Automated grading of student responses still faces numerous challenges, particularly when dealing with complex and ambiguous answers. In particular, large models are prone to scoring bias when handling uncertain responses, and few-shot reasoning methods often lack stability, which limits their applicability in real educational scenarios. To tackle these challenges, we propose the Contrastive Error Mining and FineTuning (CEM-FT) framework, which automatically identifies high-value hard samples by analyzing scoring disagreements between a full fine-tuned model and a few-shot model. A lightweight LoRA adapter is then trained on these samples to refine model performance with minimal computational overhead. Experiments on the SciEntsbank, Beetle, and Mohler datasets show that CEM-FT can improve QWK by up to 3.9% compared to the fine-tuned Qwen model on SciEntsbank datasets, which is a significant improvement over the few-shot baseline. The proposed framework substantially enhances both scoring accuracy and consistency, providing a practical, robust solution for reliable automated assessment with large language models.

Creating a well-structured lesson plan is essential for improving classroom efficiency, yet it is often a labor-intensive process. Recently, many studies have leveraged large language models (LLMs) to generate lesson plans automatically. However, existing methods heavily rely on LLMs that are pre-trained on large-scale universal corpora, which often lack critical educational theory and textbook-specific information. This can lead to inconsistencies and misalignments with textbook content. To address these challenges, we propose CE-LessonPlan, a novel compress-expand framework to generate lesson plans by effectively combining external lesson plan references and textbook information. The framework consists of two key components: a compressor, which synthesizes multiple retrieved references into a cohesive document, and an expander, which integrates textbook-specific information with the parametric knowledge of LLMs to produce another enriched lesson plan. The outputs of the compressor and expander are then seamlessly integrated to create a comprehensive golden context, further enhancing the lesson plan generation process with LLMs. We conduct extensive experiments to demonstrate that CE-LessonPlan outperforms existing methods for generating lesson plans.

Cognitive diagnosis aims to infer students' mastery levels over knowledge components from their learning interactions, supporting personalized education applications. However, existing models encode responses as binary correctness labels, discarding information about which specific option a student selected. This input-level information loss limits their ability to distinguish qualitatively different error types. As a result, providing interpretable diagnostic outputs becomes challenging. To address these limitations, we propose SACD, a semantic-augmented cognitive diagnosis framework that integrates LLM-based semantic analysis with student behavioral modeling. SACD comprises the following key components. First, an LLM-based exercise diagnostic generator analyzes exercise content and produces structured semantic annotations for each answer choice, capturing the specific misconceptions each option represents. Second, a kernel-based alignment mechanism projects semantic embeddings and behavioral representations into a unified kernel space, enabling effective fusion of heterogeneous information. Third, an interpretable diagnosis layer predicts student performance and generates fine-grained mastery estimates, which LLMs further process to produce actionable learning plans. Extensive experiments on three real-world datasets demonstrate that SACD achieves superior prediction accuracy while enabling interpretable, actionable diagnostics.

Knowledge tracing (KT) aims to personalize online education on large-scale web-based platforms by modeling students' evolving knowledge states from their interaction sequences. However, most KT models rely on a single encoder architecture (e.g., self-attention or RNN), with fixed inductive biases that fails to capture the diversity of learning behaviors. Specifically, student learning unfolds across multiple timescales, and interaction sequences contain diverse frequency components ranging from short-term variations to long-term trends. Our data-driven analysis reveals that existing encoders exhibit characteristic frequency biases (e.g., self-attention tends to emphasize low-frequency patterns), highlighting the limitations of any single architecture. To address this problem, we propose FA-KT, a frequency-aware mixture of heterogeneous experts framework. FA-KT combines self-attention, Mamba, CNN, and LSTM experts, each with complementary frequency biases. A frequency-aware router analyzes each sequence's frequency characteristics and adaptively combines experts to create dynamic, personalized encoders for individual students. Across five benchmark datasets, FA-KT consistently outperforms 20 strong KT baselines in predicting future performance. Code is available at https://pykt.org/.

Knowledge tracing (KT) is a problem of modeling students’ knowledge states to predict their future performance by observing their historical learning interactions. The collection of educational data presents significant challenges, as students’ limited learning engagement restricts the generation of large-scale interaction data, while stringent privacy regulations further limit the availability of student learning sequences from online platforms. Hence, it is crucial to enhance the capabilities of deep learning-based KT (DLKT) models by constructing large-scale datasets through the integration of student interaction data across multiple subjects and sources. The success of ChatGPT demonstrates that the decoder-only Transformer architecture is highly effective in capturing complex information from large-scale sequential data. Against this background, we propose a novel decoder-only Transformer architecture-based model, named Unified DLKT ( UniKT ), to learn coherent and unified representations across a wide range of data sources. Specifically, we combine student learning sequences from six educational scenarios and utilize a multi-source encoding to learn unified representations of interactions from mixed data. UniKT is a stack of Transformer decoder layers for handling long-term dependencies among students’ historical interactions and future performance. We evaluate UniKT on six publicly available real-world educational datasets, and experimental results demonstrate that our method outperforms the majority of existing DLKT models in terms of AUC and accuracy. Furthermore, the empirical analysis shows the strong transferability and adaptability of UniKT in learning from multiple sources. To encourage reproducible research, we make our data and code publicly available at https://pykt.org/ .

Knowledge tracing (KT) models students' knowledge states and predicts their future performance based on their historical interaction data. However, attention based KT models struggle to accurately capture diverse forgetting behaviors in ever-growing interaction sequences. First, existing models use uniform time decay matrices, conflating forgetting representations with problem relevance. Second, the fixed-length window prediction paradigm fails to model continuous forgetting processes in expanding sequences. To address these challenges, this paper introduces LefoKT, a unified architecture that enhances attention based KT models by incorporating proposed relative forgetting attention. LefoKT improves forgetting modeling through relative forgetting attention to decouple forgetting patterns from problem relevance. It also enhances attention based KT models' length extrapolation capability for capturing continuous forgetting processes in ever-growing interaction sequences. Extensive experimental results on three datasets validate the effectiveness of LefoKT.

Knowledge tracing (KT) involves using the historical records of student-learning interactions to anticipate their performance on forthcoming questions. Central to this process is the modeling of human cognition to gain deeper insights into how knowledge is acquired and retained. Human cognition is characterized by two key features: long-term cognitive trends, reflecting the gradual accumulation and stabilization of knowledge over time, and short-term cognitive fluctuations, which arise from transient factors such as forgetting or momentary lapses in attention. Although existing attention-based KT models effectively capture long-term cognitive trends, they often fail to adequately address short-term cognitive fluctuations. These limitations lead to overly smoothed cognitive features and reduced model performance, especially when the test data length exceeds the training data length. To address these problems, we propose FlucKT, a novel short-term cognitive fluctuations enhanced attention network for KT tasks. FlucKT improves the attention mechanism in two ways: First, by using a decomposition-based layer with causal convolution to separate and dynamically reweight long-term and short-term cognitive features. Second, by introducing a kernelized bias attention score penalty to enhance focus on short-term fluctuations, improving length generalization capabilities. Our contributions are validated through extensive experiments on three real-world datasets, demonstrating significant improvements in length generalization and prediction performance.

Sequential Recommender Systems (SRS) has stood out as a highly promising technique in numerous domains due to its impressive capability of capturing complex user preferences. Current SRS have employed transformer-based models to give the next-item prediction. Nevertheless, its quadratic computational complexity has often resulted in notable inefficiencies, posing a significant obstacle to real-time recommendation processes. Recently, Mamba has demonstrated its exceptional effectiveness in time series prediction, delivering substantial improvements in both efficiency and effectiveness. However, directly applying Mamba to SRS poses certain challenges. Its unidirectional structure may impede the ability to capture contextual information in user-item interactions, while its instability in state estimation may hinder the ability to capture short-term patterns in interaction sequences. To address these issues, we propose a novel framework called Selective Gated Mamba for Sequential Recommendation (SIGMA). By introducing the Partially Flipped Mamba (PF-Mamba), we construct a special bi-directional structure to address the context modeling challenge. Then, to consolidate PF-Mamba's performance, we employed an input-dependent Dense Selective Gate (DS Gate) to allocate the weights of the two directions and further filter the sequential information. Moreover, for short sequence modeling, we devise a Feature Extract GRU (FE-GRU) to capture the short-term dependencies. Experimental results demonstrate that SIGMA significantly outperforms existing baselines across five real-world datasets. Our implementation code is available in Supplementary Material to ease reproducibility.

Step-level reward models (SRMs) can significantly enhance mathematical reasoning performance through process supervision or step-level preference alignment based on reinforcement learning. The performance of SRMs is pivotal, as they serve as critical guidelines, ensuring that each step in the reasoning process is aligned with desired outcomes. Recently, AlphaZero-like methods, where Monte Carlo Tree Search (MCTS) is employed for automatic step-level preference annotation, have proven particularly effective. However, the precise mechanisms behind the success of SRMs remain largely unexplored. To address this gap, this study delves into the counterintuitive aspects of SRMs, particularly focusing on MCTS-based approaches. Our findings reveal that the removal of natural language descriptions of thought processes has minimal impact on the efficacy of SRMs. Furthermore, we demonstrate that SRMs are adept at assessing the complex logical coherence present in mathematical language while having difficulty in natural language. These insights provide a nuanced understanding of the core elements that drive effective step-level reward modeling in mathematical reasoning. By shedding light on these mechanisms, this study offers valuable guidance for developing more efficient and streamlined SRMs, which can be achieved by focusing on the crucial parts of mathematical reasoning.

Mathematical reasoning remains a challenging area for large language models (LLMs), prompting the development of math-specific LLMs such as LLEMMA, DeepSeekMath, and Qwen2-Math, among others. These models typically follow a two-stage training paradigm: pre-training with math-related corpora and post-training with problem datasets for supervised fine-tuning (SFT). Despite these efforts, the improvements in mathematical reasoning achieved through continued pre-training (CPT) are often less significant compared to those obtained via SFT. This study addresses this discrepancy by exploring alternative strategies during the pre-training phase, focusing on the use of problem-solving data over general mathematical corpora. We investigate three primary research questions: (1) Can problem-solving data enhance the model's mathematical reasoning capabilities more effectively than general mathematical corpora during CPT? (2) Are synthetic data from the same source equally effective, and which synthesis methods are most efficient? (3) How do the capabilities developed from the same problem-solving data differ between the CPT and SFT stages, and what factors contribute to these differences? Our findings indicate that problem-solving data significantly enhances the model's mathematical capabilities compared to general mathematical corpora. We also identify effective data synthesis methods, demonstrating that the tutorship amplification synthesis method achieves the best performance. Furthermore, while SFT facilitates instruction-following abilities, it underperforms compared to CPT with the same data, which can be partially attributed to its poor learning capacity for more challenging problem-solving data. These insights provide valuable guidance for optimizing the mathematical reasoning capabilities of LLMs, culminating in our development of a powerful mathematical base model called MathGPT-8B.

Knowledge tracing (KT) is an essential task in online education systems. It aims to predict the future performance of students based on their historical learning interaction data. Despite significant advancements in attention-based KT models, they still face some limitations: inaccurate input representation and excessive student forgetting modeling. These limitations often lead to the attention noise problem: the model assigns non-negligible attention weight to some information that is cognitively irrelevant in nature, thereby generating interference signals. To address this problem, we propose a novel KT model, i.e., DenoiseKT. DenoiseKT effectively models the difficulty of the questions and utilizes graph neural network to capture the complex relationship between questions, thereby refining the representations of input features. Additionally, the denoised attention mechanism introduces a weight factor to reduce the model's attention weight distribution on irrelevant information. We extensively compare DenoiseKT with 22 state-of-the-art KT models on 4 widely-used public datasets. Experimental results show that DenoiseKT can effectively solve the attention noise problem and outperform other models. The source code of DenoiseKT is available at https://pykt.org.

Transformer-based models have gained significant traction in sequential recommender systems (SRSs) for their ability to capture user-item interactions effectively. However, these models often suffer from high computational costs and slow inference. Meanwhile, existing efficient SRS approaches struggle to embed high-quality semantic and positional information into latent representations. To tackle these challenges, this paper introduces GLINT-RU, a lightweight and efficient SRS leveraging a single-layer dense selective Gated Recurrent Units (GRU) module to accelerate inference. By incorporating a dense selective gate, GLINT-RU adaptively captures temporal dependencies and fine-grained positional information, generating high-quality latent representations. Additionally, a parallel mixing block infuses fine-grained positional features into user-item interactions, enhancing both recommendation quality and efficiency. Extensive experiments on three datasets demonstrate that GLINT-RU achieves superior prediction accuracy and inference speed, outperforming baselines based on RNNs, Transformers, MLPs, and SSMs. These results establish GLINT-RU as a powerful and efficient solution for SRSs. The implementation code is publicly available for reproducibility. https://github.com/szhang-cityu/GLINT-RU.

Recent advancements in Large Language Models (LLMs) have emphasized the critical role of fine-tuning (FT) techniques in adapting LLMs to specific tasks, especially when retraining from scratch is computationally infeasible. Fine-tuning enables LLMs to leverage task- or domain-specific data, producing models that more effectively meet the requirements of targeted applications. However, conventional FT approaches often suffer from catastrophic forgetting and suboptimal data efficiency, limiting their real-world applicability. To address these challenges, this paper proposes DEAL, a novel framework that integrates Low-Rank Adaptation (LoRA) with a continuous fine-tuning strategy. By incorporating knowledge retention and adaptive parameter update modules, the framework mitigates the limitations of existing FT methods while maintaining efficiency. Experiments on 15 diverse datasets show that DEAL consistently outperforms baseline methods, yielding substantial gains in task accuracy and resource efficiency. These findings demonstrate the potential of our approach to advance continual adaptation in LLMs by enhancing task performance while improving resource efficiency. The source code is publicly available at https://github.com/Applied-Machine-Learning-Lab/DEAL.

Recent deep sequential recommendation models often struggle to effectively model key characteristics of user behaviors, particularly in handling sequence length variations and capturing diverse interaction patterns. We propose STAR-Rec, a novel architecture that synergistically combines preference-aware attention and state-space modeling through a sequence-level mixture-of-experts framework. STAR-Rec addresses these challenges by: (1) employing preference-aware attention to capture both inherently similar item relationships and diverse preferences (2) utilizing state-space modeling to efficiently process variable-length sequences with linear complexity, and (3) incorporating a mixture-of-experts component that adaptively routes different behavioral patterns to specialized experts, handling both focused category-specific browsing and diverse category exploration patterns. We theoretically demonstrate how the state space model and attention mechanisms can be naturally unified in recommendation scenarios, where SSM captures temporal dynamics through state compression while attention models both similar and diverse item relationships. Extensive experiments on four real-world datasets demonstrate that STAR-Rec consistently outperforms state-of-the-art sequential recommendation methods, particularly in scenarios involving diverse user behaviors and varying sequence lengths. The implementation code is available anonymously online for easy reproducibility.

Knowledge tracing (KT) aims to predict students' future performance based on their past learning activities. However, no one is perfect. Factors such as carelessness, fatigue, and stress often cause students to make mistakes on problems they have already mastered, leading to anomalies in their historical learning data. These anomalies disrupt inherent patterns in the data, misleading the KT model. Extracting cognitive patterns that accurately reflect students' knowledge mastery from such error-prone data remains a significant challenge. Against this background, this paper proposes a novel KT method named RoubstKT, inspired by educational measurement theory and frequency-based decomposition. A cognitive decoupling analyzer is proposed to decouple the student's cognitive pattern and random factors from the data through smoothing and subtraction operations, then recombine them using a gating mechanism or adaptive parameter fusion strategy. To more effectively diagnose students' knowledge mastery, we employ a decay-based attention mechanism that focuses on random behaviors at adjacent time steps. We conducted comprehensive experiments based on real-world datasets and targeted datasets with added random noise. The experimental results demonstrated the effectiveness of the proposed method.

WebAssembly (Wasm), a binary instruction format designed for efficient cross-platform execution, has rapidly become a foundational web standard, widely adopted in browsers, client-side, and server-side applications. However, its growing popularity has led to an increase in Wasm-targeted malware, including cryptojackers and obfuscated malicious scripts, which pose significant threats to web security. In spite of progress in deep learning based detection methods for Wasm malware, such as MINOS, these approaches face substantial performance degradation in adversarial environments. In our experiments, MINOS's detection accuracy dropped to 49.90% under adversarial attacks, revealing critical vulnerabilities. To address this, we introduce WasmGuard, a robust malware detection framework tailored for Wasm. WasmGuard employs FGSM-based adversarial training with prior-based initialization for perturbation bytes in customized sections, coupled with a novel adversarial contrastive learning objective. Using our large-scale dataset, WasmMal-15K (publicly available at https://github.com/Yuxia-Sun/WasmMal GitHub), WasmGuard outperforms six competing methods, achieving up to 99.20% Robust Accuracy and 99.93% Standard Accuracy under PGD-50 adversarial attacks, while maintaining low training overhead. Additionally, we have released WebChecker, a WasmGuard-powered browser plugin, providing real-time protection against malicious Wasm files, at https://github.com/Yuxia-Sun/WasmGuard.

Delivering superior search services is crucial for enhancing customer experience and driving revenue growth in e-commerce. Conventionally, search systems model user behaviors by combining user preference and query-item relevance statically, often through a fixed logical 'and' relationship. This paper reexamines existing approaches through a unified lens using causal graphs and Venn diagrams, uncovering two prevalent yet significant issues: entangled preference and relevance effects, and a collapsed modeling space. To surmount these challenges, our research introduces a novel framework, DRP, which enhances search accuracy through two components to reconstruct the behavior modeling space. Specifically, we implement preference editing to proactively remove the relevance effect from preference predictions, yielding untainted user preferences. Additionally, we employ adaptive fusion, which dynamically adjusts fusion criteria to align with the varying patterns of relevance and preference, facilitating more nuanced and tailored behavior predictions within the reconstructed modeling space. Empirical validation on two public datasets and a proprietary e-commerce search dataset underscores the superiority of our proposed methodology, demonstrating marked improvements in performance over existing approaches. The code is available at https://github.com/Applied-Machine-Learning-Lab/DRP.

Knowledge tracing (KT) plays a crucial role in predicting students’ future performance by analyzing their historical learning processes. Deep neural networks (DNNs) have shown great potential in solving the KT problem. However, there still exist some important challenges when applying deep learning techniques to model the KT process. The first challenge lies in modeling the individual question information. This is crucial because students’ knowledge acquisition on questions that share the same set of knowledge components (KCs) may vary significantly. However, due to the large question bank, the average number of interactions per question may not be sufficient. This limitation can potentially result in overfitting of the question embedding and inaccurate question knowledge acquisition state that relies on its corresponding question representation. Furthermore, there is a considerable portion of questions receiving relatively less interaction from students in comparison to the majority of questions. This can further increase the risk of overfitting and lower the accuracy of the obtained question knowledge acquisition state. The second challenge lies in interpreting the prediction results from existing deep learning-based KT models. In real-world applications, while it may not be necessary to have complete transparency and interpretability of the model parameters, it is crucial to present the model’s prediction results in a manner that teachers find interpretable. This makes teachers accept the rationale behind the prediction results and utilize them to design teaching activities and tailored learning strategies for students. However, the inherent black-box nature of deep learning techniques often poses a hurdle for teachers to fully embrace the model’s prediction results. To address these challenges, we propose a Question-centric Multi-experts Contrastive Learning framework for KT called Q-MCKT. This framework explicitly models students’ knowledge acquisition state at both the question and concept levels. It leverages the mixture of experts technique to capture a more robust and accurate knowledge acquisition state in both question and concept levels for prediction. Additionally, a fine-grained question-centric contrastive learning task is introduced to enhance the representations of less interactive questions and improve the accuracy of their corresponding question knowledge acquisition states. Moreover, Q-MCKT utilizes an item response theory-based prediction layer to generate interpretable prediction results based on the knowledge acquisition states obtained from the question and concept knowledge acquisition modules. We evaluate the proposed Q-MCKT framework on four public real-world educational datasets. The experimental results demonstrate that our approach outperforms a wide range of deep learning-based KT models in terms of prediction accuracy while maintaining better model interpretability. To ensure reproducibility, we have provided all the datasets and code on our website at https://github.com/rattlesnakey/Q-MCKT .

Knowledge tracing (KT) involves utilizing historical data from students’ learning interactions to model their mastery of knowledge over time, with the aim of predicting their future performance in interactions. Recently, significant advancements have been achieved through the application of various deep learning methodologies to address the KT challenge. However, a considerable proportion of deep learning-based knowledge tracing (DLKT) approaches exhibit striking similarities in their methodologies, and model designs, and even the outcomes demonstrate minimal divergence. In addition, the evaluation procedures employed in current DLKT studies are not standardized, resulting in substantial inconsistencies in the reported area under the curve (AUC) outcomes, despite analyzing the same model on identical datasets. To address the two aforementioned problems, this paper proposes a generalized DLKT framework and represents the existing DLKT models with five components, i.e., multimodal data encoder, student knowledge memory, auxiliary knowledge base, learning outcome objective, and computational efficiency and scalability. Furthermore, we develop and open source a standardized DLKT benchmark platform namedpyKT,1that consists of a standardized set of integrated data preprocessing procedures on 9 popular datasets across different domains, and 21 frequently compared DLKT model implementations. WithpyKT, we conduct empirical and reproducible research to assess the performance of prevalent DLKT algorithms in an unbiased and clear setting over multiple data sources. Finally, we discuss the applications of KT techniques in the educational sector and their future development directions.

Sketching in Virtual Reality (VR) is challenging mainly due to the absence of physical surface support and virtual depth perception cues, which induce high cognitive and sensorimotor load. This paper presents WieldingCanvas, an interactive VR sketching platform that integrates canvas manipulations to draw lines and curves in 3D. Informed by real-life examples of two-handed creative activities, WieldingCanvas interprets users’ spatial gestures to move, swing, rotate, transform, or fold a virtual canvas, whereby users simply draw primitive strokes on the canvas, which are turned into finer and more sophisticated shapes via the manipulation of the canvas. We evaluated the capability and user experience of WieldingCanvas with two studies where participants were asked to sketch target shapes. A set of freehand sketches of high aesthetic qualities were created, and the results demonstrated that WieldingCanvas can assist users with creating 3D sketches.

Accurately judging student on-going performance is crucial for adaptive teaching. In this work, we focus on the task of automatically predicting students' levels of mastery of math questions from teacher-student classroom dialogue data in online one-on-one classes. As a step toward this direction, we introduce the Multi-turn Classroom Dialogue (MCD) dataset as a benchmark testing the capabilities of machine learning models in classroom conversation understanding of student performance judgment. Our dataset contains aligned multi-turn spoken language of 5000+ unique samples of solving grade-8 math questions collected from 500+ hours' worth of online one-on-one tutoring classes. In our experiments, we assess various state-of-the-art models on the MCD dataset, highlighting the importance of understanding multi-turn dialogues and handling noisy ASR transcriptions. Our findings demonstrate the dataset's utility in advancing research on automated student performance assessment. To encourage reproducible research, we make our data publicly available at https://github.com/ai4ed/MCD.

Recommender systems play a pivotal role across practical scenarios, showcasing remarkable capabilities in user preference modeling. However, the centralized learning paradigm predominantly used raises serious privacy concerns. The federated recommender system (FedRS) addresses this by updating models on clients, while a central server orchestrates training without accessing private data. Existing FedRS approaches, however, face unresolved challenges, including non-convex optimization, vulnerability, potential privacy leakage risk, and communication inefficiency. This paper addresses these challenges by reformulating the federated recommendation problem as a convex optimization issue, ensuring convergence to the global optimum. Based on this, we devise a novel method, RFRec, to tackle this optimization problem efficiently. In addition, we propose RFRecF, a highly efficient version that incorporates non-uniform stochastic gradient descent to improve communication efficiency. In user preference modeling, both methods learn local and global models, collaboratively learning users' common and personalized interests under the federated learning setting. Moreover, both methods significantly enhance communication efficiency, robustness, and privacy protection, with theoretical support. Comprehensive evaluations on four benchmark datasets demonstrate RFRec and RFRecF's superior performance compared to diverse baselines. The code is available to ease reproducibility1.

Knowledge tracing (KT) is a prediction task that aims to predict students’ future performance based on their past learning data. The rapid progress in attention mechanisms has led to the emergence of various high-performing attention based KT models. However, in online or personalized education settings, students’ varying learning paths result in different lengths of student interaction sequences, which poses a significant challenge for attention based KT models as their context window sizes are fixed during both training and prediction stages. We refer to this as the length extrapolation of KT model. In this paper, we propose extraKT to facilitate better extrapolation that learn from student interactions with a short context window and continue to perform well across various longer context window sizes at prediction stage. Specifically, we negatively bias attention scores with linearly decreasing penalties that are proportional to query-key distance, which efficiently represents short-term forgetting characteristics of student knowledge states. We conduct comprehensive and rigorous experiments on three real-world educational datasets. The results show that our extraKT model exhibits robust length extrapolation capability and outperforms state-of-the-art baseline models in terms of AUC and accuracy. To encourage reproducible research, we merge our data and code to the publicly available pyKT benchmark at https://github.com/pykt-team/pykt-toolkit.

Recent advanced AI technologies, especially large language models (LLMs) like GPTs, have significantly advanced the field of data mining and led to the development of various LLM-based applications. AI for education (AI4EDU) is a vibrant multi-disciplinary field of data mining, machine learning, and education, with increasing importance and extraordinary potential. In this field, LLM and adaptive learning-based models can be utilized as interfaces in human-in-the-loop education systems, where the model serves as a mediator among the teacher, students, and machine capabilities, including its own. This perspective has several benefits, including the ability to personalize interactions, allow unprecedented flexibility and adaptivity for human-AI collaboration and improve the user experience. However, several challenges still exist, including the need for more robust and efficient algorithms, designing effective user interfaces, and ensuring ethical considerations are addressed. This workshop aims to bring together researchers and practitioners from academia and industry to explore cutting-edge AI technologies for personalized education, especially the potential of LLMs and adaptive learning technologies.

Deep generative models allow the synthesis of realistic human faces from freehand sketches or semantic maps. However, although they are flexible, sketches and semantic maps provide too much freedom for manipulation, and thus, are not easy for novice users to control. In this study, we present DeepFaceReshaping, a novel landmark-based deep generative framework for interactive face reshaping. To edit the shape of a face realistically by manipulating a small number of face landmarks, we employ neural shape deformation to reshape individual face components. Furthermore, we propose a novel Transformer-based partial refinement network to synthesize the reshaped face components conditioned on the edited landmarks, and fuse the components to generate the entire face using a local-to-global approach. In this manner, we limit possible reshaping effects within a feasible component-based face space. Thus, our interface is intuitive even for novice users, as confirmed by a user study. Our experiments demonstrate that our method outperforms traditional warping-based approaches and recent deep generative techniques.

As Virtual Reality (VR) applications gain popularity, the need for a secure, usable, and memorable user authentication method becomes crucial. However, security and privacy in such VR applications are often ignored. Current methods are insufficient in preventing man-in-the-room (MITR) attacks, which allow attackers to observe user interactions in VR while remaining invisible, and inputted passwords can easily be stolen. In this study, we propose a dynamic combination of multi-attribute authentication methods for VR, where various 3D objects and their attributes can be created and displayed. Users must select combinations of 3D objects and their attributes provided by our designed principles for identity authentication. We explore the impact of method parameters on security and provide three specific parameter schemes to deploy the practical authentication system. We designed three user studies to evaluate the usability, security, and memorability of our authentication system. The results show that the proposed scheme can effectively resist both shoulder surfing and MITR attacks with unsuccessful attack rates of 100% and 95.83%, respectively. Furthermore, this research provides suggestions to secure VR applications while maintaining usability and enhancing the memorability of the authentication method.

Spatial search tasks are common and crucial in many Virtual Reality (VR) applications. Traditional methods to enhance the performance of spatial search often employ sensory cues such as visual, auditory, or haptic feedback. However, the design and use of bimanual haptic feedback with two VR controllers for spatial search in VR remains largely unexplored. In this work, we explored bimanual haptic feedback with various combinations of haptic properties, where four types of bimanual haptic feedback were designed, for spatial search tasks in VR. Two experiments were designed to evaluate the effectiveness of bimanual haptic feedback on spatial direction guidance and search in VR. The results from the first experiment reveal that our proposed bimanual haptic schemes significantly enhanced the recognition of spatial directions in terms of accuracy and speed compared to spatial audio feedback. The second experiment's findings suggest that the performance of bimanual haptic feedback was comparable to or even better than the visual arrow, especially in reducing the angle of head movement and enhancing searching targets behind the participants, which was supported by subjective feedback as well. Based on these findings, we have derived a set of design recommendations for spatial search using bimanual haptic feedback in VR.

Knowledge tracing (KT) is a crucial technique to predict students’ future performance by observing their historical learning processes. Due to the powerful representation ability of deep neural networks, remarkable progress has been made by using deep learning techniques to solve the KT problem. The majority of existing approaches rely on the homogeneous question assumption that questions have equivalent contributions if they share the same set of knowledge components. Unfortunately, this assumption is inaccurate in real-world educational scenarios. Furthermore, it is very challenging to interpret the prediction results from the existing deep learning based KT models. Therefore, in this paper, we present QIKT, a question-centric interpretable KT model to address the above challenges. The proposed QIKT approach explicitly models students’ knowledge state variations at a fine-grained level with question-sensitive cognitive representations that are jointly learned from a question-centric knowledge acquisition module and a question-centric problem solving module. Meanwhile, the QIKT utilizes an item response theory based prediction layer to generate interpretable prediction results. The proposed QIKT model is evaluated on three public real-world educational datasets. The results demonstrate that our approach is superior on the KT prediction task, and it outperforms a wide range of deep learning based KT models in terms of prediction accuracy with better model interpretability. To encourage reproducible results, we have provided all the datasets and code at https://pykt.org/.

Accurately judging students' ongoing performance is very crucial for real-world educational scenarios. In this work, we focus on the task of automatically predicting students' levels of mastery of math questions from teacher-student classroom dialogue data in the online learning environment. We propose a novel neural network armed with a multi-granularity attention mechanism to capture the personalized pedagogical instructions from the very noisy teacher-student dialogue transcriptions. We conduct experiments on a real-world educational dataset and the results demonstrate the superiority and availability of our model in terms of various evaluation metrics.

With the acceleration of urbanization, traffic forecasting has become an essential role in smart city construction. In the context of spatio-temporal prediction, the key lies in how to model the dependencies of sensors. However, existing works basically only consider the micro relationships between sensors, where the sensors are treated equally, and their macroscopic dependencies are neglected. In this paper, we argue to rethink the sensor's dependency modeling from two hierarchies: regional and global perspectives. Particularly, we merge original sensors with high intra-region correlation as a region node to preserve the inter-region dependency. Then, we generate representative and common spatio-temporal patterns as global nodes to reflect a global dependency between sensors and provide auxiliary information for spatio-temporal dependency learning. In pursuit of the generality and reality of node representations, we incorporate a Meta GCN to calibrate the regional and global nodes in the physical data space. Furthermore, we devise the cross-hierarchy graph convolution to propagate information from different hierarchies. In a nutshell, we propose a Hierarchical Information Enhanced Spatio-Temporal prediction method, HIEST, to create and utilize the regional dependency and common spatio-temporal patterns. Extensive experiments have verified the leading performance of our HIEST against state-of-the-art baselines. We publicize the code to ease reproducibility1.

Traffic prediction is a typical spatio-temporal data mining task and has great significance to the public transportation system. Considering the demand for its grand application, we recognize key factors for an ideal spatio-temporal prediction method: efficient, lightweight, and effective. However, the current deep model-based spatio-temporal prediction solutions generally own intricate architectures with cumbersome optimization, which can hardly meet these expectations. To accomplish the above goals, we propose an intuitive and novel framework, MLPST, a pure multi-layer perceptron architecture for traffic prediction. Specifically, we first capture spatial relationships from both local and global receptive fields. Then, temporal dependencies in different intervals are comprehensively considered. Through compact and swift MLP processing, MLPST can well capture the spatial and temporal dependencies while requiring only linear computational complexity, as well as model parameters that are more than an order of magnitude lower than baselines. Extensive experiments validated the superior effectiveness and efficiency of MLPST against advanced baselines, and among models with optimal accuracy, MLPST achieves the best time and space efficiency.

In the era of information explosion, spatio-temporal data mining serves as a critical part of urban management. Considering the various fields demanding attention, e.g., traffic state, human activity, and social event, predicting multiple spatio-temporal attributes simultaneously can alleviate regulatory pressure and foster smart city construction. However, current research can not handle the spatio-temporal multi-attribute prediction well due to the complex relationships between diverse attributes. The key challenge lies in how to address the common spatio-temporal patterns while tackling their distinctions. In this paper, we propose an effective solution for spatio-temporal multi-attribute prediction, PromptST. We devise a spatio-temporal transformer and a parameter-sharing training scheme to address the common knowledge among different spatio-temporal attributes. Then, we elaborate a spatio-temporal prompt tuning strategy to fit the specific attributes in a lightweight manner. Through the pretrain and prompt tuning phases, our PromptST is able to enhance the specific spatio-temoral characteristic capture by prompting the backbone model to fit the specific target attribute while maintaining the learned common knowledge. Extensive experiments on real-world datasets verify that our PromptST attains state-of-the-art performance. Furthermore, we also prove PromptST owns good transferability on unseen spatio-temporal attributes, which brings promising application potential in urban computing. The implementation code is available to ease reproducibility.

The studies on adversarial attacks and defenses have greatly improved the robustness of Deep Neural Networks (DNNs). Most advanced approaches have been overwhelmingly designed for continuous data such as images. However, these achievements are still hard to be generalized to categorical data. To bridge this gap, we propose a novel framework, Probabilistic Categorical Adversarial Attack (or PCAA). It transfers the discrete optimization problem of finding categorical adversarial examples to a continuous problem that can be solved via gradient-based methods. We analyze the optimality (attack success rate) and time complexity of PCAA to demonstrate its significant advantage over current search-based attacks. More importantly, through extensive empirical studies, we demonstrate that the well-established defenses for continuous data, such as adversarial training and TRADES, can be easily accommodated to defend DNNs for categorical data.

Recent studies suggest that "memorization" is one necessary factor for overparameterized deep neural networks (DNNs) to achieve optimal performance. Specifically, the perfectly fitted DNNs can memorize the labels of many atypical samples, generalize their memorization to correctly classify test atypical samples and enjoy better test performance. While, DNNs which are optimized via adversarial training algorithms can also achieve perfect training performance by memorizing the labels of atypical samples, as well as the adversarially perturbed atypical samples. However, adversarially trained models always suffer from poor generalization, with both relatively low clean accuracy and robustness on the test set. In this work, we study the effect of memorization in adversarial trained DNNs and disclose two important findings: (a) Memorizing atypical samples is only effective to improve DNN's accuracy on clean atypical samples, but hardly improve their adversarial robustness and (b) Memorizing certain atypical samples will even hurt the DNN's performance on typical samples. Based on these two findings, we propose Benign Adversarial Training (BAT) which can facilitate adversarial training to avoid fitting "harmful" atypical samples and fit as more "benign" atypical samples as possible. In our experiments, we validate the effectiveness of BAT, and show that it can achieve better clean accuracy vs. robustness trade-off than baseline methods, in benchmark datasets for image classification.

Knowledge tracing (KT) is a task that predicts students' future performance based on their historical learning interactions. With the rapid development of deep learning techniques, existing KT approaches follow a data-driven paradigm that uses massive problem-solving records to model students' learning processes. However, although the educational contexts contain various factors that may have an influence on student learning outcomes, existing public KT datasets mainly consist of anonymized ID-like features, which may hinder the research advances towards this field. Therefore, in this work, we present, \emph{XES3G5M}, a large-scale dataset with rich auxiliary information about questions and their associated knowledge components (KCs)\footnote{\label{ft:kc}A KC is a generalization of everyday terms like concept, principle, fact, or skill.}. The XES3G5M dataset is collected from a real-world online math learning platform, which contains 7,652 questions, and 865 KCs with 5,549,635 interactions from 18,066 students. To the best of our knowledge, the XES3G5M dataset not only has the largest number of KCs in math domain but contains the richest contextual information including tree structured KC relations, question types, textual contents and analysis and student response timestamps. Furthermore, we build a comprehensive benchmark on 19 state-of-the-art deep learning based knowledge tracing (DLKT) models. Extensive experiments demonstrate the effectiveness of leveraging the auxiliary information in our XES3G5M with DLKT models. We hope the proposed dataset can effectively facilitate the KT research work.

Deep recommender systems typically involve numerous feature fields for users and items, with a large number of low-frequency features. These low-frequency features would reduce the prediction accuracy with large storage space due to their vast quantity and inadequate training. Some pioneering studies have explored embedding compression techniques to address this issue of the trade-off between storage space and model predictability. However, these methods have difficulty compacting the embedding of low-frequency features in various feature fields due to the high demand for human experience and computing resources during hyper-parameter searching. In this paper, we propose the AutoDPQ framework, which automatically compacts low-frequency feature embeddings for each feature field to an adaptive magnitude. Experimental results indicate that AutoDPQ can significantly reduce the parameter space while improving recommendation accuracy. Moreover, AutoDPQ is compatible with various deep CTR models by improving their performance significantly with high efficiency.

Knowledge tracing (KT) is the problem of predicting students' future performance based on their historical interaction sequences. With the advanced capability of capturing contextual long-term dependency, attention mechanism becomes one of the essential components in many deep learning based KT (DLKT) models. In spite of the impressive performance achieved by these attentional DLKT models, many of them are often vulnerable to run the risk of overfitting, especially on small-scale educational datasets. Therefore, in this paper, we propose sparseKT, a simple yet effective framework to improve the robustness and generalization of the attention based DLKT approaches. Specifically, we incorporate a k-selection module to only pick items with the highest attention scores. We propose two sparsification heuristics: (1) soft-thresholding sparse attention and (2) top-K sparse attention. We show that our sparseKT is able to help attentional KT models get rid of irrelevant student interactions and improve the predictive performance when compared to 11 state-of-the-art KT models on three publicly available real-world educational datasets. To encourage reproducible research, we make our data and code publicly available at https://github.com/pykt-team/pykt-toolkit1..

Transformer models have achieved remarkable success in sequential recommender systems (SRSs). However, computing the attention matrix in traditional dot-product attention mechanisms results in a quadratic complexity with sequence lengths, leading to high computational costs for long-term sequential recommendation. Motivated by the above observation, we propose a novel L2-Normalized Linear Attention for the Transformer-based Sequential Recommender Systems (LinRec), which theoretically improves efficiency while preserving the learning capabilities of the traditional dot-product attention. Specifically, by thoroughly examining the equivalence conditions of efficient attention mechanisms, we show that LinRec possesses linear complexity while preserving the property of attention mechanisms. In addition, we reveal its latent efficiency properties by interpreting the proposed LinRec mechanism through a statistical lens. Extensive experiments are conducted based on two public benchmark datasets, demonstrating that the combination of LinRec and Transformer models achieves comparable or even superior performance than state-of-the-art Transformer-based SRS models while significantly improving time and memory efficiency. The implementation code is available online at https://github.com/Applied-Machine-Learning-Lab/LinRec.>

Knowledge tracing (KT) is the task of using students' historical learning interaction data to model their knowledge mastery over time so as to make predictions on their future interaction performance. Recently, remarkable progress has been made of using various deep learning techniques to solve the KT problem. However, the success behind deep learning based knowledge tracing (DLKT) approaches is still left somewhat unknown and proper measurement and analysis of these DLKT approaches remain a challenge.

Knowledge tracing (KT) is the problem of predicting students’ future performance based on their historical interactions with intelligent tutoring systems. Recent studies have applied multiple types of deep neural networks to solve the KT problem. However, there are two important factors in real-world educational data that are not well represented. First, most existing works augment input representations with the co-occurrence matrix of questions and knowledge components1 (KCs) but fail to explicitly integrate such intrinsic relations into the final response prediction task. Second, the individualized historical performance of students has not been well captured. In this paper, we proposed AT-DKT to improve the prediction performance of the original deep knowledge tracing model with two auxiliary learning tasks, i.e., question tagging (QT) prediction task and individualized prior knowledge (IK) prediction task. Specifically, the QT task helps learn better question representations by predicting whether questions contain specific KCs. The IK task captures students’ global historical performance by progressively predicting student-level prior knowledge that is hidden in students’ historical learning interactions. We conduct comprehensive experiments on three real-world educational datasets and compare the proposed approach to both deep sequential KT models and non-sequential models. Experimental results show that AT-DKT outperforms all sequential models with more than 0.9% improvements of AUC for all datasets, and is almost the second best compared to non-sequential models. Furthermore, we conduct both ablation studies and quantitative analysis to show the effectiveness of auxiliary tasks and the superior prediction outcomes of AT-DKT. To encourage reproducible research, we make our data and code publicly available at https://github.com/pykt-team/pykt-toolkit 2.

Pairwise learning strategies are prevalent for optimizing recommendation models on implicit feedback data, which usually learns user preference by discriminating between positive (i.e., clicked by a user) and negative items (i.e., obtained by negative sampling). However, the size of different item groups (specified by item attribute) is usually unevenly distributed. We empirically find that the commonly used uniform negative sampling strategy for pairwise algorithms (e.g., BPR) can inherit such data bias and oversample the majority item group as negative instances, severely countering group fairness on the item side. In this paper, we propose a Fairly adaptive Negative sampling approach (FairNeg), which improves item group fairness via adaptively adjusting the group-level negative sampling distribution in the training process. In particular, it first perceives the model’s unfairness status at each step and then adjusts the group-wise sampling distribution with an adaptive momentum update strategy for better facilitating fairness optimization. Moreover, a negative sampling distribution Mixup mechanism is proposed, which gracefully incorporates existing importance-aware sampling techniques intended for mining informative negative samples, thus allowing for achieving multiple optimization purposes. Extensive experiments on four public datasets show our proposed method’s superiority in group fairness enhancement and fairness-utility tradeoff.

Sequential recommendation aims to offer potentially interesting products to users by capturing their historical sequence of interacted items. Although it has facilitated extensive physical scenarios, sequential recommendation for multi-modal sequences has long been neglected. Multi-modal data that depicts a user’s historical interactions exists ubiquitously, such as product pictures, textual descriptions, and interacted item sequences, providing semantic information from multiple perspectives that comprehensively describe a user’s preferences. However, existing sequential recommendation methods either fail to directly handle multi-modality or suffer from high computational complexity. To address this, we propose a novel Multi-Modal Multi-Layer Perceptron (MMMLP) for maintaining multi-modal sequences for sequential recommendation. MMMLP is a purely MLP-based architecture that consists of three modules - the Feature Mixer Layer, Fusion Mixer Layer, and Prediction Layer - and has an edge on both efficacy and efficiency. Extensive experiments show that MMMLP achieves state-of-the-art performance with linear complexity. We also conduct ablating analysis to verify the contribution of each component. Furthermore, compatible experiments are devised, and the results show that the multi-modal representation learned by our proposed model generally benefits other recommendation models, emphasizing our model’s ability to handle multi-modal information. We have made our code available online to ease reproducibility1.

We propose a new method for realistic human motion transfer using a generative adversarial network (GAN), which generates a motion video of a target character imitating actions of a source character, while maintaining high authenticity of the generated results. We tackle the problem by decoupling and recombining the posture information and appearance information of both the source and target characters. The innovation of our approach lies in the use of the projection of a reconstructed 3D human model as the condition of GAN to better maintain the structural integrity of transfer results in different poses. We further introduce a detail enhancement net to enhance the details of transfer results by exploiting the details in real source frames. Extensive experiments show that our approach yields better results both qualitatively and quantitatively than the state-of-the-art methods.

Online dialogic instructions are a set of pedagogical instructions used in real-world online educational contexts to motivate students, help understand learning materials, and build effective study habits. In spite of the popularity and advantages of online learning, the education technology and educational data mining communities still suffer from the lack of large-scale, high-quality, and well-annotated teaching instruction datasets to study computational approaches to automatically detect online dialogic instructions and further improve the online teaching effectiveness. Therefore, in this paper, we present a dataset of online dialogic instruction detection, DialogID, which contains 30,431 effective dialogic instructions. These teaching instructions are well annotated into 8 categories. Furthermore, we utilize the prevalent pre-trained language models (PLMs) and propose a simple yet effective adversarial training learning paradigm to improve the quality and generalization of dialogic instruction detection. Extensive experiments demonstrate that our approach outperforms a wide range of baseline methods. The data and our code are available for research purposes from: https://github.com/ai4ed/DialogID.

Knowledge tracing (KT) is the task of using students' historical learning interaction data to model their knowledge mastery over time so as to make predictions on their future interaction performance. Recently, remarkable progress has been made of using various deep learning techniques to solve the KT problem. However, the success behind deep learning based knowledge tracing (DLKT) approaches is still left somewhat unknown and proper measurement and analysis of these DLKT approaches remain a challenge. First, data preprocessing procedures in existing works are often private and custom, which limits experimental standardization. Furthermore, existing DLKT studies often differ in terms of the evaluation protocol and are far away real-world educational contexts. To address these problems, we introduce a comprehensive python based benchmark platform, \textsc{pyKT}, to guarantee valid comparisons across DLKT methods via thorough evaluations. The \textsc{pyKT} library consists of a standardized set of integrated data preprocessing procedures on 7 popular datasets across different domains, and 10 frequently compared DLKT model implementations for transparent experiments. Results from our fine-grained and rigorous empirical KT studies yield a set of observations and suggestions for effective DLKT, e.g., wrong evaluation setting may cause label leakage that generally leads to performance inflation; and the improvement of many DLKT approaches is minimal compared to the very first DLKT model proposed by Piech et al. \cite{piech2015deep}. We have open sourced \textsc{pyKT} and our experimental results at \url{https://pykt.org/}. We welcome contributions from other research groups and practitioners.

Crowdsourcing has proven to be a cost-effective way to meet the demands for labeled training data in supervised deep learning models. However, crowdsourced labels are often inconsistent and noisy due to cognitive and expertise differences among crowd workers. Existing approaches either infer latent true labels from noisy crowdsourced labels or learn a discriminative model directly from the crowdsourced labeled data, assuming the latent true label distribution is class-balanced. Unfortunately, in many real-world applications, the true label distribution typically is imbalanced across classes. Therefore, in this paper, we address the problem of learning from crowdsourced labeled data with an imbalanced true label distribution. We propose a new framework, named “Learning from Imbalanced Crowdsourced Labeled Data” (ICED), which simultaneously infers true labels from imbalanced crowdsourced labeled data and achieves high accuracy on downstream tasks such as classification. The ICED framework consists of two modules, i.e., a true label inference module and a synthetic data generation module, that augment each other iteratively. Extensive experiments conducted on both synthetic and real-world datasets demonstrate the effectiveness of the ICED framework. Our code is available at https://github.com/wentao-repo/ICED.

Recently we have seen a rapid rise in the amount of education data available through the digitization of education. This huge amount of education data usually exhibits in a mixture form of images, videos, speech, texts, etc. It is crucial to consider data from different modalities to build successful applications in AI in education (AIED). This talk targets AI researchers and practitioners who are interested in applying state-of-the-art multimodal machine learning techniques to tackle some of the hard-core AIED tasks. These include tasks such as automatic short answer grading, student assessment, class quality assurance, knowledge tracing, etc.

Representation learning has been proven to play an important role in the unprecedented success of machine learning models in numerous tasks, such as machine translation, face recognition and recommendation. The majority of existing representation learning approaches often require a large number of consistent and noise-free labels. However, due to various reasons such as budget constraints and privacy concerns, labels are very limited in many real-world scenarios. Directly applying standard representation learning approaches on small labeled data sets will easily run into over-fitting problems and lead to sub-optimal solutions. Even worse, in some domains such as education, the limited labels are usually annotated by multiple workers with diverse expertise, which yields noises and inconsistency in such crowdsourcing settings. In this paper, we propose a novel framework which aims to learn effective representations from limited data with crowdsourced labels. Specifically, we design a grouping based deep neural network to learn embeddings from a limited number of training samples and present a Bayesian confidence estimator to capture the inconsistency among crowdsourced labels. Furthermore, to expedite the training process, we develop a hard example selection procedure to adaptively pick up training examples that are misclassified by the model. Extensive experiments conducted on three real-world data sets demonstrate the superiority of our framework on learning representations from limited data with crowdsourced labels, comparing with various state-of-the-art baselines. In addition, we provide a comprehensive analysis on each of the main components of our proposed framework and also introduce the promising results it achieved in our real production to fully understand the proposed framework. To encourage reproducible results, we make our code available online athttps://github.com/tal-ai/RECLE.

In this paper, we propose an online clustering method called Contrastive Clustering (CC) which explicitly performs the instance- and cluster-level contrastive learning. To be specific, for a given dataset, the positive and negative instance pairs are constructed through data augmentations and then projected into a feature space. Therein, the instance- and cluster-level contrastive learning are respectively conducted in the row and column space by maximizing the similarities of positive pairs while minimizing those of negative ones. Our key observation is that the rows of the feature matrix could be regarded as soft labels of instances, and accordingly the columns could be further regarded as cluster representations. By simultaneously optimizing the instance- and cluster-level contrastive loss, the model jointly learns representations and cluster assignments in an end-to-end manner. Besides, the proposed method could timely compute the cluster assignment for each individual, even when the data is presented in streams. Extensive experimental results show that CC remarkably outperforms 17 competitive clustering methods on six challenging image benchmarks. In particular, CC achieves an NMI of 0.705 (0.431) on the CIFAR-10 (CIFAR-100) dataset, which is an up to 19% (39%) performance improvement compared with the best baseline. The code is available at https://github.com/XLearning-SCU/2021-AAAI-CC.

Jian Guan, Xiaoxi Mao, Changjie Fan, Zitao Liu, Wenbiao Ding, Minlie Huang. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Jian Guan, Zhexin Zhang, Zhuoer Feng, Zitao Liu, Wenbiao Ding, Xiaoxi Mao, Changjie Fan, Minlie Huang. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

In this paper, we study two challenging problems in incomplete multi-view clustering analysis, namely, i) how to learn an informative and consistent representation among different views without the help of labels and ii) how to recover the missing views from data. To this end, we propose a novel objective that incorporates representation learning and data recovery into a unified framework from the view of information theory. To be specific, the informative and consistent representation is learned by maximizing the mutual information across different views through contrastive learning, and the missing views are recovered by minimizing the conditional entropy of different views through dual prediction. To the best of our knowledge, this could be the first work to provide a theoretical framework that unifies the consistent representation learning and cross-view data recovery. Extensive experimental results show the proposed method remarkably outperforms 10 competitive multi-view clustering methods on four challenging datasets. The code is available at https://pengxi.me.

In real-world applications, it is common that only a portion of data is aligned across views due to spatial, temporal, or spatiotemporal asynchronism, thus leading to the so-called Partially View-aligned Problem (PVP). To solve such a less-touched problem without the help of labels, we propose simultaneously learning representation and aligning data using a noise-robust contrastive loss. In brief, for each sample from one view, our method aims to identify its within-category counterparts from other views, and thus the cross-view correspondence could be established. As the contrastive learning needs data pairs as input, we construct positive pairs using the known correspondences and negative pairs using random sampling. To alleviate or even eliminate the influence of the false negatives caused by random sampling, we propose a noise-robust contrastive loss that could adaptively prevent the false negatives from dominating the network optimization. To the best of our knowledge, this could be the first successful attempt of enabling contrastive learning robust to noisy labels. In fact, this work might remarkably enrich the learning paradigm with noisy labels. More specifically, the traditional noisy labels are defined as incorrect annotations for the supervised tasks such as classification. In contrast, this work proposes that the view correspondence might be false, which is remarkably different from the widely-accepted definition of noisy label. Extensive experiments show the promising performance of our method comparing with 10 state-of-the-art multi-view approaches in the clustering and classification tasks. The code will be publicly released at https://pengxi.me.

Existing approaches for audio-language taskspecific prediction focus on building complicated late-fusion mechanisms.However, these models face challenges of overfitting with limited labels and poor generalization.In this paper, we present a Cross-modal Transformer for Audio-and-Language, i.e., CTAL, which aims to learn the intra-and inter-modalities connections between audio and language through two proxy tasks from a large number of audio-and-language pairs: masked language modeling and masked cross-modal acoustic modeling.After fine-tuning our CTAL model on multiple downstream audioand-language tasks, we observe significant improvements on different tasks, including emotion classification, sentiment analysis, and speaker verification.Furthermore, we design a fusion mechanism in the fine-tuning phase, which allows CTAL to achieve better performance.Lastly, we conduct detailed ablation studies to demonstrate that both our novel cross-modality fusion component and audiolanguage pre-training methods contribute to the promising results.The code and pretrained models are available at https:// github.com/tal-ai/CTAL_EMNLP2021.

There is an increasing interest in the use of mathematical word problem (MWP) generation in educational assessment.Different from standard natural question generation, MWP generation needs to maintain the underlying mathematical operations between quantities and variables, while at the same time ensuring the relevance between the output and the given topic.To address above problem, we develop an end-to-end neural model to generate diverse MWPs in real-world scenarios from commonsense knowledge graph and equations.The proposed model (1) learns both representations from edge-enhanced Levi graphs of symbolic equations and commonsense knowledge; (2) automatically fuses equation and commonsense knowledge information via a self-planning module when generating the MWPs.Experiments on an educational gold-standard set and a large-scale generated MWP set show that our approach is superior on the MWP generation task, and it outperforms the SOTA models in terms of both automatic evaluation metrics, i.e., BLEU-4, ROUGE-L, Self-BLEU, and human evaluation metrics, i.e., equation relevance, topic relevance, and language coherence.To encourage reproducible results, we make our code and MWP dataset public available at https:// github.com/tal-ai/MaKE_EMNLP2021.

Data labelling is very important in many database and machine learning applications. Traditional methods rely on humans (workers or experts) to acquire labels. However, the human cost is rather expensive for a large dataset. Active learning based methods only label a small set of data with large uncertainty, train a model on these labelled data, and use the trained model to label the remainder unlabelled data. However they have two limitations. First, they cannot judiciously select appropriate data (task selection) and assign the tasks to proper humans (task assignment). Moreover, they independently process task selection and task assignment, which cannot capture the correlation between them. Second, they simply infer the truth of a task based on the answers from humans and the trained model (truth inference) by independently modeling humans and models. In other words, they ignore the correlation between them (the labelled data may have noise caused by humans with biases, and the model trained by the noisy labels may bring additional biases), and thus lead to poor inference results. To address these limitations, in this paper, we propose CrowdRL, an end-to-end reinforcement learning (RL) based framework for data labelling. To the best of our knowledge, CrowdRL is the first RL framework designed for the data labelling workflow by seamlessly integrating task selection, task assignment and truth inference together. CrowdRL fully utilizes the power of heterogeneous annotators (experts and crowdsourcing workers) and machine learning models together to infer the truth, which highly improves the quality of data labelling. CrowdRL uses RL to model task assignment and task selection, and designs an agent to judiciously assign tasks to appropriate workers. CrowdRL jointly models the answers of workers, experts and models, and designs a joint inference model to infer the truths. Experimental results on real datasets show that CrowdRL outperforms state-of-the-art approaches with the same (even fewer) monetary cost while achieving 5%-20% higher accuracy.

Speech emotion recognition is a challenging task because the emotion expression is complex, multimodal and fine-grained. In this paper, we propose a novel multimodal deep learning approach to perform fine-grained emotion recognition from real-life speeches. We design a temporal alignment mean-max pooling mechanism to capture the subtle and fine-grained emotions implied in every utterance. In addition, we propose a cross modality excitement module to conduct sample-specific adjustment on cross modality embeddings and adaptively recalibrate the corresponding values by its aligned latent features from the other modality. Our proposed model is evaluated on two well-known real-world speech emotion recognition datasets. The results demonstrate that our approach is superior on the prediction tasks for multimodal speech utterances, and it outperforms a wide range of baselines in terms of prediction accuracy. Further more, we conduct detailed ablation studies to show that our temporal alignment mean-max pooling mechanism and cross modality excitement significantly contribute to the promising results. In order to encourage the research reproducibility, we make the code publicly available at \url{https://github.com/tal-ai/FG_CME.git}.

Graph neural networks (GNNs) have shown the power in graph representation learning for numerous tasks. In this work, we discover an interesting phenomenon that although residual connections in the message passing of GNNs help improve the performance, they immensely amplify GNNs' vulnerability against abnormal node features. This is undesirable because in real-world applications, node features in graphs could often be abnormal such as being naturally noisy or adversarially manipulated. We analyze possible reasons to understand this phenomenon and aim to design GNNs with stronger resilience to abnormal features. Our understandings motivate us to propose and derive a simple, efficient, interpretable, and adaptive message passing scheme, leading to a novel GNN with Adaptive Residual, AirGNN. Extensive experiments under various abnormal feature scenarios demonstrate the effectiveness of the proposed algorithm.

Graph Neural Networks (GNNs) have achieved tremendous success in various real-world applications due to their strong ability in graph representation learning. GNNs explore the graph structure and node features by aggregating and transforming information within node neighborhoods. However, through theoretical and empirical analysis, we reveal that the aggregation process of GNNs tends to destroy node similarity in the original feature space. There are many scenarios where node similarity plays a crucial role. Thus, it has motivated the proposed framework SimP-GCN that can effectively and efficiently preserve node similarity while exploiting graph structure. Specifically, to balance information from graph structure and node features, we propose a feature similarity preserving aggregation which adaptively integrates graph structure and node features. Furthermore, we employ self-supervised learning to explicitly capture the complex feature similarity and dissimilarity relations between nodes. We validate the effectiveness of SimP-GCN on seven benchmark datasets including three assortative and four disassorative graphs. The results demonstrate that SimP-GCN outperforms representative baselines. Further probe shows various advantages of the proposed framework. The implementation of SimP-GCN is available at https://github.com/ChandlerBang/SimP-GCN.

The failure of landing a job for college students could cause serious social consequences such as drunkenness and suicide. In addition to academic performance, unconscious biases can become one key obstacle for hunting jobs for graduating students. Thus, it is necessary to understand these unconscious biases so that we can help these students at an early stage with more personalized intervention. In this paper, we develop a framework, i.e., MAYA (Multi-mAjor emploYment stAtus) to predict students' employment status while considering biases. The framework consists of four major components. Firstly, we solve the heterogeneity of student courses by embedding academic performance into a unified space. Then, we apply a generative adversarial network (GAN) to overcome the class imbalance problem. Thirdly, we adopt Long Short-Term Memory (LSTM) with a novel dropout mechanism to comprehensively capture sequential information among semesters. Finally, we design a bias-based regularization to capture the job market biases. We conduct extensive experiments on a large-scale educational dataset and the results demonstrate the effectiveness of our prediction framework.

Robust language processing systems are becoming increasingly important given the recent awareness of dangerous situations where brittle machine learning models can be easily broken with the presence of noises. In this paper, we introduce a robust word recognition framework that captures multi-level sequential dependencies in noised sentences. The proposed framework employs a sequence-to-sequence model over characters of each word, whose output is given to a word-level bi-directional recurrent neural network. We conduct extensive experiments to verify the effectiveness of the framework. The results show that the proposed framework outperforms state-of-the-art methods by a large margin and they also suggest that character-level dependencies can play an important role in word recognition. The code of the proposed framework and the major experiments are publicly available1.

Recently there are increasing concerns about the fairness of Artificial Intelligence (AI) in realworld applications such as computer vision and recommendations.For example, recognition algorithms in computer vision are unfair to black people such as poorly detecting their faces and inappropriately identifying them as "gorillas".As one crucial application of AI, dialogue systems have been extensively applied in our society.They are usually built with real human conversational data; thus they could inherit some fairness issues which are held in the real world.However, the fairness of dialogue systems has not been well investigated.In this paper, we perform a pioneering study about the fairness issues in dialogue systems.In particular, we construct a benchmark dataset and propose quantitative measures to understand fairness in dialogue models.Our studies demonstrate that popular dialogue models show significant prejudice towards different genders and races.Besides, to mitigate the bias in dialogue systems, we propose two simple but effective debiasing methods.Experiments show that our methods can reduce the bias in dialogue systems significantly.The dataset and the implementation are released to foster fairness research in dialogue systems 1 .

The automatic feedback of school assignments is an important application of AI in education.In this work, we focus on the task of personalized multimodal feedback generation, which aims to generate personalized feedback for teachers to evaluate students' assignments involving multimodal inputs such as images, audios, and texts.This task involves the representation and fusion of multimodal information and natural language generation, which presents the challenges from three aspects: (1) how to encode and integrate multimodal inputs; (2) how to generate feedback specific to each modality; and (3) how to fulfill personalized feedback generation.In this paper, we propose a novel Personalized Multimodal Feedback Generation Network (PMFGN) armed with a modality gate mechanism and a personalized bias mechanism to address these challenges.Extensive experiments on real-world K-12 education data show that our model significantly outperforms baselines by generating more accurate and diverse feedback.In addition, detailed ablation experiments are conducted to deepen our understanding of the proposed framework.

Numerous sequence learning models have been proposed to capture the sequential dependencies and have achieved remarkable success in a wide range of applications. In addition to sequential dependencies, real sequential data also exhibits side dependencies. For example, in session-based recommender systems, items are naturally related because of their intrinsic attributes such as brand, category, and function; and in document modeling, words are inherently related since they can share similar syntactic functions. Intuitively, such side dependencies provide rich information beyond the sequence (or sequential dependencies); thus they have great potential to advance traditional sequence modeling. However, research on exploring side dependencies for sequence learning is rather limited. In this paper, we study the problem of sequence learning with side dependencies. In particular, we propose a novel sequence learning framework SEE, which can simultaneously capture both sequential and side dependencies. Extensive experiments on real sequential data demonstrate the effectiveness of the proposed framework SEE and the advantages of integrating side dependencies.

Dialogue systems play an increasingly important role in various aspects of our daily life.It is evident from recent research that dialogue systems trained on human conversation data are biased.In particular, they can produce responses that reflect people's gender prejudice.Many debiasing methods have been developed for various NLP tasks, such as word embedding.However, they are not directly applicable to dialogue systems because they are likely to force dialogue models to generate similar responses for different genders.This greatly degrades the diversity of the generated responses and immensely hurts the performance of the dialogue models.In this paper, we propose a novel adversarial learning framework Debiased-Chat to train dialogue models free from gender bias while keeping their performance.Extensive experiments on two real-world conversation datasets show that our framework significantly reduces gender bias in dialogue models while maintaining the response quality.The implementation of the proposed framework is released 1 .

Classroom activity detection (CAD) focuses on accurately classifying whether the teacher or student is speaking and recording both the length of individual utterances during a class. A CAD solution helps teachers get instant feedback on their pedagogical instructions. This greatly improves educators’ teaching skills and hence leads to students’ achievement. However, CAD is very challenging because (1) the CAD model needs to be generalized well enough for different teachers and students; (2) data from both vocal and language modalities has to be wisely fused so that they can be complementary; and (3) the solution shouldn’t heavily rely on additional recording device. In this paper, we address the above challenges by using a novel attention based neural framework. Our framework not only extracts both speech and language information, but utilizes attention mechanism to capture long-term semantic dependence. Our framework is device-free and is able to take any classroom recording as input. The proposed CAD learning framework is evaluated in two real-world education applications. The experimental results demonstrate the benefits of our approach on learning attention based neural network from classroom data with different modalities, and show our approach is able to outperform state-of-the-art baselines in terms of various evaluation metrics.

Positive and unlabeled (PU) learning aims to obtain a well-performed classifier via an incomplete binary training set, in which only a part of labels of one category is known while the rest are unknown. However, in many real-world applications such as image recognition, the collected data samples often involve more than two categories. Moreover, only a small portion of the collected samples might have associated labels due to some practical reasons, and these labeled samples cannot always cover all the categories. We refer to this type of data as incomplete labeled data. In this paper, we first formally define the incomplete labeled data learning problem and then aim to tackle it via adversarial data generation. Specifically, we propose a novel generative framework LILA, which can produce synthetic labeled samples for both partially labeled categories and unlabeled categories. To enforce that the generated samples for unlabeled categories can associate with correct labels, we integrate two active learning processes into the LILA framework for selecting unlabeled samples in the collected sample set to query their labels effectively. After LILA has been well trained, a classifier can be trained on the balanced augmented data set consisting of both generated and original labeled samples. Extensive experiments on real image data demonstrate the effectiveness of our proposed framework. We release the implementation of the proposed framework via https://github.com/wentao-repo/LILA.

Recently we have seen a rapid rise in the amount of education data available through the digitization of education. This huge amount of education data usually exhibits in a mixture form of images, videos, speech, texts, etc. It is crucial to consider data from different modalities to build successful applications in AI in education (AIED). This tutorial targets AI researchers and practitioners who are interested in applying state-of-the-art multimodal machine learning techniques to tackle some of the hard-core AIED tasks. These include tasks such as automatic short answer grading, student assessment, class quality assurance, knowledge tracing, etc.

Multi-scale neural networks have shown effectiveness in image restoration tasks, which are usually designed and integrated in a handcrafted manner. Different from the existing labor-intensive handcrafted architecture design paradigms, we present a novel method, termed as multi-sCaLe nEural ARchitecture sEarch for image Restoration (CLEARER), which is a specifically designed neural architecture search (NAS) for image restoration. Our contributions are twofold. On one hand, we design a multi-scale search space that consists of three task-flexible modules. Namely, 1) Parallel module that connects multi-resolution neural blocks in parallel, while preserving the channels and spatial-resolution in each neural block, 2) Transition module remains the existing multi-resolution features while extending them to a lower resolution, 3) Fusion module integrates multi-resolution features by passing the features of the parallel neural blocks to the current neural blocks. On the other hand, we present novel losses which could 1) balance the tradeoff between the model complexity and performance, which is highly expected to image restoration; and 2) relax the discrete architecture parameters into a continuous distribution which approximates to either 0 or 1. As a result, a differentiable strategy could be employed to search when to fuse or extract multi-resolution features, while the discretization issue faced by the gradient-based NAS could be alleviated. The proposed CLEARER could search a promising architecture in two GPU hours. Extensive experiments show the promising performance of our method comparing with nine image denoising methods and eight image deraining approaches in quantitative and qualitative evaluations. The codes are available at https://github.com/limit-scu.

In many real-world classification applications such as fake news detection, the training data can be extremely imbalanced, which brings challenges to existing classifiers as the majority classes dominate the loss functions of classifiers. Oversampling techniques such as SMOTE are effective approaches to tackle the class imbalance problem by producing more synthetic minority samples. Despite their success, the majority of existing oversampling methods only consider local data distributions when generating minority samples, which can result in noisy minority samples that do not fit global data distributions or interleave with majority classes. Hence, in this paper, we study the class imbalance problem by simultaneously exploring local and global data information since: (i) the local data distribution could give detailed information for generating minority samples; and (ii) the global data distribution could provide guidance to avoid generating outliers or samples that interleave with majority classes. Specifically, we propose a novel framework GL-GAN, which leverages the SMOTE method to explore local distribution in a learned latent space and employs GAN to capture the global information, so that synthetic minority samples can be generated under even extremely imbalanced scenarios. Experimental results on diverse real data sets demonstrate the effectiveness of our GL-GAN framework in producing realistic and discriminative minority samples for improving the classification performance of various classifiers on imbalanced training data. Our code is available at https://github.com/wentao-repo/GL-GAN.

Spoken language proficiency is critically important for children’s growth and personal development. Due to the limited and imbalanced educational resources in China, elementary students barely have chances to improve their oral language skills in classes. Verbal fluency tasks (VFTs) were invented to let the students practice their spoken language proficiency after school. VFTs are simple but concrete math related questions that ask students to not only report answers but speak out the entire thinking process. In spite of the great success of VFTs, they bring a heavy grading burden to elementary teachers. To alleviate this problem, we develop Dolphin, a spoken language proficiency assessment system for Chinese elementary education. Dolphin is able to automatically evaluate both phonological fluency and semantic relevance of students’ VFT answers. We conduct a wide range of offline and online experiments to demonstrate the effectiveness of Dolphin. In our offline experiments, we show that Dolphin improves both phonological fluency and semantic relevance evaluation performance when compared to state-of-the-art baselines on real-world educational data sets. In our online A/B experiments, we test Dolphin with 183 teachers from 2 major cities (Hangzhou and Xi’an) in China for 10 weeks and the results show that VFT assignments grading coverage is improved by 22%.

In this paper, we study two less-touched challenging problems in single image dehazing neural networks, namely, how to remove haze from a given image in an unsupervised and zeroshot manner. To the ends, we propose a novel method based on the idea of layer disentanglement by viewing a hazy image as the entanglement of several "simpler" layers, i.e., a hazy-free image layer, transmission map layer, and atmospheric light layer. The major advantages of the proposed ZID are two-fold. First, it is an unsupervised method that does not use any clean images including hazy-clean pairs as the ground-truth. Second, ZID is a "zero-shot" method, which just uses the observed single hazy image to perform learning and inference. In other words, it does not follow the conventional paradigm of training deep model on a large scale dataset. These two advantages enable our method to avoid the labor-intensive data collection and the domain shift issue of using the synthetic hazy images to address the real-world images. Extensive comparisons show the promising performance of our method compared with 15 approaches in the qualitative and quantitive evaluations. The source code could be found at www.pengxi.me.

Learning representation has been proven to be helpful in numerous machine learning tasks. The success of the majority of existing representation learning approaches often requires a large amount of consistent and noise-free labels. However, labels are not accessible in many real-world scenarios and they are usually annotated by the crowds. In practice, the crowdsourced labels are usually inconsistent among crowd workers given their diverse expertise and the number of crowdsourced labels is very limited. Thus, directly adopting crowdsourced labels for existing representation learning algorithms is inappropriate and suboptimal. In this paper, we investigate the above problem and propose a novel framework of Representation Learning with crowdsourced Labels, i.e., "RLL", which learns representation of data with crowdsourced labels by jointly and coherently solving the challenges introduced by limited and inconsistent labels. The proposed representation learning framework is evaluated in two real-world education applications. The experimental results demonstrate the benefits of our approach on learning representation from limited labeled data from the crowds, and show RLL is able to outperform state-of-the-art baselines. Moreover, detailed experiments are conducted on RLL to fully understand its key components and the corresponding performance.

In modern recommender systems, both users and items are associated with rich side information, which can help understand users and items. Such information is typically heterogeneous and can be roughly categorized into flat and hierarchical side information. While side information has been proved to be valuable, the majority of existing systems have exploited either only flat side information or only hierarchical side information due to the challenges brought by the heterogeneity. In this paper, we investigate the problem of exploiting heterogeneous side information for recommendations. Specifically, we propose a novel framework jointly captures flat and hierarchical side information with mathematical coherence. We demonstrate the effectiveness of the proposed framework via extensive experiments on various real-world datasets. Empirical results show that our approach is able to lead a significant performance gain over the state-of-the-art methods.

In this work, we focus on models and analysis of multivariate time series data that are organized in hierarchies. Such time series are referred to as hierarchical time series (HTS) and they are very common in business, management, energy consumption, social networks, or web traffic modeling and analysis. We propose a new flexible hierarchical forecasting framework, that takes advantage of the hierarchical relational structure to predict individual time series. Our new forecasting framework is able to (1) handle HTS modeling and forecasting problems; (2) make accurate forecasting for HTS with seasonal patterns; (3) incorporate various individual forecasting models and combine heuristics based on the HTS datasets' own characterization. The proposed framework is evaluated on a real-world web traffic data set. The results demonstrate that our approach is superior when applied to hierarchical web traffic prediction problems, and it outperforms alternative time series prediction models in terms of accuracy

User experience in modern content discovery applications critically depends on high-quality personalized recommendations. However, building systems that provide such recommendations presents a major challenge due to a massive pool of items, a large number of users, and requirements for recommendations to be responsive to user actions and generated on demand in real-time. Here we present Pixie, a scalable graph-based real-time recommender system that we developed and deployed at Pinterest. Given a set of user-specific pins as a query, Pixie selects in real-time from billions of possible pins those that are most related to the query. To generate recommendations, we develop Pixie Random Walk algorithm that utilizes the Pinterest object graph of 3 billion nodes and 17 billion edges. Experiments show that recommendations provided by Pixie lead up to 50% higher user engagement when compared to the previous Hadoop-based production system. Furthermore, we develop a graph pruning strategy at that leads to an additional 58% improvement in recommendations. Last, we discuss system aspects of Pixie, where a single server executes 1,200 recommendation requests per second with 60 millisecond latency. Today, systems backed by Pixie contribute to more than 80% of all user engagement on Pinterest.

Hierarchical time series (HTS) is a special class of multivariate time series where many related time series are organized in a hierarchical tree structure and they are consistent across hierarchy levels. HTS modeling is crucial and serves as the basis for business planning and management in many areas such as manufacturing inventory, energy and traffic management. However, due to machine failures, network disturbances or human maloperation, HTS data suffer from missing values across different hierarchical levels. In this paper, we study the missing value estimation problem under hierarchical web traffic settings, where the user-visit traffic are organized in various hierarchical structures, such as geographical structure and website structure. We develop an efficient algorithm, HTSImpute, to accurately estimate the missing value in multivariate noisy web traffic time series with specific hierarchical consistency in HTS settings. Our HTSImpute is able to (1) utilize the temporal dependence information within each individual time series, (2) exploit the intra-relations between time series through hierarchy, (3) guarantee the satisfaction of hierarchical consistency constraints. Results on three synthetic HTS datasets and three real-world hierarchical web traffic datasets demonstrate that our approach is able to provide more accurate and hierarchically consistent estimations than other baselines.