tracing methods.We also conduct a series of quantitative and qualitative experiments on a real-world student code dataset to validate OKT and demonstrate its promise in educational applications.

Exploiting the ever-growing corpus of scientific content calls for new ways and means to effectively organize, search, and retrieve scientific formulae. We propose a new data-driven framework for retrieving similar scientific formulae via learned formula representations based on tree embeddings. FORTE (for FOrmula Representation learning via Tree Embeddings) leverages operator tree representations of symbolic scientific formulae (such as math equations) to explicitly capture their inherent structural and semantic properties. FORTE employs i) a tree encoder that encodes the formula’s operator tree into an embedding vector and ii) a tree decoder that directly generates a formula’s operator tree from the embedding vector. We also develop a novel tree beam search algorithm that improves the quality of the decoded operator trees. We demonstrate that FORTE (sometimes significantly) outperforms various baseline methods on formula reconstruction and retrieval using a real-world dataset comprising 770k scientific formulae collected on-line.

We study the problem of predicting student knowledge acquisition in online courses from clickstream behavior. Motivated by the proliferation of eLearning lecture delivery, we specifically focus on student in-video activity in lectures videos, which consist of content and in-video quizzes. Our methodology for predicting in-video quiz performance is based on three key ideas we develop. First, we model students’ clicking behavior via time-series learning architectures operating on raw event data, rather than defining hand-crafted features as in existing approaches that may lose important information embedded within the click sequences. Second, we develop a self-supervised clickstream pre-training to learn informative representations of clickstream events that can initialize the prediction model effectively. Third, we propose a clustering guided meta-learning-based training that optimizes the prediction model to exploit clusters of frequent patterns in student clickstream sequences. Through experiments on three real-world datasets, we demonstrate that our method obtains substantial improvements over two base-line models in predicting students’ in-video quiz performance. Further, we validate the importance of the pre-training and meta-learning components of our framework through ablation studies. Finally, we show how our methodology reveals insights on video-watching behavior associated with knowledge acquisition for useful learning analytics.

We study the problem of generating arithmetic math word problems (MWPs) given a math equation that specifies the mathematical computation and a context that specifies the problem scenario.Existing approaches are prone to generating MWPs that are either mathematically invalid or have unsatisfactory language quality.They also either ignore the context or require manual specification of a problem template, which compromises the diversity of the generated MWPs.In this paper, we develop a novel MWP generation approach that leverages i) pre-trained language models and a context keyword selection model to improve the language quality of the generated MWPs and ii) an equation consistency constraint for math equations to improve the mathematical validity of the generated MWPs.Extensive quantitative and qualitative experiments on three realworld MWP datasets demonstrate the superior performance of our approach compared to various baselines.

Computerized adaptive testing (CAT) refers to a form of tests that are personalized to every student/test taker. CAT methods adaptively select the next most informative question/item for each student given their responses to previous questions, effectively reducing test length. Existing CAT methods use item response theory (IRT) models to relate student ability to their responses to questions and static question selection algorithms designed to reduce the ability estimation error as quickly as possible; therefore, these algorithms cannot improve by learning from large-scale student response data. In this paper, we propose BOBCAT, a Bilevel Optimization-Based framework for CAT to directly learn a data-driven question selection algorithm from training data. BOBCAT is agnostic to the underlying student response model and is computationally efficient during the adaptive testing process. Through extensive experiments on five real-world student response datasets, we show that BOBCAT outperforms existing CAT methods (sometimes significantly) at reducing test length.

Despite the abundance of data generated from students’ activities in virtual learning environments, the use of supervised machine learning in learning analytics is limited by the availability of labeled data, which can be difficult to collect for complex educational constructs. In a previous study, a subfield of machine learning called Active Learning (AL) was explored to improve the data labeling efficiency. AL trains a model and uses it, in parallel, to choose the next data sample to get labeled from a human expert. Due to the complexity of educational constructs and data, AL has suffered from the cold-start problem where the model does not have access to sufficient data yet to choose the best next sample to learn from. In this paper, we explore the use of past data to warm start the AL training process. We also critically examine the implications of differing contexts (urbanicity) in which the past data was collected. To this end, we use authentic affect labels collected through human observations in middle school mathematics classrooms to simulate the development of AL-based detectors of engaged concentration. We experiment with two AL methods (uncertainty sampling, L-MMSE) and random sampling for data selection. Our results suggest that using past data to warm start AL training could be effective for some methods based on the target population's urbanicity. We provide recommendations on the data selection method and the quantity of past data to use when warm starting AL training in the urban and suburban schools.

To adapt materials for an individual learner, intelligent tutoring systems must estimate their knowledge or abilities. Depending on the content taught by the tutor, there have historically been different approaches to student modeling. Unlike common skill-based models used by math and science tutors, second language acquisition (SLA) tutors use memory-based models since there are many tasks involving memorization and retrieval, such as learning the meaning of a word in a second language. Based on estimated memory strengths provided by these memory-based models, SLA tutors are able to identify the optimal timing and content of retrieval practices for each learner to improve retention. In this work, we seek to determine whether skill-based models can be combined with memory-based models to improve student modeling and especially retrieval practice performance for SLA. In order to define skills in the context of SLA, we develop methods that can automatically extract multiple types of linguistic features from words. Using these features as skills, we apply skill-based models to a real-world SLA dataset. Our main findings are as follows. First, incorporating lexical features to represent individual words as skills in skill-based models outperforms existing memory-based models in terms of recall probability prediction. Second, incorporating additional morphological and syntactic features of each word via multiple-skill tagging of each word further improves the skill-based models. Third, incorporating semantic features, like word embeddings, to model similarities between words in a learner’s practice history and their effects on memory also improves the models and appears to be a promising direction for future research.

Learning with labels noise has gained significant traction recently due to the sensitivity of deep neural networks under label noise under common loss functions. Losses that are theoretically robust to label noise, however, often makes training difficult. Consequently, several recently proposed methods, such as Meta-Weight-Net (MW-Net), use a small number of unbiased, clean samples to learn a weighting function that downweights samples that are likely to have corrupted labels under the meta-learning framework. However, obtaining such a set of clean samples is not always feasible in practice. In this paper, we analytically show that one can easily train MW-Net without access to clean samples simply by using a loss function that is robust to label noise, such as mean absolute error, as the meta objective to train the weighting network. We experimentally show that our method beats all existing methods that do not use clean samples and performs on-par with methods that use gold samples on benchmark datasets across various noise types and noise rates.

Conventional radio-frequency (RF) sensing systems rely on either frequency diversity or spatial diversity to ensure high sensing accuracy. Such reliance introduces several practical limitations that hinder the pervasive deployment of existing solutions. To circumvent this prevalent reliance, we present MetaSense, a system that leverages antenna pattern diversity for fine-grained RF sensing. MetaSense incorporates the dynamic metasurface antenna (DMA) and the auxiliary-assisted ensemble multimask learning (AEMML) framework in its design. The DMA is a novel type of antenna that can provide a diverse set of uncorrelated radiation patterns in a low-cost and low-complexity manner. The AEMML is a quality-aware learning framework that can dynamically assess and aggregate the heterogeneous channel measurements from different antenna patterns to ensure high sensing accuracy. It also incorporates a transfer learning model that allows it to generalize to new sensing conditions with few training instances required. We prototype MetaSense and demonstrate its effectiveness on a writing motion recognition task using a custom-designed 2-D DMA. The results show that MetaSense achieves 92% to 98% accuracy in classifying ten miniature writing motions, outperforming a nontunable antenna by 20% in all scenarios. Moreover, when deployed in new sensing positions where limited training instances are available, MetaSense requires as few as five training instances per class to achieve over 90% accuracy.

The development of new technologies at an unprecedented rate is rapidly changing the landscape of the labor market. Therefore, for workers who want to build a successful career, acquiring new skills required by new jobs through lifelong learning is crucial. In this paper, we propose a novel and interpretable monotonic nonlinear state-space model to analyze online user professional profiles and provide actionable feedback and recommendations to users on how they can reach their career goals. Specifically, we use a series of binary-valued and non-decreasing latent states to represent the expanding skill set of each user throughout their career and propose an efficient inference method under our model. Using a series of experiments on two large real-world datasets, we show that our model (sometimes significantly) outperforms existing methods on the tasks of company, job title, and skill prediction. More importantly, our model is interpretable and can be used for other important tasks including skill gap identification and career path planning. Using a series of case studies, we show that our model can provide i) actionable feedback to users and guide them through their upskilling and reskilling processes and ii) recommendations of feasible paths for users to reach their career goals.

Knowledge tracing (KT) refers to the problem of predicting future learner performance given their past performance in educational applications. Recent developments in KT using flexible deep neural network-based models excel at this task. However, these models often offer limited interpretability, thus making them insufficient for personalized learning, which requires using interpretable feedback and actionable recommendations to help learners achieve better learning outcomes. In this paper, we propose attentive knowledge tracing (AKT), which couples flexible attention-based neural network models with a series of novel, interpretable model components inspired by cognitive and psychometric models. AKT uses a novel monotonic attention mechanism that relates a learner's future responses to assessment questions to their past responses; attention weights are computed using exponential decay and a context-aware relative distance measure, in addition to the similarity between questions. Moreover, we use the Rasch model to regularize the concept and question embeddings; these embeddings are able to capture individual differences among questions on the same concept without using an excessive number of parameters. We conduct experiments on several real-world benchmark datasets and show that AKT outperforms existing KT methods (by up to $6%$ in AUC in some cases) on predicting future learner responses. We also conduct several case studies and show that AKT exhibits excellent interpretability and thus has potential for automated feedback and personalization in real-world educational settings.

The 2016 United States presidential election has been characterized as a period of extreme divisiveness that was exacerbated on social media by the influence of fake news, trolls, and social bots. However, the extent to which the public became more polarized in response to these influences over the course of the election is not well understood. In this paper we propose IdeoTrace, a framework for (i) jointly estimating the ideology of social media users and news websites and (ii) tracing changes in user ideology over time. We apply this framework to the last two months of the election period for a group of 47508 Twitter users and demonstrate that both liberal and conservative users became more polarized over time.

Over the past decade low graduation and retention rates has plagued higher education institutions. To assist students in choosing a sequence of courses, choosing majors and successful academic pathways; many institutions provide several on-site academic advising services supported by data driven educational technologies. Accurate performance prediction can serve as the backbone for degree planning software, personalized advising systems and early warning systems that can identify students at-risk of dropping from their field of study. In this work, we present a deep learning based recommender system approach called Neural Collaborative Filtering (NCF) for predicting the grade a student will earn in a course that he/she plans to take in the next-term. Prior grade prediction methods are based on matrix factorization (MF) where students and courses are represented in a latent "knowledge" space. The deep learning inspired approach provides added flexibility in learning the latent spaces in comparison to MF approaches. The proposed approach also incorporates instructor information besides student and course information. Moreover, for proper analysis of the learned model parameters, we assume the embeddings obtained for students, courses and instructors should be non-negative. This non-negative NCF model referred by NCFnn model adds a rectified linear units (ReLU) on the embedding layer of NCF. The experimental results on datasets from George Mason University, a large, public university in the United States, demonstrate that the proposed NCF approaches significantly outperform competitive baselines across different test sets.

We consider the problem of jointly predicting the quality and timing of responses to questions asked in online discussion forums. While prior work has focused on identifying users most likely to answer and/or to provide the highest quality answers to a question, the promptness of the response is also a key factor of user satisfaction. To address this, we propose point process and neural network-based algorithms for three prediction tasks regarding a user's response to a question: whether the user will answer, the net votes that will be received on the answer, and the time that will elapse before the answer. These algorithms learn over a set of 20 features we define for each pair of user and question that quantify both topical and structural aspects of the forums, including discussion post similarities and social centrality measures. Through evaluation on a Stack Overflow dataset consisting of 20,000 question threads, we find that our method outperforms baselines on each prediction task by more than 20%. We also find that the importance of the features varies depending on the task and the amount of historical data available for inference. At the end, we design a question recommendation system that incorporates these predictions to jointly optimize response quality and timing in forums subject to user constraints.

We study a multi-agent partially observable environment in which autonomous agents aim to coordinate their actions, while also learning the parameters of the unknown environment through repeated interactions. In particular, we focus on the role of communication in a multi-agent reinforcement learning problem. We consider a learning algorithm in which agents make decisions based on their own observations of the environment, as well as the observations of other agents, which are collected through communication between agents. We first identify two potential benefits of this type of information sharing when agents' observation quality is heterogeneous: (1) it can facilitate coordination among agents, and (2) it can enhance the learning of all participants, including the better informed agents. We show however that these benefits of communication depend in general on its timing, so that delayed information sharing may be preferred in certain scenarios.

It is of crucial importance to simultaneously protect against sensitive attributes in data while building predictive models. In this paper, we tackle the problem of learning representations from raw data that are i) informative and predictive of desirable variables, and ii) private and protect against adversaries that attempt to recover sensitive variables. We cast this problem under the generative adversarial network (GAN) framework and design three components: an encoder, an ally that predicts the desired variables, and an adversary that predicts the sensitive ones. As a use case, we apply our approach to learn representations of raw student clickstream event data captured as they watch lecture videos in massive open online courses (MOOCs). Through experiments on a real-world dataset collected from a MOOC, we demonstrate that our method can learn a low-dimensional representation of each user that i) excels at classifying whether a user will answer a quiz question correctly, and ii) prevents an adversary from recovering each user's identity. Our results indicate that our approach is effective in learning representations that are both informative and private.

Sensor selection refers to the problem of intelligently selecting a small subset of a collection of available sensors to reduce the sensing cost while preserving signal acquisition performance. The majority of sensor selection algorithms find the subset of sensors that best recovers an arbitrary signal from a number of linear measurements that is larger than the dimension of the signal. In this paper, we develop a new sensor selection algorithm for sparse (or near sparse) signals that finds a subset of sensors that best recovers such signals from a number of measurements that is much smaller than the dimension of the signal. Existing sensor selection algorithms cannot be applied in such situations. Our proposed Incoherent Sensor Selection (Insense) algorithm minimizes a coherence-based cost function that is adapted from recent results in sparse recovery theory. Using three datasets, including a real-world dataset on microbial diagnostics, we demonstrate the superior performance of Insense for sparse-signal sensor selection.

Phase retrieval refers to the problem of recovering real- or complex-valued vectors from magnitude measurements. The best-known algorithms for this problem are iterative in nature and rely on so-called spectral initializers that provide accurate initialization vectors. We propose a novel class of estimators suitable for general nonlinear measurement systems, called linear spectral estimators (LSPEs), which can be used to compute accurate initialization vectors for phase retrieval problems. The proposed LSPEs not only provide accurate initialization vectors for noisy phase retrieval systems with structured or random measurement matrices, but also enable the derivation of sharp and nonasymptotic mean-squared error bounds. We demonstrate the efficacy of LSPEs on synthetic and real-world phase retrieval problems, and we show that our estimators significantly outperform existing methods for structured measurement systems that arise in practice.

The Rasch model is widely used for item response analysis in applications ranging from recommender systems to psychology, education, and finance. While a number of estimators have been proposed for the Rasch model over the last decades, the associated analytical performance guarantees are mostly asymptotic. This paper provides a framework that relies on a novel linear minimum mean-squared error (L-MMSE) estimator which enables an exact, nonasymptotic, and closed-form analysis of the parameter estimation error under the Rasch model. The proposed framework provides guidelines on the number of items and responses required to attain low estimation errors in tests or surveys. We furthermore demonstrate its efficacy on a number of real-world collaborative filtering datasets, which reveals that the proposed L-MMSE estimator performs on par with state-of-the-art nonlinear estimators in terms of predictive performance.

As infrastructure-as-a-service clouds become more popular, cloud providers face the complicated problem of maximizing their resource utilization by handling the dynamics of user demand. Auction-based pricing, such as Amazon EC2 spot pricing, provides an option for users to use idle resources at highly reduced yet dynamic prices; under such a pricing scheme, users place bids for cloud resources, and the provider chooses a threshold “spot” price above which bids are admitted. In this paper, we propose a nonlinear dynamical system model for the time-evolution of the spot price as a function of latent states that characterize user demand in the spot and on-demand markets. This model enables us to adaptively predict future spot prices given past spot price observations, allowing us to derive user bidding strategies for heterogeneous cloud resources that minimize the cost to complete a job with negligible probability of interruption. Along the way, the model also yields novel, empirically verifiable insights into cloud provider behavior. We experimentally validate our model and bidding strategy on two months of Amazon EC2 spot price data and find that our proposed bidding strategy is up to 4 times closer to the optimal strategy in hindsight compared to a baseline regression approach while incurring the same negligible probability of interruption.

The ever growing amount of educational content renders it increasingly difficult to manually generate sufficient practice or quiz questions to accompany it. This paper introduces QG-Net, a recurrent neural network-based model specifically designed for automatically generating quiz questions from educational content such as textbooks. QG-Net, when trained on a publicly available, general-purpose question/answer dataset and without further fine-tuning, is capable of generating high quality questions from textbooks, where the content is significantly different from the training data. Indeed, QG-Net outperforms state-of-the-art neural network-based and rules-based systems for question generation, both when evaluated using standard benchmark datasets and when using human evaluators. QG-Net also scales favorably to applications with large amounts of educational content, since its performance improves with the amount of training data.

Optimizing the selection of learning resources and practice questions to address each individual student's needs has the potential to improve students' learning efficiency. In this paper, we study the problem of selecting a personalized learning action for each student (e.g. watching a lecture video, working on a practice question, etc.), based on their prior performance, in order to maximize their learning outcome. We formulate this problem using the contextual multi-armed bandits framework, where students' prior concept knowledge states (estimated from their responses to questions in previous assessments) correspond to contexts, the personalized learning actions correspond to arms, and their performance on future assessments correspond to rewards. We propose three new Bayesian policies to select personalized learning actions for students that each exhibits advantages over prior work, and experimentally validate them using real-world datasets.

Hashing is an important tool in large-scale machine learning. Unfortunately, current data-dependent hashing algorithms are not robust to small perturbations of the data points, which degrades the performance of nearest neighbor (NN) search. The culprit is the minimization of the L_2-norm, average distortion among pairs of points to find the hash function. Inspired by recent progress in robust optimization, we develop a novel hashing algorithm, dubbed RHash, that instead minimizes the L_1-norm, worst-case distortion among pairs of points. We develop practical and efficient implementations of RHash that couple the alternating direction method of multipliers (ADMM) framework with column generation to scale well to large datasets. A range of experimental evaluations demonstrate the superiority of RHash over ten state-of-the-art binary hashing schemes. In particular, we show that RHash achieves the same retrieval performance as the state-of-the-art algorithms in terms of average precision while using up to 60% fewer bits.

An important, yet largely unstudied, problem in student data analysis is to detect misconceptions from students' responses to open-response questions. Misconception detection enables instructors to deliver more targeted feedback on the misconceptions exhibited by many students in their class, thus improving the quality of instruction. In this paper, we propose a new natural language processing (NLP) framework to detect the common misconceptions among students' textual responses to open-response, short-answer questions. We introduce a probabilistic model for students' textual responses involving misconceptions and experimentally validate it on a real-world student-response dataset. Preliminary experimental results show that our proposed framework excels at classifying whether a response exhibits one or more misconceptions. More importantly, it can also automatically detect the common misconceptions exhibited across responses from multiple students to multiple questions; this is especially important at large scale, since instructors will no longer need to manually specify all possible misconceptions that students might exhibit.

Statistical models of student responses on assessment questions, such as those in homeworks and exams, enable educators and computer-based personalized learning systems to gain insights into students’ knowledge using machine learning. Popular student-response models, including the Rasch model and item response theory models, represent the probability of a student answering a question correctly using an affine function of latent factors. While such models can accurately predict student responses, their ability to interpret the underlying knowledge structure (which is certainly nonlinear) is limited. In response, we develop a new, nonlinear latent variable model that we call the dealbreaker model, in which a student’s success probability is determined by their weakest concept mastery. We develop efficient parameter inference algorithms for this model using novel methods for nonconvex optimization. We show that the dealbreaker model achieves comparable or better prediction performance as compared to affine models with real-world educational datasets. We further demonstrate that the parameters learned by the dealbreaker model are interpretable—they provide key insights into which concepts are critical (i.e., the “dealbreaker”) to answering a question correctly. We conclude by reporting preliminary results for a movie-rating dataset, which illustrate the broader applicability of the dealbreaker model.

While computer and communication technologies have provided effective means to scale up many aspects of education, the submission and grading of assessments such as homework assignments and tests remains a weak link. In this paper, we study the problem of automatically grading the kinds of open response mathematical questions that figure prominently in STEM (science, technology, engineering, and mathematics) courses. Our data-driven framework for mathematical language processing (MLP) leverages solution data from a large number of learners to evaluate the correctness of their solutions, assign partial-credit scores, and provide feedback to each learner on the likely locations of any errors. MLP takes inspiration from the success of natural language processing for text data and comprises three main steps. First, we convert each solution to an open response mathematical question into a series of numerical features. Second, we cluster the features from several solutions to uncover the structures of correct, partially correct, and incorrect solutions. We develop two different clustering approaches, one that leverages generic clustering algorithms and one based on Bayesian nonparametrics. Third, we automatically grade the remaining (potentially large number of) solutions based on their assigned cluster and one instructor-provided grade per cluster. As a bonus, we can track the cluster assignment of each step of a multistep solution and determine when it departs from a cluster of correct solutions, which enables us to indicate the likely locations of errors to learners. We test and validate MLP on real-world MOOC data to demonstrate how it can substantially reduce the human effort required in large-scale educational platforms.

This paper deals with the recovery of an unknown, low-rank matrix from quantized and (possibly) corrupted measurements of a subset of its entries. We develop statistical models and corresponding (multi-)convex optimization algorithms for quantized matrix completion (Q-MC) and quantized robust principal component analysis (Q-RPCA). In order to take into account the quantized nature of the available data, we jointly learn the underlying quantization bin boundaries and recover the low-rank matrix, while removing potential (sparse) corruptions. Experimental results on synthetic and two real-world collaborative filtering datasets demonstrate that directly operating with the quantized measurements - rather than treating them as real values - results in (often significantly) lower recovery error if the number of quantization bins is less than about 10.

We propose SPARFA-Trace, a new machine learning-based framework for time-varying learning and content analytics for educational applications. We develop a novel message passing-based, blind, approximate Kalman filter for sparse factor analysis (SPARFA) that jointly traces learner concept knowledge over time, analyzes learner concept knowledge state transitions (induced by interacting with learning resources, such as textbook sections, lecture videos, etc., or the forgetting effect), and estimates the content organization and difficulty of the questions in assessments. These quantities are estimated solely from binary-valued (correct/incorrect) graded learner response data and the specific actions each learner performs (e.g., answering a question or studying a learning resource) at each time instant. Experimental results on two online course datasets demonstrate that SPARFA-Trace is capable of tracing each learner's concept knowledge evolution over time, analyzing the quality and content organization of learning resources, and estimating the question--concept associations and the question difficulties. Moreover, we show that SPARFA-Trace achieves comparable or better performance in predicting unobserved learner responses compared to existing collaborative filtering and knowledge tracing methods.

We develop a new model and algorithm for machine learning-based learning analytics, which estimate a learner's knowledge of the concepts underlying a domain. Our model represents the probability that a learner provides the correct response to a question in terms of three factors: their understanding of a set of underlying concepts, the concepts involved in each question, and each question's intrinsic difficulty. We estimate these factors given the graded responses to a set of questions. We develop a bi-convex algorithm to solve the resulting SPARse Factor Analysis (SPARFA) problem. We also incorporate user-defined tags on questions to facilitate the interpretability of the estimated factors. Experiments with synthetic and real-world data demonstrate the efficacy of our approach.