In large-scale sensor networks, Multivariate Time Series Classification (MTSC) is a pivotal task for identifying events dependent on longitudinal data at the edge. However, existing methods focus on neither the inherent ability of convolutional networks to perceive subsequence features, nor the prolonged processing pipeline and the model deployment overhead brought by the highly parameterized models. To resolve these difficulties, we present EdgeMTSC, a lightweight large-kernel ConvNet for MTSC, which naturally extracts and learns features of diverse subsequences. Specifically, a novel module named Inter-channel Message Passing-driven Kernel Block (IMP-KB) is proposed, which maintains a learnable correlation matrix to propagate and merge inter-channel messages, and fuses miscellaneous patterns learned by parallel conv kernels of different sizes. EdgeMTSC sequences two modules of different receptive fields to aggregate local features using small kernels and study long-term representation provided by large kernels, respectively. For inference parameter reduction and accelerating inference without performance loss, the conv blocks in IMP-KBs are structurally reparameterizable. The performance of our model (76.2%) is benchmarked on 26 UEA MTSC datasets and is superior to the SOTA model (MPTSNet, 75%). At the same time, EdgeMTSC uses the fewest parameters and achieves the minimum inference time, applicable on any machine (8 devices ranging from large-scale distributed AI computing servers to resource-constrained edge devices) and in any application scenario.

Multimodal Emotion Recognition in Conversation aims to identify emotions within a dialogue with multimodal data, including audio, visual, and textual features.While existing methods have made significant improvements, there are two fundamental limitations to be addressed.From the modality fusion perspective, current approaches treat all modalities as functionally equivalent during fusion, overlooking their distinct communicative roles and information capacities, in which text conveys explicit semantic meaning while audio provides paralinguistic cues.From the emotion label perspective, many works ignore the continuous structure of emotion characterized by psychological theory and apply uniform penalties regardless of affective proximity.To address these limitations, we propose EMART, EMotion-Wheel-Guided Audio-Referred Text Representation for ERC, specifically focusing on audio and text modalities.First, we propose a modality-aware fusion strategy capturing linguistic features from text as the primary source and audio as a complementary component.Secondly, we propose an emotion-wheelguided supervised contrastive loss to encode emotional proximity based on Russell's circumplex model.Experimental results on IEMO-CAP and MELD demonstrate outstanding performance.The code is available at: https: //github.com/DILAB-HYU/EMART.git.

Accurate thyroid nodule segmentation in ultrasound images is essential for effective diagnosis and treatment planning. While multitask learning has shown promise in improving segmentation performance, several challenges remain unresolved: (a) scarcity of labeled data, (b) lack of integration of domain-specific prior knowledge, and (c) limited robustness in real-world clinical scenarios. To address these issues, we propose SSMT-Net, a Semi-Supervised Multitask Transformer-based Network that leverages unlabeled data for an initial unsupervised pre-training phase. In the subsequent supervised phase, our model jointly optimizes thyroid nodule segmentation, thyroid gland segmentation, and nodule size estimation, effectively integrating both local and global contextual cues. This multitask formulation enables the model to generalize better and remain robust across variable clinical conditions. Evaluated on two public datasets, TN3K and DDTI, SSMT-Net sets a new benchmark in thyroid nodule segmentation, achieving up to 3.38% and 1.23% absolute improvements in IoU and DSC, respectively, compared to existing state-of-the-art methods. Our code is available at: https://github.com/Umar-Faroq/SSMT-Net.

Deep learning models often struggle to maintain performance when the training and testing data come from different distributions. Test-time adaptation (TTA) addresses this by adapting a pre-trained model to an unlabeled target domain under distribution shifts. A more challenging setting is open-set TTA (OSTTA), where the target domain may contain unknown samples outside the source classes. Existing OSTTA methods primarily detect and discard such unknowns, relying only on known samples for adaptation. In this work, we argue that unknown samples can also provide valuable cues for improving adaptation. We propose LU-OSTTA (learning from unknown for OSTTA), a simple yet effective framework that leverages both in-distribution and semantically useful out-of-distribution samples. Our approach introduces: (i) a class-conditioned dynamic energy threshold to separate OOD samples more reliably, (ii) an optimal transport–based pseudo-label refinement to mitigate noise under distribution shifts, and (iii) an adaptive prototype weighting strategy that emphasizes semantically aligned target samples while down-weighting harmful ones. Experimental results demonstrate that our LUOSTTA consistently outperforms state-of-the-art TTA and OSTTA methods, highlighting the benefits of utilizing rather than discarding unknown samples. Our code is available at: https://github.com/takihasan/LU-OSTTA.

Examining Drug-Drug Interactions (DDIs) is a pivotal element in the process of drug development. DDIs occur when one drug's properties are affected by the inclusion of other drugs. Detecting favorable DDIs has the potential to pave the way for creating and advancing innovative medications applicable in practical settings. However, existing DDI prediction models continue to face challenges related to generalization in extreme cases, robust feature extraction, and real-life application possibilities. We aim to address these challenges by leveraging the effectiveness of context-aware deep graph learning by introducing a novel framework named CADGL. Based on a customized variational graph autoencoder (VGAE), we capture critical structural and physio-chemical information using two context pre-processors for feature extraction from two different perspectives- local neighborhood and molecular context, in a heterogeneous graphical structure. Our customized VGAE consists of a graph encoder, a latent information encoder, and an MLP decoder. CADGL surpasses other state-of-the-art DDI prediction models, excelling in predicting clinically valuable novel DDIs, supported by rigorous case studies. Our code is available at: https://github.com/azminewasi/CADGL.

Modern vehicle navigation systems can greatly benefit from waypoint point-of-interest (POI) recommendation, which suggests personalized intermediate stops along a driving route. This paper defines the novel waypoint POI recommendation problem: given a starting point and a destination, recommend one or more personalized POIs to visit en route. This scenario (e.g., suggesting a lunch stop during a road trip) differs from the conventional ''next POI'' recommendation in that it infers waypoint POIs from only two (origin and destination) inputs and predicts multiple intermediate stops rather than a single next location. To solve this problem, we propose WayPOI, a novel recommender model for Waypoint POI suggestion based on hierarchical graph based contrastive learning (WayPOI). WayPOI constructs a hierarchical graph that captures both individual and group-level behavioral patterns of users and POIs, and it employs a contrastive learning strategy to learn effective user and POI representations from sparse data. Through experiments on real-world driving data provided by Hyundai as well as on three public datasets, we demonstrate that WayPOI significantly outperforms several recent POI recommendation models, even though these baselines were carefully re-formed and retrained to perform waypoint recommendation for a fair comparison. Our ablation study confirms the benefit of each proposed component.

Test-Time Adaptation (TTA) aims to adapt an unseen target domain utilizing the unlabeled target data using a pre-trained source model. Continual TTA is a more challenging paradigm that deals with non-stationary environments during the test data adaptation. Most existing continual TTA methods are based on pseudo-labeling, but often (1) rely on overconfident pseudo-labels and (2) remain unstable under continual distribution shifts leading to error accumulation and catastrophic forgetting. To tackle these limitations, we propose Neighbor-Filtration based Continual Test-Time Adaptation (NF-CTTA), a reliable and memory-aware adaptation framework that addresses these challenges. NF-CTTA first calibrates pseudo-labels using class-conditional calibration error to correct over/under-confidence of the model. To further ensure reliability, we introduce an OOD Neighbor Filtration technique that selects a subset of high-confidence samples based on entropy and neighbor similarity, ensuring consistency within the semantic neighborhood. Finally, we propose a priority-guided memory buffer that retains the most informative low-entropy samples for replay, mitigating catastrophic forgetting across evolving test distributions. Extensive experiments across multiple domain shift benchmarks demonstrate that NF-CTTA achieves superior performance and stability compared to existing TTA and CTTA methods. The code is available at: https://github.com/takihasan/NF-CTTA.

The olfaction is hardly mentioned in the studies of multi-modal Large Language Models (LLMs). This demo presents a prototypical framework to embody prevalent LLMs with smelling ability using a plug-and-play olfactory signal processing service. To this end, we collect a dataset on Korean beers by self-developed electronic noses (e-noses) and an open-source dataset. An olfaction-related question-answering corpus is also generated to fine-tune LLMs. A gas classification model is applied to identify the smelling liquor upon the e-nose data. We then adopt and fine-tune LLMs on the generated datasets. The results show that LLMs under this framework can interact with the environment by its `nose' and provide olfaction-related answers augmented by our dataset. To the best of our knowledge, this is the first work on embodying LLMs with artificial olfaction. We additionally deployed the gas classification model and the trained LLM in a simple web-based system to show the feasibility of our prototype. Our demo video can be found at: https://bit.ly/4j8x6ZY.

With the widespread adoption of IoT, wearable devices, and sensors, time series data from human subjects are significantly increasing in the healthcare domain. Due to the laborious nature of manual annotation in time series data and the requirement for human experts, self-supervised learning methods are attempted to alleviate the limited label situations. While existing self-supervised methods have been successful to achieve comparable performance to the fully supervised methods, there are still some limitations that need to be addressed, considering the nature of time series data from human subjects: In real-world clinical settings, data labels (e.g., sleep stages) are usually annotated by subject-level, and there is a substantial variation in patterns between subjects. Thus, a model should be designed to deal with not only the label scarcity but also subject-wise nature of data to ensure high performance in real-world scenarios. To mitigate these issues, we propose a novel self-supervised learning framework for human subject time series data: Subject-Aware Time Series Clustering (SA-TSC). In the unsupervised representation learning phase, SA-TSC adopts a subject-wise learning strategy rather than instance-wise learning which randomly samples data instances from different subjects within the batch during training. Specifically, we generate subject-graphs with our graph construction method based on Gumbel-Softmax and perform graph spectral clustering on each subject-graph. In addition, we utilize graph neural networks to capture dependencies between channels and design our own graph learning module motivated from self-supervised loss. Experimental results show the outstanding performance of our SA-TSC with the limited & subject-wise label setting, leading to its high applicability to the healthcare industry. The code is available at: https://github.com/DILAB-HYU/SA-TSC

Beyond the traditional CNN structure, we have recently witnessed lots of breakthroughs in computer vision architectures such as Vision Transformer, MLP-Mixer, SNN-MLP, and so on. However, many efforts in developing novel architectures for vision tasks are heavily focused on achieving powerful performances, and how to attain interpretability in a trained neural network remains an open question. Inspired by the imaginary system GLOM, we present HiLite : Hierarchical Level-implemented Architecture attaining Part-Whole Interpretability, where islands of identical vectors can provide unprecedented interpretability. In our column-like structure, each level is a layer of a part-whole hierarchy composed of multiple neurons, and the function to define the neural field along an image input patch is initialized as the level vector inside the model. We propose two-column networks (Top-Down (TD) and Bottom-Up (BU)) that allow inter-level communication between adjacent levels on a specific patch and propose Gated Consensus Attention to perform intra-level communication on different patches within the level. At each time step, the level vector and outputs from different networks are combined into a weighted sum and passed to the next step, and outputs from the final time step are utilized as representation vectors. Here, supervised contrastive learning is used to find the relationship of meaningful patches in each class, where negative examples contribute to preventing representation collapse between neighboring patches. HiLite shows a possibility of performance through a quantitative evaluation on four image classification datasets as well as two metrics for assessing representation quality and showcases the intrinsic interpretability by simply generating a visual cue. We believe that our work is a solid step towards novel research on neural architectures attaining interpretability.

Deep Neural Networks (DNNs) tend to perform poorly on unseen domains due to domain shifts. Domain Generalization (DG) aims to improve the performance on such scenarios by minimizing the distribution discrepancy between source domains. Among many studies, dropout-based DG approaches which remove domain-specific features have gained attention. However, they are limited in minimizing the upper bound of generalization risk because they do not explicitly consider the distribution discrepancy when discarding features. In this paper, we propose a novel Discrepancy-guided Channel Dropout (DgCD) for DG that explicitly derives the discrepancy between domains and drops the channels with significant distribution discrepancy. Given a training batch, we perform two ways of standardization: (1) based on the variance/mean of the batch (i.e., sampled from all source domains) and (2) based on the variance/mean of domain-wise samples in the batch. With the two normal distributions, we explicitly derive the discrepancy using KL-divergence and backpropagate it towards each channel. A channel with a higher contribution to the discrepancy is more likely to be dropped. Experimental results show the superiority of DgCD over the state-of-the-art DG baselines, demonstrating the effectiveness of our dropout strategy which is directly coupled to reducing the domain discrepancy. Our code is available at: https://github.com/gyeomo/DgCD

Recently, sentiment-aware pre-trained language models (PLMs) demonstrate impressive results in downstream sentiment analysis tasks. However, they neglect to evaluate the quality of their constructed sentiment representations; they just focus on improving the fine-tuning performance, which overshadows the representation quality. We argue that without guaranteeing the representation quality, their downstream performance can be highly dependent on the supervision of the fine-tuning data rather than representation quality. This problem would make them difficult to foray into other sentiment-related domains, especially where labeled data is scarce. We first propose Sentiment-guided Textual Similarity (SgTS), a novel metric for evaluating the quality of sentiment representations, which is designed based on the degree of equivalence in sentiment polarity between two sentences. We then propose SentiCSE, a novel Sentiment-aware Contrastive Sentence Embedding framework for constructing sentiment representations via combined word-level and sentence-level objectives, whose quality is guaranteed by SgTS. Qualitative and quantitative comparison with the previous sentiment-aware PLMs shows the superiority of our work. Our code is available at: https://github.com/nayohan/SentiCSE

Knowledge Graphs (KGs) have proven essential in information processing and reasoning applications because they link related entities and give context-rich information, supporting efficient information retrieval and knowledge discovery; presenting information flow in a very effective manner. Despite being widely used globally, Bangla is relatively underrepresented in KGs due to a lack of comprehensive datasets, encoders, NER (named entity recognition) models, POS (part-of-speech) taggers, and lemmatizers, hindering efficient information processing and reasoning applications in the language. Addressing the KG scarcity in Bengali, we propose BanglaAutoKG, a pioneering framework that is able to automatically construct Bengali KGs from any Bangla text. We utilize multilingual LLMs to understand various languages and correlate entities and relations universally. By employing a translation dictionary to identify English equivalents and extracting word features from pre-trained BERT models, we construct the foundational KG. To reduce noise and align word embeddings with our goal, we employ graph-based polynomial filters. Lastly, we implement a GNN-based semantic filter, which elevates contextual understanding and trims unnecessary edges, culminating in the formation of the definitive KG. Empirical findings and case studies demonstrate the universal effectiveness of our model, capable of autonomously constructing semantically enriched KGs from any text. Data and code are available here: https://github.com/azminewasi/BanglaAutoKG

Category-specific models are provenly valuable methods in 3D single object tracking (SOT) regardless of Siamese or motion-centric paradigms. However, such over-specialized model designs incur redundant parameters, thus limiting the broader applicability of 3D SOT task. This paper first introduces unified models that can simultaneously track objects across all categories using a single network with shared model parameters. Specifically, we propose to explicitly encode distinct attributes associated to different object categories, enabling the model to adapt to cross-category data. We find that the attribute variances of point cloud objects primarily occur from the varying size and shape (e.g., large and square vehicles v.s. small and slender humans). Based on this observation, we design a novel point set representation learning network inheriting transformer architecture, termed AdaFormer, which adaptively encodes the dynamically varying shape and size information from cross-category data in a unified manner. We further incorporate the size and shape prior derived from the known template targets into the model’s inputs and learning objective, facilitating the learning of unified representation. Equipped with such designs, we construct two category-unified models SiamCUT and MoCUT. Extensive experiments demonstrate that SiamCUT and MoCUT exhibit strong generalization and training stability. Furthermore, our category-unified models outperform the category-specific counterparts by a significant margin (e.g., on KITTI dataset, $\sim$12\% and $\sim$3\% performance gains on the Siamese and motion paradigms).

Hypergraphs are effective in learning high-order relationships between nodes, which naturally represent group interactions as hyperedges (i.e., arbitrary-sized subsets of nodes). However, most approaches currently used for learning hypergraph representations do not consider pairwise relationships between nodes. While high-order relationships provide insight into the general connections among nodes in a group, they do not reveal the pairwise relationships between individual nodes within that group. Considering that it is unlikely for all nodes in the same group to share identical relationships, we argue that considering pairwise relationships is a critical aspect. In this paper, we propose Multi-view Mixed Attention for Contrastive Learning (MMACL) to address the aforementioned problem. MMACL proposes Mixed-Attention, which blends high-order relationships derived from the hypergraph attention network and pairwise relationships derived from the graph attention network. Then, it performs node-level contrastive learning to the graph structure with different views learned at each layer to finally obtain an expressive node representation. Our extensive experimental results on several popular datasets validate the effectiveness of the proposed MMACL for hypergraph node classification. Our code is available at: https://github.com/JongsooLee-HYU/MMACL

Model explainability is one of the crucial ingredients for building trustable AI systems, especially in the applications requiring reliability such as automated driving and diagnosis. Many explainability methods have been studied in the literature. Among many others, this article focuses on a research line that tries to visually explain a pre-trained image classification model such as Convolutional Neural Network by discovering concepts learned by the model, which is so-called the concept-based explanation. Previous concept-based explanation methods rely on the human definition of concepts (e.g., the Broden dataset) or semantic segmentation techniques like Slic (Simple Linear Iterative Clustering). However, we argue that the concepts identified by those methods may show image parts which are more in line with a human perspective or cropped by a segmentation method, rather than purely reflect a model's own perspective. We propose Model-Oriented Concept Extraction (MOCE), a novel approach to extracting key concepts based solely on a model itself, thereby being able to capture its unique perspectives which are not affected by any external factors. Experimental results on various pre-trained models confirmed the advantages of extracting concepts by truly representing the model's point of view.

Recently, Graph Neural Networks (GNNs) have achieved state-of-the-art performance on the multivariate time-series anomaly detection task by learning relationships between variables (sensors). However, they show limitations in capturing temporal dependencies due to lack of sufficient consideration on the characteristics of time to their graph structure. Several studies constructed a time-oriented graph, where each node represents a timestamp within a certain sliding window, to model temporal dependencies, but they failed to learn the trend of changes in time-series. This paper proposes Dual time-oriented Graph ATtention networks (DuoGAT) that resolves the aforementioned problems. Unlike previous work that uses the simple complete undirected structure for time-oriented graphs, our work models directed graphs with weighted edges that only connect from prior events to posterior events, and the edges that connect nearby events are given higher weights. In addition, another time-oriented graph is used to model time series stationary via differencing, which especially focuses on capturing the series of changes. Empirically, our method outperformed the existing state-of-the-art work with the highest F1-score for the four real-world dataset while maintaining low training cost. We also proposed a novel explanation method for anomaly detection using DuoGAT, which provides time-oriented reasoning via hierarchically tracking time points critical in a specific anomaly detection. Our code is available at: https://github.com/ByeongtaePark/DuoGAT

The explosive popularity of smart manufacturing has caught the attention of researchers in terms of intelligent mold processing and management. Machining mold components is a crucial step in the mold production process for many industries, which creates (e.g., cutting, drilling, and shaping a metal) the individual parts (e.g., core pins, ejector pins, cavities, slides, and lifters) that make up a mold used in manufacturing. We present IMPsys, an AI-based system that automatically explores machining jobs, infers their processing time and schedules them on machines, given numerous 3D modelling files of mold components. Our demo video can be found at: http://bit.ly/3EeKnyL.

Time-series forecasting has been actively studied and adopted in various real-world domains. Recently there have been two research mainstreams in this area: building Transformer-based architectures such as Informer, Autoformer and Reformer, and developing time-series representation learning frameworks based on contrastive learning such as TS2Vec and CoST. Both efforts have greatly improved the performance of time series forecasting. In this paper, we investigate a novel direction towards improving the forecasting performance even more, which is orthogonal to the aforementioned mainstreams as a model-agnostic scheme. We focus on time stamp embeddings that has been less-focused in the literature. Our idea is simple-yet-effective: based on given current time stamp, we predict embeddings of its near future time stamp and utilize the predicted embeddings in the time-series (value) forecasting task. We believe that if such future time information can be previewed at the time of prediction, they can be utilized by any time-series forecasting models as useful additional information. Our experimental results confirmed that our method consistently and significantly improves the accuracy of the recent Transformer-based models and time-series representation learning frameworks. Our code is available at: https://github.com/sunsunmin/Look_Ahead

Collaborative Filtering (CF) methods for recommender systems commonly suffer from the data sparsity issue. Data imputation has been widely adopted to deal with this issue. However, existing studies have limitations in the sense that both uncertainty and robustness of imputation have not been taken into account, where there is a high risk that the imputed values are likely to be far from the true values. This paper explores a novel imputation framework, named Uncertainty-Aware Multiple Imputation (UA-MI), which can effectively solve the sparsity issue. Given a (sparse) user-item interaction matrix, our key idea is to quantify uncertainty on each missing entry and then the cells with the lowest uncertainty are selectively imputed. Here, we suggest three strategies for measuring uncertainty in missing user-item interactions, each of which is based on sampling, dropout, and ensemble, respectively. They successfully obtain element-wise mean and variance on the missing entries, where the variance helps determine where in the matrix should be imputed and the corresponding mean values are imputed. Experiments show that our UA-MI framework significantly outperformed the existing imputation strategies

This paper focuses on the problem of forecasting daily traffic of new roads, where very little data is available for prediction. We propose a novel prediction model based on Generative Adversarial Networks (GAN) that learns the subtle patterns of the changes in the traffic flow according to the various contextual factors. Then the trained generator makes a prediction via generating a realistic traffic flow data of a target new road given its weather and day type. Both the quantitative and qualitative results of our extensive experiments indicate the effectiveness of our method.

Sleep staging is an important task that enables sleep quality assessment and disorder diagnosis. Due to dependency on manually labeled data, many researches have turned from supervised approaches to self-supervised learning (SSL) for sleep staging. While existing SSL methods have made significant progress in terms of its comparable performance to supervised methods, there are still some limitations. Contrastive learning could potentially lead to false negative pair assignments in sleep signal data. Moreover, existing data augmentation techniques directly modify the original signal data, making it likely to lose important information. To mitigate these issues, we propose Self-Supervised Learning with Attention-aided Positive Pairs (SSLAPP). Instead of the contrastive learning, SSLAPP carefully draws high-quality positive pairs and exploits them in representation learning. Here, we propose attention-based latent signal augmentation, which plays a key role by capturing important features without losing valuable signal information. Experimental results show that our proposed method achieves state-of-the-art performance in sleep stage classification with limited labeled data. The code is available at: https://github.com/DILAB-HYU/SSLAPP

Triangular meshes have been actively used in computer graphics to represent 3D shapes. However, due to their non-uniform and irregular nature, learning such data with a Deep Neural Network is not straightforward. Transforming mesh data to simpler structures (e.g., voxel grids, point clouds, or multi-view 2D images) leads to other issues including spatial information loss and scalability. Traditional descriptors for mesh data simply extract hand-crafted features, which might not be effective in various environments. Several deep architectures that directly consume mesh data have been proposed, but their input features are still heuristic and unable to fully capture both geodesic and geometric characteristics of a mesh. In addition, their model architectures are not designed to be capable of providing visual explanations of their decision making.

Explainable recommendation has gained great attention in recent years. A lot of work in this research line has chosen to use the knowledge graphs (KG) where relations between entities can serve as explanations. However, existing studies have not considered sentiment on relations in KG, although there can be various types of sentiment on relations worth considering (e.g., a user's satisfaction on an item). In this paper, we propose a novel recommendation framework based on KG integrated with sentiment analysis for more accurate recommendation as well as more convincing explanations. To this end, we first construct a Sentiment-Aware Knowledge Graph (namely, SAKG) by analyzing reviews and ratings on items given by users. Then, we perform item recommendation and reasoning over SAKG through our proposed Sentiment-Aware Policy Learning (namely, SAPL) based on a reinforcement learning strategy. To enhance the explainability for end-users, we further developed an interactive user interface presenting textual explanations as well as a collection of reviews related with the discovered sentiment. Experimental results on three real-world datasets verified clear improvements on both the accuracy of recommendation and the quality of explanations.

In this demo, we show MANIAC, a MAN-machIne collaborative system for malware Author Classification. It is developed to fight a number of author groups who have been generating lots of new malwares by sharing source code within a group and exploiting evasive schemes such as polymorphism and metamorphism. Notably, MANIAC allows users to intervene in the model's classification of malware authors with high uncertainty. It also provides effective interfaces and visualizations with users to achieve maximum classification accuracy with minimum human labor.

For these few decades malwares have been posing a major concern in the cyber security. Recently, a number of "author groups" have been generating lots of newmalwares by sharing source code within a group and exploiting evasive schemes such as polymorphism and metamorphism. This motivates us to study the problem of identifying the author group of a given malware, which would be able to work for not only blocking malwares but also legally punishing suspected malware authors. In this paper, we propose a human-machine collaborative approach for classifying author groups of malwares accurately. We also propose a visualization method for helping human experts to make the decision easily. We verify the superiority of our framework through extensive experiments using real-world malware data.

Cold-start problems are arguably the biggest challenges faced by collaborative filtering (CF) used in recommender systems. When few ratings are available, CF models typically fail to provide satisfactory recommendations for cold-start users or to display cold-start items on users' top-N recommendation lists. Data imputation has been a popular choice to deal with such problems in the context of CF, filling empty ratings with inferred scores. Different from (and complementary to) data imputation, this paper presents AR-CF, which stands for Augmented Reality CF, a novel framework for addressing the cold-start problems by generating virtual, but plausible neighbors for cold-start users or items and augmenting them to the rating matrix as additional information for CF models. Notably, AR-CF not only directly tackles the cold-start problems, but is also effective in improving overall recommendation qualities. Via extensive experiments on real-world datasets, AR-CF is shown to (1) significantly improve the accuracy of recommendation for cold-start users, (2) provide a meaningful number of the cold-start items to display in top-N lists of users, and (3) achieve the best accuracy as well in the basic top-N recommendations, all of which are compared with recent state-of-the-art methods.

Summary Learning and analyzing graph data is one of the most fundamental research areas in machine learning and data mining. Among numerous graph‐based data structures, this paper focuses on a graph bag (simply, bag), which corresponds to a training object containing one or more graphs, and a label is available only for a bag. This type of a bag can represent various real‐world objects such as drugs, web pages, XML documents, and images, among many others, and there have been many researches on models for learning this type of bag data. Within this research context, we define a novel problem of dynamic graph bag classification, and propose an algorithm to solve this problem. Dynamic bag classification aims to build a classification model for bags, which are presented in a streaming fashion, ie, frequent emerging of new bags or graphs over time. Given such changes made to the bag dataset, our proposed algorithm aims to update incrementally the top‐m most discriminative features instead of searching for them from scratch. Incremental gSpan and incremental gScore are proposed as core parts of our algorithm to deal with a stream of bags efficiently. We evaluate our algorithm on two real‐world datasets in terms of both feature selection time and classification accuracy. The experimental results demonstrate that our algorithm derives an informative feature set much faster than the existing one originally designed for targeting static bag data, with little accuracy loss.

Generative Adversarial Networks (GAN) have not only achieved a big success in various generation tasks such as images, but also boosted the accuracy of classification tasks by generating additional labeled data, which is called data augmentation. In this paper, we propose a Rating Augmentation framework with GAN, named RAGAN, aiming to alleviate the data sparsity problem in collaborative filtering (CF), eventually improving recommendation accuracy significantly. We identify a unique challenge that arises when applying GAN to CF for rating augmentation: naive RAGAN tends to generate values biased towards high ratings. Then, we propose a refined version of RAGAN, named RAGANBT, which addresses this challenge successfully. Via our extensive experiments, we validate that our RAGANBT is really effective to solve the data sparsity problem, thereby providing existing CF models with great improvement in accuracy under various situations such as basic top-N recommendation, long-tail item recommendation, and recommendation to cold-start users.

Generative Adversarial Networks (GAN) have achieved big success in various domains such as image generation, music generation, and natural language generation. In this paper, we propose a novel GAN-based collaborative filtering (CF) framework to provide higher accuracy in recommendation. We first identify a fundamental problem of existing GAN-based methods in CF and highlight it quantitatively via a series of experiments. Next, we suggest a new direction of vector-wise adversarial training to solve the problem and propose our GAN-based CF framework, called CFGAN, based on the direction. We identify a unique challenge that arises when vector-wise adversarial training is employed in CF. We then propose three CF methods realized on top of our CFGAN that are able to address the challenge. Finally, via extensive experiments on real-world datasets, we validate that vector-wise adversarial training employed in CFGAN is really effective to solve the problem of existing GAN-based CF methods. Furthermore, we demonstrate that our proposed CF methods on CFGAN provide recommendation accuracy consistently and universally higher than those of the state-of-the-art recommenders.

C-Rank (contribution-based ranking) is a state-of-the-art algorithm for ranking web pages. It combines content and link information by introducing the concept of contribution, which implies how much a page contributes to improving the content quality of other pages. However, C-Rank suffers from taking infeasible time to reflect changes in World Wide Web to every page's C-Rank score. In this paper, we propose an effective and efficient method to incrementally maintain C-Rank scores of terms in web pages. Our proposed method is carefully designed to selectively update C-Rank scores of specific pages without any accuracy loss, rather than re-computing all the C-Rank scores of all the terms in all the pages from the scratch. Our experimental results on real-world dataset confirm the effectiveness and efficiency of our proposed method.

This paper proposes a novel birthmark, a dynamic API authority vector (DAAV), for detecting software theft. DAAV satisfies four essential requirements for good birthmarks--credibility, resiliency, scalability, and packing-free--while existing birthmarks fail to satisfy all of them together. In particular, existing static birthmarks are unable to handle the packed programs and existing dynamic birthmarks do not satisfy credibility and resiliency. Our experimental results demonstrate that DAAV provides satisfying credibility and resiliency compared with existing dynamic birthmarks and also can cover packed programs.

As plagiarism of software increases rapidly, there are growing needs for software plagiarism detection systems. In this paper, we propose a software plagiarism detection system using an API-labeled control flow graph (A-CFG) that abstracts the functionalities of a program. The A-CFG can reflect both the sequence and the frequency of APIs, while previous work rarely considers both of them together. To perform a scalable comparison of a pair of A-CFGs, we use random walk with restart (RWR) that computes an importance score for each node in a graph. By the RWR, we can generate a single score vector for an A-CFG and can also compare A-CFGs by comparing their score vectors. Extensive evaluations on a set of Windows applications demonstrate the effectiveness and the scalability of our proposed system compared with existing methods.