Deep research is a retrieval-intensive task that requires iteratively retrieving evidence, reading across sources, and synthesizing source-grounded outputs. In practice, real-world deep research applications are of high workloads that require generating massive reports or conducting large-scale literature surveys. Such applications increasingly require batch processing capabilities, which are missing from traditional chat-oriented agents, limiting throughput when processing large volumes of structurally similar jobs. We introduce C3Flow (Concurrent Claude Code Workflow), a framework that transforms Claude Code from an interactive assistant into a SIMD-style (Single Instruction, Multiple Data) concurrent compute engine. C³Flow treats each agent instance as an isolated, schedulable unit capable of handling declarative multi-step tasks, multi-model routing, and comprehensive trajectory logging. On BrowseComp-zh, C³Flow improves pass@1 from 48.44% to 61.59% and pass@3 from 70.24% to 77.51% compared to standard function calling, while reducing average latency. For multi-hop fact verification, C3Flow achieves a 5.9 speedup over human annotators while maintaining 87.5% accuracy, demonstrating its effectiveness for production-scale deep research pipelines. Code is available at~ https://github.com/RAGenius/C3Flow.

Retrieval-Augmented Generation (RAG) effectively grounds Large Language Models (LLMs) with external knowledge and is widely applied to Web-related tasks. However, its scalability is hindered by excessive context length and redundant retrievals. Recent research on soft context compression aims to address this by encoding long documents into compact embeddings, yet they often underperform non-compressed RAG due to their reliance on auto-encoder-like full-compression that forces the encoder to compress all document information regardless of relevance to the input query.

Community detection is a longstanding topic in graph and Web algorithms, and semi-supervised local community detection, identifying the community to which the given user belongs, garners increasing research attention in recent years. While achieving encouraging results, existing solutions often encounter accumulated errors due to the weak supervision in the community expansion process, and are undermined by the initial seed sensitivity that a suboptimal or boundary seed node can easily misguide community generation. To fill these gaps, we propose a fresh generative perspective on hyperbolic space, which recasts this problem as the seed-conditioned sequence generation, and reformulates community generation as a continuous transport of probability distributions in the manifold measure space. In this paper, we present a novel Semi-supervised Local community detection framework based on hyperbolic Flow Matching (SLFM). Specifically, it leverages a geometric-aware Seed Selector that refines initial seeds with hyperbolic angular and radial priors, and a Hyperbolic Flow Transporter that learns a vector field to map a source distribution to a target community distribution, generating a robust set of anchors. Finally, a Community Expander is introduced to utilize these anchors as surrogate supervision to effectively recover the full community. Experimental results on four real-world datasets demonstrate that SLFM significantly outperforms existing methods in both local and global semi-supervised settings.

Recent advancements in Large Language Models (LLMs) have significantly extended context windows, igniting discussions about the necessity of Retrieval-Augmented Generation (RAG). U-NIAH, a unified Needle-in-a-Haystack (NIAH) framework, systematically evaluates LLMs and RAG methods in controlled long-context settings. It extends beyond traditional NIAH by incorporating more practical and complex scenarios like multi-needle, long-needle, and needle-in-needle configurations and leveraging the synthetic dataset to mitigate LLM biases. The experiments aim to address three research questions in long-context scenarios: (1) performance tradeoffs between LLMs and RAG, (2) error patterns in RAG, and (3) RAG’s limitations in complex settings. Results show that smaller LLMs benefit more from RAG. In all settings, RAG achieves a win rate of 82.58% over direct answers. Additionally, it is found that retrieval noise and chunk ordering degrade RAG performance, and we further summarized typical error patterns, including omissions due to noise, hallucinations under high noise critical conditions, and self-doubt behaviors, as well as how these phenomena vary with context length. Finally, in some challenging scenarios, experiments show that deep reasoning models are more easily affected by distractors. These findings highlight the complementary roles of RAG and LLMs and offer actionable insights for optimizing deployment strategies ( https://github.com/Tongji-KGLLM/U-NIAH ).

Efficiently modeling user behavior on online payment platforms is crucial for accurately identifying potential financial risks. With the rapid growth of online payment platforms, the volume of user transaction data has significantly increased. Moreover, users' payment behaviors often encompass diverse activities and interactions across multiple scenarios. Based on observations from online payment platforms, we identify three key challenges: scarce labels and poor representation robustness, long user payment behavior sequences, and complex and heterogeneous amount-aware scenarios.

Explainable transaction risk analysis is a challenge for traditional deep learning models, which only predict suspicious transactions without explanations. Current explainable methods rely on hand-crafted rules and lack the ability to automatically generate language-based explanations. Large Language Models (LLMs) offer promise due to their reasoning and text generation abilities but struggle with domain knowledge and hallucinations, making risk analysis difficult. Specifically, LLMs face: (1) insufficient adaptation to transaction data analysis, and (2) ineffective knowledge retrieval methods that ignore the rich graph structure of transaction data. To address these issues, we propose the Dual Graph Retrieval-Augmented Generation (Dual-gRAG) framework, which utilizes dual retrieval: expert knowledge and reasoning case retrieval. Expert knowledge compensates for domain gaps, while reasoning case retrieval provides step-wise analysis guidance. We incorporate both graph-structured features and semantic features into the retrieval process to enhance the effectiveness of the retrieval. Extensive experiments show that Dual-gRAG improves LLMs' risk analysis capabilities, achieving a 15% increase in different metrics.

High Click-Through Rate (CTR) imagery has proven commercial value for food delivery platforms, driving a need for strategies to generate visually compelling images. Our investigations reveal a positive correlation between appropriate food backgrounds and subsequent user engagement. Despite advancements in diffusion models, inpainting new backgrounds does not guarantee high CTR, and fine-tuning diffusion models for this purpose is prohibitively expensive for the fast-paced online food delivery advertising sector. Consequently, there is a lack of cost-effective, transferable generation frameworks tailored to high-CTR food images. In this paper, we propose FoRAGe, a novel high-CTR Food image Retrieval-Augmented Generation pipeline leveraging ControlNet based on Stable Diffusion. Specifically, we construct a comprehensive food image database encompassing a diverse range of background environments. During image generation, FoRAGe retrieves high-quality background exemplars featuring analogous food subjects from the database and employs the retrieved backgrounds as conditions to guide image synthesis via the ControlNet model. Subsequently, a multimodal CTR prediction model is utilized to identify and select optimal images for deployment. Extensive online experiments demonstrate a significant increase in CTR for images generated by our proposed pipeline, and ablation studies further elucidate the impact of different strategies and configurations. Code is available at https://github.com/jiaxu-feng/FoRAGe.

Time series forecasting is a vital task with widespread applications.While recent advancements have adopted patching to enrich shortterm context, existing encoding methods often struggle to capture the diverse semantics within patches, resulting in semantic information loss and limited model performance.Moreover, most long-term dynamics modeling approaches rely on homogeneous architectures with fixed receptive fields, inevitably sacrificing either performance or efficiency.To address these challenges, we propose Mixture of Universals (MoU), a novel framework designed to prevent semantic loss during patch encoding and efficiently enhance long-term dynamics through a hybrid approach.Specifically, MoU is consist of two novel designs: Mixture of Feature Extractors (MoF) and Mixture of Architectures (MoA).MoF introduces a semantics-aware encoding mechanism that selectively activates the corresponding subextractor based on the semantic context of input patches, preserving diverse temporal patterns and mitigating information loss.MoA, on the other hand, hierarchically captures long-term dependency with progressively expanded receptive field, improving model performance while maintaining relatively low computational costs.We conducted extensive experiments on seven real-world datasets, and the results demonstrate the superiority of our model.Our Code is available at https://github.com/lunaaa95/mou/.

Risk scoring systems have been widely deployed in many applications, which assign risk scores to users according to their behavior sequences. Though many deep learning methods with sophisticated designs have achieved promising results, the black-box nature hinders their applications due to fairness, explainability, and compliance consideration. Rule-based systems are considered reliable in these sensitive scenarios. However, building a rule system is labor-intensive. Experts need to find informative statistics from user behavior sequences, design rules based on statistics and assign weights to each rule. In this paper, we bridge the gap between effective but black-box models and transparent rule models. We propose a two-stage framework, CauseRuDi, that distills the knowledge of black-box teacher models into rule-based student models. We design a Monte Carlo tree search-based statistics generation method that maximizes the correlation or dependence between the generated statistics and the teacher model's outputs. We formulate a sequential move game and a simultaneous move coalitional game to generate multiple statistics. Then statistics are composed into logical rules with our proposed neural logical networks by mimicking the outputs of teacher models. We evaluate CauseRuDi on three real-world public datasets and an industrial dataset to demonstrate its effectiveness.

Discrete-Time Dynamic Graphs (DTDGs), which are prevalent in real-world implementations and notable for their ease of data acquisition, have garnered considerable attention from both academic researchers and industry practitioners. The representation learning of DTDGs has been extensively applied to model the dynamics of temporally changing entities and their evolving connections. Currently, DTDG representation learning predominantly relies on GNN+RNN architectures, which manifest the inherent limitations of both Graph Neural Networks (GNNs) and Recurrent Neural Networks (RNNs). GNNs suffer from the over-smoothing issue as the models architecture goes deeper, while RNNs struggle to capture long-term dependencies effectively. GNN+RNN architectures also grapple with scaling to large graph sizes and long sequences. Additionally, these methods often compute node representations separately and focus solely on individual node characteristics, thereby overlooking the behavior intersections between the two nodes whose link is being predicted, such as instances where the two nodes appear together in the same context or share common neighbors.

Large Language Models (LLMs) have shown impressive performance in various domains, prompting researchers to explore their potential application in recommendation systems. However, directly applying LLMs to recommendation tasks has proven to be less effective due to the significant gap between the data used for pre-training LLMs and the specific requirements of recommendation tasks. In this study, we propose Direct Multi-Preference Optimization (DMPO), a streamlined framework to bridge this gap and enhance the alignment of LLMs for recommendation tasks. DMPO can be viewed as a pair-wise ranking loss to distinguish between positive and negative samples in recommendation tasks. Furthermore, DMPO improves the performance of LLM-based recommenders by maximizing the probability of positive samples and minimizing the probability of multiple negative samples at the same time. Experimental evaluations are conducted to compare DMPO with traditional recommendation methods and other LLM-based recommendation methods. The results reveal that DMPO significantly enhances the recommendation capabilities of LLMs across three real-world public datasets in few-shot scenarios. Furthermore, the experiments also demonstrate that DMPO exhibits superior generalization ability in cross-domain recommendation. A case study elucidates the reasons behind these consistent improvements and also underscores DMPO's potential as an explainable recommendation system. Our code and data are available at https://github.com/BZX667/DMPO.

Urban undisciplined events (UUE) are of increasing concern to urban officials because they reduce the quality of life and cause societal disorder. How to accurately predict future occurrences is a key point in preventing these events. However, existing supervised methods struggle to perform well on sparse UUEs while self-supervised MAE-based methods adopt a traditional random masking strategy which leads to limited performance on UUE forecasting. Fortunately, we have designed an innovative spatiotemporal masking strategy and its corresponding pre-training task called Masked Spatio-Temporal Event Series Modeling (MSTEM). Through Cluster-assisted region masking, MSTEM efficiently distributes masked regions evenly among different clusters, enhancing the model's ability to capture spatial correlation and heterogeneity while addressing sparse region distribution of UUEs. Frequency-enhanced patch masking helps the model to sufficiently extract the temporal features of UUEs by reconstructing multiple views. Additionally, we propose future merge and cluster label modeling to enhance the extraction of spatiotemporal dependencies, thereby improving the performance of MSTEM on downstream prediction tasks. Experimental evaluations on four real-world datasets including crimes and disorderly conduct show that our masked autoencoder with MSTEM outperforms most of the state-of-the-art baselines.

As online transactions rapidly increase, money laundering has become more difficult to detect, rendering traditional rule-based algorithms inadequate for the current severe laundering landscape. Although efforts have been made to model user behavior sequences for detecting money laundering, these approaches still fall short in scenarios with extremely low anomaly rates. In our anti-money laundering practices, we have identified the following three challenges: weak perception of intensity, scarce labels, poor representation robustness. In this paper, we present CLeAR, a novel robust sequence-based self-supervised Representation Learning framework for Anti-Money Laundering. To address the weak perception of intensity, we devise an Intensity-Aware Transformer to better capture the nuances of user behavior sequences. By introducing sequence-based Contrastive Learning into this task, we effectively tackle the issue of scarce labels and enhance sequence modeling. Additionally, we developed two self-supervised learning tasks-next behavior matching and sub-sequence matching-that significantly enhance the overall robustness of representation. After rigorous experiments across datasets of various scales, CLeAR consistently delivers exceptional performance, even under the extremely low anomaly rates that closely mimic real-world conditions.

Drug combinations can cause adverse drug-drug interactions(DDIs). Identifying specific effects is crucial for developing safer therapies. Previous works on DDI event prediction have typically been limited to using labels of specific events as supervision, which renders them insufficient to address two significant challenges: (1) the bias caused by highly imbalanced event distribution where certain interaction types are vastly underrepresented. (2) the scarcity of labeled data for rare events, a pervasive issue where rare yet potentially critical interactions are often overlooked or under-explored due to limited available data. In response, we offer "DDIPrompt", an innovative solution inspired by the recent advancements in graph prompt learning. Our framework aims to address these issues by leveraging the intrinsic knowledge from pre-trained models, which can be efficiently deployed with minimal downstream data. Specifically, to solve the first challenge, DDIPrompt features a hierarchical pre-training strategy to foster a generalized and comprehensive understanding of drug properties. It captures intra-molecular structures through augmented links based on structural proximity between drugs, further learns inter-molecular interactions emphasizing edge connections rather than concrete catagories. For the second challenge, we implement a prototype-enhanced prompting mechanism during inference. This mechanism, refined by few-shot examples from each category, effectively harnesses the rich pre-training knowledge to enhance prediction accuracy, particularly for these rare but crucial interactions. Comprehensive evaluations on two benchmark datasets demonstrate DDIPrompt's SOTA performance, especially for those rare DDI events.

Community detection plays a pivotal role in network analysis, with applications in recommendation systems, anomaly detection, and biochemistry. However, traditional methods, while computationally efficient, often fall short in managing the complexities of real-world network structures. In contrast, deep learning approaches enhance accuracy but require substantial computational resources and task-specific architectures. This paper introduce GetCom, a novel three-phase "pre-train, generate, prompt" framework that integrates traditional methods and deep learning techniques. In the pre-training phase, GetCom acquires comprehensive understanding of community structures, which provides a solid foundation for the subsequent phases. During the generation phase, traditional community detection methods are employed to efficiently identify potential communities, which are subsequently refined in the prompt learning phase. This integration offers an efficient, accurate, and generalizable solution for community detection. Experiments on five real-world network datasets demonstrate that GetCom achieves state-of-the-art performance, with strong efficiency and generalization capabilities across diverse datasets and tasks.

Accurately predicting Vehicle Energy Consumption (VEC) is crucial for estimating a vehicle's total energy requirements along a predetermined trajectory. Current research mainly focuses on personalized models that enhance VEC prediction accuracy by leveraging driving behavior features extracted from historical trajectory data. However, there are still two significant limitations. First, existing algorithms predominantly model trajectories with coarse granularity, focusing solely on overall characteristics and neglecting the crucial interplay between vehicles, drivers, and the environments, which fundamentally shape trajectory dynamics. Second, current models predict driver behavior preferences solely from vehicle operational states in historical trajectories, often overlooking the influence of external environmental factors. To overcome these limitations, we introduce a Spatial-Temporal Framework for Energy Consumption Prediction of Vehicle Trajectories (ST-ECP). Specifically, we construct a heterogeneous interaction graph that captures the complex relationships between vehicles, environments, and drivers, effectively characterizing the dynamic attributes of trajectories across various conditions. Additionally, we design a personalized pattern aggregation module to extract personalized driving behavior features. Extensive experimental on real-world datasets demonstrate the effectiveness and efficiency of ST-ECP.

Time series forecasting (TSF) consists of point prediction and probabilistic forecasting. Unlike point forecasting which predicts an expected value of a future target, probabilistic time series forecasting models the uncertainty in data by predicting the distribution of future values, which enhances decision-making flexibility and improves risk management. Traditional probabilistic forecasting methods usually assume a fixed distribution of data, which is not always true for time series. Recently, there have been efforts to adapt diffusion models for time series owing to their exceptional ability to model the distribution of data without prior assumptions. However, how to apply advantages of diffusion models to time series forecasting remains a substantial challenge due to specific issues in time series such as distribution drift and complex dynamic temporal patterns.

Temporal Graph Networks (TGNs) have demonstrated their remarkable performance in modeling temporal interaction graphs. These works can generate temporal node representations by encoding the surrounding neighborhoods for the target node. However, an inherent limitation of existing TGNs is their reliance onfixed, hand-crafted rules for neighborhood encoding, overlooking the necessity for an adaptive and learnable neighborhood that can accommodate both personalization and temporal evolution across different timestamps. In this paper, we aim to enhance existing TGNs by introducing anadaptive neighborhood encoding mechanism. We present SEAN (Selective Encoding for Adaptive Neighborhood), a flexible plug-and-play model that can be seamlessly integrated with existing TGNs, effectively boosting their performance. To achieve this, we decompose the adaptive neighborhood encoding process into two phases: (i) representative neighbor selection, and (ii) temporal-aware neighborhood information aggregation. Specifically, we propose the Representative Neighbor Selector component, which automatically pinpoints the most important neighbors for the target node. It offers a tailored understanding of each node's unique surrounding context, facilitating personalization. Subsequently, we propose a Temporal-aware Aggregator, which synthesizes neighborhood aggregation by selectively determining the utilization of aggregation routes and decaying the outdated information, allowing our model to adaptively leverage both the contextually significant and current information during aggregation. We conduct extensive experiments by integrating SEAN into three representative TGNs, evaluating their performance on four public datasets and one financial benchmark dataset introduced in this paper. The results demonstrate that SEAN consistently leads to performance improvements across all models, achieving SOTA performance and exceptional robustness.

Urban public nuisance events, like garbage exposure, illegal parking, facilities damage, and etc., impair the quality of life for city residents. Predicting and controlling these nuisances is crucial but complicated due to their ties to subjective and psychological factors. In this study, we reveal a significant correlation between such nuisances and meteorological indicators, influenced by the impact of climate on people's psychological states. We employ meteorology predictions that are integrated in Hawkes processes to enhance the accuracy of predicting the category and timing of these nuisances. To this end, we propose Spatial-Temporal Two-Tower Transformer (ST-T3), which simultaneously considers spatial data and further improves the prediction accuracy. Evaluated by about three-year data from both downtown and suburban Shanghai, our method outperforms both traditional and advanced prediction systems. We share a portion of the de-identified dataset for open research.

Targeted community detection aims to distinguish a particular type of community in the network. This is an important task with a lot of real-world applications, e.g., identifying fraud groups in transaction networks. Traditional community detection methods fail to capture the specific features of the targeted community and detect all types of communities indiscriminately. Semi-supervised community detection algorithms, emerged as a feasible alternative, are inherently constrained by their limited adaptability and substantial reliance on a large amount of labeled data, which demands extensive domain knowledge and manual effort.

Online groups have become increasingly prevalent, providing users with space to share experiences and explore interests. Therefore, user-centric group discovery task, i.e., recommending groups to users can help both users' online experiences and platforms' long-term developments. Existing recommender methods can not deal with this task as modeling user-group participation into a bipartite graph overlooks their item-side interests. Although there exist a few works attempting to address this task, they still fall short in fully preserving the social context and ensuring effective interest representation learning.

Continuous-time dynamic graph modeling is a crucial task for many real-world applications, such as financial risk management and fraud detection. Though existing dynamic graph modeling methods have achieved satisfactory results, they still suffer from three key limitations, hindering their scalability and further applicability. i) Indiscriminate updating. For incoming edges, existing methods would indiscriminately deal with them, which may lead to more time consumption and unexpected noisy information. ii) Ineffective node-wise long-term modeling. They heavily rely on recurrent neural networks (RNNs) as a backbone, which has been demonstrated to be incapable of fully capturing node-wise long-term dependencies in event sequences. iii) Neglect of re-occurrence patterns. Dynamic graphs involve the repeated occurrence of neighbors that indicates their importance, which is disappointedly neglected by existing methods.

Temporal Graph Networks (TGNs) have shown remarkable performance in learning representation for continuous-time dynamic graphs. However, real-world dynamic graphs typically contain diverse and intricate noise. Noise can significantly degrade the quality of representation generation, impeding the effectiveness of TGNs in downstream tasks. Though structure learning is widely applied to mitigate noise in static graphs, its adaptation to dynamic graph settings poses two significant challenges. i) Noise dynamics. Existing structure learning methods are ill-equipped to address the temporal aspect of noise, hampering their effectiveness in such dynamic and ever-changing noise patterns. ii) More severe noise. Noise may be introduced along with multiple interactions between two nodes, leading to the re-pollution of these nodes and consequently causing more severe noise compared to static graphs.

In real-world scenarios, dynamic signed networks are ubiquitous where edges have positive and negative sign semantics and evolve over time. Encoding the dynamics and sign semantics of the network simultaneously is challenging. Moreover, over-smoothing is inevitably introduced by the learning of network dynamics. Targeting this gap, we propose Dynamic Signed Network Embedding (DynamiSE), which effectively integrates the balance theory and ordinary differential equation (ODE) into node representation learning to construct a deeper dynamic signed graph neural network and capture the complex sign semantics formed by the two types of edges.

Uncivilized urban events disrupt urban order and have a detrimental impact on daily life. Recognizing the significant implications of these events, urban managers strive to proactively prevent them by accurately predicting their future occurrence. However, existing methods overlook crucial contextual information within urban scenarios while mining spatio-temporal dependencies in single event series. Fortunately, we discovered a connection between meteorological conditions and uncivilized events. To leverage this relationship, we propose a novel approach named the Meteorology-Assisted Spatio-Temporal Graph Neural Network (MAST) which integrates meteorological information into the spatio-temporal dependency modeling for predicting urban uncivilized events. Additionally, our approach captures latent regularities in human behavior by explicitly modeling individuals’ psychological states based on meteorological information. We also adopt cross-view contrastive learning between urban regions to dynamically capture the informative components of meteorological information for precise prediction of urban uncivilized events. Experimental evaluations on a real-world dataset demonstrate the superiority of MAST over state-of-theart baselines in terms of predictive performance.

Pre-trained language models (PLMs) such as GPTs have been revealed to be biased towards certain target classes because of the prompt and the model's intrinsic biases. In contrast to the fully supervised scenario where there are a large number of costly labeled samples that can be used to fine-tune model parameters to correct for biases, there are no labeled samples available for the zero-shot setting. We argue that a key to calibrating the biases of a PLM on a target task in zero-shot setting lies in detecting and estimating the biases, which remains a challenge. In this paper, we first construct probing samples with the randomly generated token sequences, which are simple but effective in detecting inputs for stimulating GPTs to show the biases; and we pursue an in-depth research on the plausibility of utilizing class scores for the probing samples to reflect and estimate the biases of GPTs on a downstream target task. Furtherly, in order to effectively utilize the probing samples and thus reduce negative effects of the biases of GPTs, we propose a lightweight model Calibration Adapter (CA) along with a self-guided training strategy that carries out distribution-level optimization, which enables us to take advantage of the probing samples to fine-tune and select only the proposed CA, respectively, while keeping the PLM encoder frozen. To demonstrate the effectiveness of our study, we have conducted extensive experiments, where the results indicate that the calibration ability acquired by CA on the probing samples can be successfully transferred to reduce negative effects of the biases of GPTs on a downstream target task, and our approach can yield better performance than state-of-the-art (SOTA) models in zero-shot settings.

Since group activities have become very common in daily life, there is an urgent demand for generating recommendations for a group of users, referred to as group recommendation task. Existing group recommendation methods usually infer groups’ preferences via aggregating diverse members’ interests. Actually, groups’ ultimate choice involves compromises between members, and finally, an agreement can be reached. However, existing individual information aggregation lacks a holistic group-level consideration, failing to capture the consensus information. Besides, their specific aggregation strategies either suffer from high computational costs or become too coarse-grained to make precise predictions.

Temporal interaction graphs (TIGs), consisting of sequences of timestamped interaction events, are prevalent in fields like e-commerce and social networks. To better learn dynamic node embeddings that vary over time, researchers have proposed a series of temporal graph neural networks for TIGs. However, due to the entangled temporal and structural dependencies, existing methods have to process the sequence of events chronologically and consecutively to ensure node representations are up-to-date. This prevents existing models from parallelization and reduces their flexibility in industrial applications. To tackle the above challenge, in this paper, we propose TIGER, a TIG embedding model that can restart at any timestamp. We introduce a restarter module that generates surrogate representations acting as the warm initialization of node representations. By restarting from multiple timestamps simultaneously, we divide the sequence into multiple chunks and naturally enable the parallelization of the model. Moreover, in contrast to previous models that utilize a single memory unit, we introduce a dual memory module to better exploit neighborhood information and alleviate the staleness problem. Extensive experiments on four public datasets and one industrial dataset are conducted, and the results verify both the effectiveness and the efficiency of our work.

Risk scoring systems have been widely deployed in many applications, which assign risk scores to users according to their behavior sequences. Though many deep learning methods with sophisticated designs have achieved promising results, the black-box nature hinders their applications due to fairness, explainability, and compliance consideration. Rule-based systems are considered reliable in these sensitive scenarios. However, building a rule system is labor-intensive. Experts need to find informative statistics from user behavior sequences, design rules based on statistics and assign weights to each rule. In this paper, we bridge the gap between effective but black-box models and transparent rule models. We propose a two-stage method, RuDi, that distills the knowledge of black-box teacher models into rule-based student models. We design a Monte Carlo tree search-based statistics generation method that can provide a set of informative statistics in the first stage. Then statistics are composed into logical rules with our proposed neural logical networks by mimicking the outputs of teacher models. We evaluate RuDi on three real-world public datasets and an industrial dataset to demonstrate its effectiveness.

Community detection refers to the task of discovering closely related subgraphs to understand the networks. However, traditional community detection algorithms fail to pinpoint a particular kind of community. This limits its applicability in real-world networks, e.g., distinguishing fraud groups from normal ones in transaction networks. Recently, semi-supervised community detection emerges as a solution. It aims to seek other similar communities in the network with few labeled communities as training data. Existing works can be regarded as seed-based: locate seed nodes and then develop communities around seeds. However, these methods are quite sensitive to the quality of selected seeds since communities generated around a mis-detected seed may be irrelevant. Besides, they have individual issues, e.g., inflexibility and high computational overhead. To address these issues, we propose CLARE, which consists of two key components, Community Locator and Community Rewriter. Our idea is that we can locate potential communities and then refine them. Therefore, the community locator is proposed for quickly locating potential communities by seeking subgraphs that are similar to training ones in the network. To further adjust these located communities, we devise the community rewriter. Enhanced by deep reinforcement learning, it suggests intelligent decisions, such as adding or dropping nodes, to refine community structures flexibly. Extensive experiments verify both the effectiveness and efficiency of our work compared with prior state-of-the-art approaches on multiple real-world datasets.

Click-through rate prediction is a critical task in online advertising. Currently, many existing methods attempt to extract user potential interests from historical click behavior sequences. However, it is difficult to handle sparse user behaviors or broaden interest exploration. Recently, some researchers incorporate the item-item co-occurrence graph as an auxiliary. Due to the elusiveness of user interests, those works still fail to determine the real motivation of user click behaviors. Besides, those works are more biased towards popular or similar commodities. They lack an effective mechanism to break the diversity restrictions. In this paper, we point out two special properties of triangles in the item-item graphs for recommendation systems: Intra-triangle homophily and Inter-triangle heterophiy. Based on this, we propose a novel and effective framework named Triangle Graph Interest Network (TGIN). For each clicked item in user behavior sequences, we introduce the triangles in its neighborhood of the item-item graphs as a supplement. TGIN regards these triangles as the basic units of user interests, which provide the clues to capture the real motivation for a user clicking an item. We characterize every click behavior by aggregating the information of several interest units to alleviate the elusive motivation problem. The attention mechanism determines users' preference for different interest units. By selecting diverse and relative triangles, \short brings in novel and serendipitous items to expand exploration opportunities of user interests. Then, we aggregate the multi-level interests of historical behavior sequences to improve CTR prediction. Extensive experiments on both of public and industrial datasets clearly verify the effectiveness of our framework.

Fuzzy Petri nets (FPNs) are a useful instrument for modelling expert systems to conduct knowledge representation and reasoning. Many studies have been carried out for improving the performance of FPNs in terms of their accurate representation of knowledge and power of approximate reasoning. Nevertheless, the current representation methods with FPNs are unable to handle the uncertain linguistic knowledge given by domain experts and the reliability of their judgments. In addition, the existing reasoning algorithms have no way to capture the interrelationship of the propositions with the same output transition. Therefore, we present a new type of FPNs, called 2-dimensional uncertain linguistic Petri nets (2DULPNs). The 2-dimensional uncertain linguistic variables (2DULVs) and Choquet integral are combined for knowledge representation and reasoning for the first time. The truth degrees of propositions, thresholds and certainty values of linguistic production rules are denoted as 2DULVs. Some new aggregated operators based on Choquet integral are proposed and used in the approximate reasoning to capture the interactions among antecedent propositions. Finally, an equipment fault diagnosis example is provided to illustrate the correctness and effectiveness of the proposed 2DULPN model.

In order to model the evolution of user preference, we should learn user/item embeddings based on time-ordered item purchasing sequences, which is defined as Sequential Recommendation~(SR) problem. Existing methods leverage sequential patterns to model item transitions. However, most of them ignore crucial temporal collaborative signals, which are latent in evolving user-item interactions and coexist with sequential patterns. Therefore, we propose to unify sequential patterns and temporal collaborative signals to improve the quality of recommendation, which is rather challenging. Firstly, it is hard to simultaneously encode sequential patterns and collaborative signals. Secondly, it is non-trivial to express the temporal effects of collaborative signals.

Chinese characters are often composed of subcharacter components which are also semantically informative, and the component-level internal semantic features of a Chinese character inherently bring with additional information that benefits the semantic representation of the character. Therefore, there have been several studies that utilized subcharacter component information (e.g. radical, fine-grained components and stroke n-grams) to improve Chinese character representation.

Human interactions with items are being constantly logged, which enables advanced representation learning and facilitates various tasks. Instead of generating static embeddings at the end of training, several temporal embedding methods were recently proposed to learn user and item embeddings as functions of time, where each entity has a trajectory of embedding vectors aiming to encode the full dynamics. However, these methods may not be optimal to encode the dynamical behaviors on the interaction graphs in that they can not generate "fully''-temporal embeddings and do not consider information propagation. In this paper, we tackle the issues and propose CoPE (Co ntinuous P ropagation and E volution). We use an ordinary differential equation based graph neural network to model information propagation and more sophisticated evolution patterns. We train CoPE on sequences of interactions with the help of meta-learning to ensure fast adaptation to the most recent interactions. We evaluate CoPE on three tasks and prove its effectiveness.

Unsupervised pre-trained language models (PLMs) have boosted the development of effective biomedical text mining models. But the biomedical texts contain a huge number of long-tail concepts and terminologies, which makes further pre-training on biomedical corpora relatively expensive (more biomedical corpora and more pre-training steps are needed). Nonetheless, this problem receives less attention in recent studies. In Chinese biomedical text, concepts and terminologies consist of Chinese characters, and Chinese characters are often composed of sub-character components which are also semantically informative; thus in order to enhance the semantics of biomedical concepts and terminologies, the use of a Chinese character’s component-level internal semantic information also appears to be reasonable.In this paper, we propose a novel hanzi-aware pre-trained language model for Chinese biomedical text mining, referred to as BioHanBERT (hanzi-aware BERT for Chinese biomedical text mining), utilizing the component-level internal semantic information of Chinese characters to enhance the semantics of Chinese biomedical concepts and terminologies, and thereby to reduce further pre-training costs. BioHanBERT first employs a Chinese character encoder to extract the component-level internal semantic feature of each Chinese character, and then fuse the character’s internal semantic feature and its contextual embedding extracted by BERT to enrich the representations of the concepts or terminologies containing the character. The results of extensive experiments show that our model is able to consistently outperform current state-of-the-art (SOTA) models in a wide range of Chinese biomedical natural language processing (NLP) tasks.

Semantic relationship information is important to the image-text retrieval task. Existing work usually extract relationship information by calculating the relationship value pairwise, which is hardly to find out a meaningful semantic relationship. A more reasonable method is to convert the modal to a scene graph, thereby explicitly modeling the relationship. Scene graph is a kind of graph data structure modeling the scene of modality. There are two concept in a scene graph, object and relationship. In image modal, object indicates the image region and relationship represents the predicate of the image regions. In text modal, object indicates the entity and relationship represents the association between entities, also known as semantic relationship. In image-text retrieval task, both object and relationship are important, and a key challenge is to obtain semantic information. In this paper, image and text are represented as two kinds of scene graphs: visual scene graph and textual scene graph, and then they are combined into Heterogeneous Scene Graph(HSG). By explicitly modeling relationships using directed graph, the information can be passed edge-wise. To further extract semantic information, we introduce the metapath, which can extract specific semantic information on specified path. Moreover, we propose Heterogeneous Message Passing(HMP) to communicate information on the metapath. After the message passing, the similarity of two modalities can be represented as the similarity of the graphs. Experiment shows that the model achieve competitive results on Flickr30K and MSCOCO, which indicates that our approach has advantages in image-text retrieval.

Citing comprehensive and correct related work is crucial in academic writing. It can not only support the author's claims but also help readers trace other related research papers. Nowadays, with the rapid increase in the number of scientific literatures, it has become increasingly challenging to search for high-quality citations and write the manuscript. In this paper, we present an automatic writing assistant model, AutoCite, which not only infers potentially related work but also automatically generates the citation context at the same time. Specifically, AutoCite involves a novel multi-modal encoder and a multi-task decoder architecture. Based on the multi-modal inputs, the encoder in AutoCite learns paper representations with both citation network structure and textual contexts. The multi-task decoder in AutoCite couples and jointly learns citation prediction and context generation in a unified manner. To effectively join the encoder and decoder, we introduce a novel representation fusion component, i.e., gated neural fusion, which feeds the multi-modal representation inputs from the encoder and creates outputs for the downstream multi-task decoder adaptively. Extensive experiments on five real-world citation network datasets validate the effectiveness of our model.

Capturing the topological and temporal information of interactions and predicting future interactions are crucial for many domains, such as social networks, financial transactions, and e-commerce. With the advent of co-evolutional models, the mutual influence between the interacted users and items are captured. However, existing models only update the interaction information of nodes along the timeline. It causes the problem of information asymmetry, where early updated nodes often have much less information than the most recently updated nodes. The information asymmetry is essentially a blockage of information flow. We propose HILI (Highly Liquid Temporal Interaction Graph Embeddings) to predict highly liquid embeddings on temporal interaction graphs. Our embedding model makes interaction information highly liquid without information asymmetry. A specific least recently used-based and frequency-based windows are used to determine the priority of the nodes that receive the latest interaction information. HILI updates node embeddings by attention layers. The attention layers learn the correlation between nodes and update node embedding simply and quickly. In addition, HILI elaborately designs, a self-linear layer, a linear layer initialized in a novel method. A self-linear layer reduces the expected space of predicted embedding of the next interacting node and makes predicted embedding focus more on relevant nodes. We illustrate the geometric meaning of a self-linear layer in the paper. Furthermore, the results of the experiments show that our model outperforms other state-of-the-art temporal interaction prediction models.

Collective classification has attracted considerable attention in the last decade, where the labels within a group of instances are correlated and should be inferred collectively, instead of independently. Conventional approaches on collective classification mainly focus on exploiting simple relational features (such as count and exists aggregators on neighboring nodes). However, many real-world applications involve complex dependencies among the instances, which are obscure/hidden in the networks. To capture these dependencies in collective classification, we need to go beyond simple relational features and extract deep dependencies between the instances. In this paper, we study the problem of deep collective classification in Heterogeneous Information Networks (HINs), which involve different types of autocorrelations, from simple to complex relations, among the instances. Different from conventional autocorrelations, which are given explicitly by the links in the network, complex autocorrelations are obscure/hidden in HINs, and should be inferred from existing links in a hierarchical order. This problem is highly challenging due to the multiple types of dependencies among the nodes and the complexity of the relational features. In this study, we proposed a deep convolutional collective classification method, called GraphInception, to learn the deep relational features in HINs. And we presented two versions of the models with different inference styles. The proposed methods can automatically generate a hierarchy of relational features with different complexities. Extensive experiments on four real-world networks demonstrate that our approach can improve the collective classification performance by considering deep relational features in HINs.

Node embedding techniques have gained prominence since they produce continuous and low-dimensional features, which are effective for various tasks. Most existing approaches learn node embeddings by exploring the structure of networks and are mainly focused on static non-attributed graphs. However, many real-world applications, such as stock markets and public review websites, involve bipartite graphs with dynamic and attributed edges, called attributed interaction graphs. Different from conventional graph data, attributed interaction graphs involve two kinds of entities (e.g. investors/stocks and users/businesses) and edges of temporal interactions with attributes (e.g. transactions and reviews). In this paper, we study the problem of node embedding in attributed interaction graphs. Learning embeddings in interaction graphs is highly challenging due to the dynamics and heterogeneous attributes of edges. Different from conventional static graphs, in attributed interaction graphs, each edge can have totally different meanings when the interaction is at different times or associated with different attributes. To tackle the above challenges, we introduce the temporal dependency and conditional proximity, which are two fundamental characteristics of interaction graphs. Then, we propose a deep node embedding method called IGE+ (Interaction Graph Embedding+). By preserving these two characteristics, IGE+ is able to produce effective node embeddings in interaction graphs. We evaluate our proposed method and various comparing methods on four real-world datasets. The experimental results prove the effectiveness of the learned embeddings by IGE+ on both node-based and edge-based tasks.

Network embedding is gaining more and more attention, aiming to map latent features into low dimensional space. Quite a lot of algorithms focus on static networks. Everything is not set in stone in the real world. Networks evolving with time are called dynamic networks. Most of the methods that regard dynamic networks as a series of snapshots, getting network embeddings and network evolving patterns separately. Furthermore, they always focus on snapshots only; few consider time interval information, including topology changes and the number of edge occurrences during each time interval.This paper proposes a model to learn dynamic network embedding named TI-GCN (Time Interval Graph Convolutional Networks). Specifically, we come up with a heuristic framework to update network embeddings with the embeddings inherited from the previous snapshot. Thus, the network embeddings are more traceable, which complements mining the evolving patterns. Through the framework, the processes of learning the network embeddings and getting evolving patterns are integrated. We also fuse time interval information so that the network embeddings are updated adapted to the network changes accordingly. The number of edge occurrences serves as supplemental information to distinguish the strength of different edges. Besides, we apply the gate mechanism for controlling the information flow. All these parts are conducted under Neural ODEs (NeuralOrdinary Differential Equations) framework. Finally, our method performs better on the temporal link prediction task than almost all the other methods on the real-world datasets.

Graph Neural Networks(GNNs), like GCN and GAT, have achieved great success in a number of supervised or semi-supervised tasks including node classification and link prediction. These existing graph neural networks can effectively encode neighborhood information of graph nodes through their message aggregating mechanisms. However, there are some unsupervised and structure-related tasks like community detection, which is a fundamental problem in network analysis that finds densely-connected groups of nodes and separates them from others in graphs. It is still difficult for these general-purposed GNNs to learn the needed structural information in these particular problems. To overcome the shortcomings of general-purposed graph representation learning methods, we propose the Community Deep Graph Infomax (CommDGI), a graph neural network designed to handle community detection problems. Inspired by the success of deep graph infomax in self-supervised graph learning, we design a novel mutual information mechanism to capture neighborhood as well as community information in graphs. A trainable clustering layer is employed to learn the community partition in an end-to-end manner. Disentangled representation learning is applied in our graph neural network so that the model can improve interpretability and generalization. Throughout the whole learning process, joint optimization is applied to learn the community-related node representations. The experimental results show that our algorithm outperforms state-of-the-art community detection methods.

Graph representation learning has attracted lots of attention recently. Existing graph neural networks fed with the complete graph data are not scalable due to limited computation and memory costs. Thus, it remains a great challenge to capture rich information in large-scale graph data. Besides, these methods mainly focus on supervised learning and highly depend on node label information, which is expensive to obtain in the real world. As to unsupervised network embedding approaches, they overemphasize node proximity instead, whose learned representations can hardly be used in downstream application tasks directly. In recent years, emerging self-supervised learning provides a potential solution to address the aforementioned problems. However, existing self-supervised works also operate on the complete graph data and are biased to fit either global or very local (1-hop neighborhood) graph structures in defining the mutual information based loss terms. In this paper, a novel self-supervised representation learning method via Sub-graph Contrast, namely Subg-Con, is proposed by utilizing the strong correlation between central nodes and their sampled subgraphs to capture regional structure information. Instead of learning on the complete input graph data, with a novel data augmentation strategy, Subg-Con learns node representations through a contrastive loss defined based on subgraphs sampled from the original graph instead. Compared with existing graph representation learning approaches, Subg-Con has prominent performance advantages in weaker supervision requirements, model learning scalability, and parallelization. Extensive experiments verify both the effectiveness and the efficiency of our work compared with both classic and state-of-the-art graph representation learning approaches on multiple realworld large-scale benchmark datasets from different domains.

Community detection is an important task with many applications. However, there is no universal definition of communities, and a variety of algorithms have been proposed based on different assumptions. In this paper, we instead study the semi-supervised community detection problem where we are given several communities in a network as training data and aim to discover more communities. This setting makes it possible to learn concepts of communities from data without any prior knowledge. We propose the Seed Expansion with generative Adversarial Learning (SEAL), a framework for learning heuristics for community detection. SEAL contains a generative adversarial network, where the discriminator predicts whether a community is real or fake, and the generator generates communities that cheat the discriminator by implicitly fitting characteristics of real ones. The generator is a graph neural network specialized in sequential decision processes and gets trained by policy gradient. Moreover, a locator is proposed to avoid well-known free-rider effects by forming a dual learning task with the generator. Last but not least, a seed selector is utilized to provide promising seeds to the generator. We evaluate SEAL on 5 real-world networks and prove its effectiveness.

In this paper, we propose a novel representation learning framework, named MEGAE, for heterogeneous information networks. To investigate the rich semantic information in heterogeneous information networks, we use metapaths to complete implicit links between nodes. A graph attention encoder is further used to learn graph structural information with shared weight parameters. The attention mechanism, on the other hand, provides us an intuition of how the representation is learned and improves the interpretability of our model. Furthermore, a multitask learning of node classification and link prediction is trained to achieve more robust generalization ability. To validate our ideas, extensive experiments on three real-world datasets show that our model achieves state-of-the-art results on node classification and link prediction tasks in HINs.

To enjoy more social network services, users nowadays are usually involved in multiple online sites at the same time. Aligned social networks provide more information to alleviate the problem of data insufficiency. In this paper, we target on the collective link prediction problem and aim to predict both the intra-network social links as well as the inter-network anchor links across multiple aligned social networks. It is not an easy task, and the major challenges involve the network characteristic difference problem and different directivity properties of the social and anchor links to be predicted. To address the problem, we propose an application oriented network embedding framework, Hierarchical Graph Attention based Network Embedding (HGANE), for collective link prediction over directed aligned networks. Very different from the conventional general network embedding models, HGANE effectively incorporates the collective link prediction task objectives into consideration. It learns the representations of nodes by aggregating information from both the intra-network neighbors (connected by social links) and inter-network partners (connected by anchor links). What's more, we introduce a hierarchical graph attention mechanism for the intra-network neighbors and inter-network partners respectively, which resolves the network characteristic differences and the link directivity challenges effectively. Extensive experiments have been conducted on real-world aligned networks datasets to demonstrate that our model outperformed the state-of-the-art baseline methods in addressing the collective link prediction problem by a large margin.

Assigning standard medical codes (e.g., ICD-9-CM) representing diagnoses or procedures to electronic health record (EHR) is an important task in the medical domain. However, automatic coding is difficult since the clinical note is composed of multiple long and heterogeneous textual narratives (e.g., discharge diagnosis, pathology reports, surgical procedure notes). Furthermore, the code label space is large and the label distribution is extremely unbalanced. The state-of-the-art methods mainly regard EHR coding as a multi-label text classification task and use shallow convolution neural network with fixed window size, which is incapable of learning variable n-gram features and the ontology structure between codes. In this paper, we leverage a densely connected convolutional neural network which is able to produce variable n-gram features for clinical note feature learning. We also incorporate a multi-scale feature attention to adaptively select multi-scale features since the most informative n-grams in clinical notes for each word can vary in length according to the neighborhood. Furthermore, we leverage graph convolutional neural network to capture both the hierarchical relationships among medical codes and the semantics of each code. Finally, We validate our method on the public dataset, and the evaluation results indicate that our method can significantly outperform other state-of-the-art models.

Collective classification has attracted considerable attention in the last decade, where the labels within a group of instances are correlated and should be inferred collectively, instead of independently. Conventional approaches on collective classification mainly focus on exploiting simple relational features (such ascount andexists aggregators on neighboring nodes). However, many real-world applications involve complex dependencies among the instances, which are obscure/hidden in the networks. To capture these dependencies in collective classification, we need to go beyond simple relational features and extract deep dependencies between the instances. In this paper, we study the problem of deep collective classification inHeterogeneous Information Networks (HINs), which involves different types of autocorrelations, from simple to complex relations, among the instances. Different from conventional autocorrelations, which are given explicitly by the links in the network, complex autocorrelations are obscure/hidden in HINs, and should be inferred from existing links in a hierarchical order. This problem is highly challenging due to the multiple types of dependencies among the nodes and the complexity of the relational features. In this study, we proposed a deep convolutional collective classification method, called GraphInception to learn the deep relational features in HINs. The proposed method can automatically generate a hierarchy of relational features with different complexities. Extensive experiments on four real-world networks demonstrate that our approach can improve the collective classification performance by considering deep relational features in HINs.

Collective inference has attracted considerable attention in the last decade, where the response variables within a group of instances are correlated and should be inferred collectively, instead of independently. Previous works on collective inference mainly focus on exploiting the autocorrelation among instances in a static network during the inference process. There are also approaches on time series prediction, which mainly exploit the autocorrelation within an instance at different time points during the inference process. However, in many real-world applications, the response variables of related instances can co-evolve over time and their evolutions are not following a static correlation across time, but are following an internal life cycle. In this paper, we study the problem of collective evolution inference, where the goal is to predict the values of the response variables for a group of related instances at the end of their life cycles. This problem is extremely important for various applications, e.g., predicting fund-raising results in crowd-funding and predicting gene-expression levels in bioinformatics. This problem is also highly challenging because different instances in the network can co-evolve over time and they can be at different stages of their life cycles and thus have different evolving patterns. Moreover, the instances in collective evolution inference problems are usually connected through heterogeneous information networks (HINs for short), which involve complex relationships among the instances interconnected by multiple types of links. We propose an approach, called NetCycle+, by incorporating information from both the correlation among related instances and their life cycles. Furthermore, in order to study the deep dependencies between nodes in the network, we extend the graph convolution model into our algorithm. We compared our approach with existing methods of collective inference and time series analysis on two real-world networks. The results demonstrate that our proposed approach can improve the inference performance by considering the autocorrelation through networks and the life cycles of the instances.

Node embedding techniques have gained prominence since they produce continuous and low-dimensional features, which are effective for various tasks. Most existing approaches learn node embeddings by exploring the structure of networks and are mainly focused on static non-attributed graphs. However, many real-world applications, such as stock markets and public review websites, involve bipartite graphs with dynamic and attributed edges, called attributed interaction graphs. Different from conventional graph data, attributed interaction graphs involve two kinds of entities (e.g. investors/stocks and users/businesses) and edges of temporal interactions with attributes (e.g. transactions and reviews). In this paper, we study the problem of node embedding in attributed interaction graphs. Learning embeddings in interaction graphs is highly challenging due to the dynamics and heterogeneous attributes of edges. Different from conventional static graphs, in attributed interaction graphs, each edge can have totally different meanings when the interaction is at different times or associated with different attributes. We propose a deep node embedding method called IGE (Interaction Graph Embedding). IGE is composed of three neural networks: an encoding network is proposed to transform attributes into a fixed-length vector to deal with the heterogeneity of attributes; then encoded attribute vectors interact with nodes multiplicatively in two coupled prediction networks that investigate the temporal dependency by treating incident edges of a node as the analogy of a sentence in word embedding methods. The encoding network can be specifically designed for different datasets as long as it is differentiable, in which case it can be trained together with prediction networks by back-propagation. We evaluate our proposed method and various comparing methods on four real-world datasets. The experimental results prove the effectiveness of the learned embeddings by IGE on both node clustering and classification tasks.

Sparse inverse covariance estimation has attracted lots of interests since it can recover the structure of the underlying Gaussian graphical model. This is a useful tool to demonstrate the connections among objects (nodes). Previous works on sparse inverse covariance estimation mainly focus on learning one single type of connections from the observed activities with a lasso, group lasso or tree-structure penalty. However, in many real-world applications, the observed activities on the nodes can be related to multiple types of connections. In this paper, we consider the problem of learning heterogeneous connectivities from the observed activities by incorporating meta paths extracted from a heterogeneous information network (HIN), an information network with multiple types of nodes and links, into the conventional graphical lasso framework. We aim at extracting the strongest type of relation between any pairs of entities and ignoring other minor relations. Specially, we introduce two novel kinds of constraints: meta path constraints and exclusive constraints, which ensure the unique type of relation among a pair of objects. This problem is highly challenging due to the non-convex optimization. We proposed a method based upon the alternating direction method of multipliers (ADMM) to efficiently solve the problem. The conducted experiments on both synthetic and real-world datasets illustrate the effectiveness of the proposed method.

The understanding of passenger flow assignment in metro network is critical for public transit management. However, the route chosen by one passenger is difficult to be directly obtained according to the transaction records only including each trip's tap-in and tap-out time stamp and stations. In this paper, a two-stage framework for calculating passenger flow assignment in complex metro networks is proposed, named PaFA (Passenger Flow Assignment), by using smart card data. First, we design an acceleration search process to obtain all routes for each O-D pair and select the candidate routes under rules. Then, inspired by topic model, we realize similar latent relationships also can be found among O-D pair, candidate routes and passenger's travel time. Along this line, we obtain the distribution of passenger flow in different candidate routes. Finally, a comprehensive evaluation with real-world data is conducted. The results demonstrate the enhanced performance of the proposed method.

Collective inference has attracted considerable attention in the last decade, where the response variables within a group of instances are correlated and should be inferred collectively, instead of independently. Previous works on collective inference mainly focus on exploiting the autocorrelation among instances in a static network during the inference process. There are also approaches on time series prediction, which mainly exploit the autocorrelation within an instance at different time points during the inference process. However, in many real-world applications, the response variables of related instances can co-evolve over time and their evolutions are not following a static correlation across time, but are following an internal life cycle. In this paper, we study the problem of collective evolution inference, where the goal is to predict the values of the response variables for a group of related instances at the end of their life cycles. This problem is extremely important for various applications, e.g., predicting fund-raising results in crowd-funding and predicting gene-expression levels in bioinformatics. This problem is also highly challenging because different instances in the network can co-evolve over time and they can be at different stages of their life cycles and thus have different evolving patterns. Moreover, the instances in collective evolution inference problems are usually connected through heterogeneous information networks, which involve complex relationships among the instances interconnected by multiple types of links. We propose an approach, called NetCycle, by incorporating information from both the correlation among related instances and their life cycles. We compared our approach with existing methods of collective inference and time series analysis on two real-world networks. The results demonstrate that our proposed approach can improve the inference performance by considering the autocorrelation through networks and the life cycles of the instances.

In this paper, we study the problem of personalized restaurant recommendations. Specifically, we develop a probabilistic factor analysis framework, named RMSQ-MF, which has the ability in exploiting multi-source information, such as the users' task, their friends' preferences, and human mobility patterns, for personalized restaurant recommendations. The rationale of this work is motivated by two observations. First, people's preferences can be affected by their friends. Second, human mobility patterns can reflect the popularity of restaurants to a certain degree. Finally, empirical studies on real-world data demonstrate that the proposed method outperforms benchmark methods with a significant margin.

As a newly emerging network model, heterogeneous information networks (HINs) have received growing attention. Many data mining tasks have been explored in HINs, including clustering, classification, and similarity search. Similarity join is a fundamental operation required for many problems. It is attracting attention from various applications on network data, such as friend recommendation, link prediction, and online advertising. Although similarity join has been well studied in homogeneous networks, it has not yet been studied in heterogeneous networks. Especially, none of the existing research on similarity join takes different semantic meanings behind paths into consideration and almost all completely ignore the heterogeneity and diversity of the HINs. In this paper, we propose a path-based similarity join (PS-join) method to return the top k similar pairs of objects based on any user specified join path in a heterogeneous information network. We study how to prune expensive similarity computation by introducing bucket pruning based locality sensitive hashing (BPLSH) indexing. Compared with existing Link-based Similarity join (LS-join) method, PS-join can derive various similarity semantics. Experimental results on real data sets show the efficiency and effectiveness of the proposed approach.

Clustering is a common natural phenomenon. In microcosms, the nucleus is formed by the aggregation of nuclear particles through strong interactions, which can be illustrated by Yukawa potential. Inspired by this clustering phenomenon, we propose a novel dynamic clustering algorithm based on Yukawa potential (Yupc). Each data object is regarded as a particle following the basic rules of movements in the Yukawa potential field. After several time intervals, similar objects gradually aggregate together and form clear clusters. Yupc neither relies on any assumption of data distribution, nor prescribes any specific number of clusters. Natural clusters of different shapes, densities, sizes, numbers and distributions can be detected by Yupc, reflecting the intrinsic structure of the original data set. In addition, we propose a framework to automatically find appropriate parameters for Yupc. Experiments performed on synthetic and real-world data show that this approach outperforms existing algorithms, especially in data sets with arbitrary kinds of clusters.

Link prediction in weighted network is an important task in Social Network Analysis. This problem aims at determining missing links in weighted networks. By taking advantage of the weights and structural information of networks, a mechanism for rating nodes' authorities in terms of the value of weight, called Benefit Rank, is defined. This mechanism can flexibly collect different order neighbors' information of nodes to complete the rating authority process for each node in weighted networks. Using Benefit Rank combined with the Weak Ties theory, similarity measures are proposed to estimate the emergence of future relationships between nodes in weighted networks. Extensive experiments were carried out on four real weighted networks. Compared with existing methods, our methods can provide higher accuracy for link prediction in weighted networks.

Behavior mining is one of the most important issues in data mining. The growing interest in the study of behavior mining has been credited to the availability of a large amount of individual behavioral data. Some objects containing common behavioral patterns in the dataset are dramatically different from other individual objects and show their peculiarities. It is very important for behavior analysis to mine these peculiar objects' groups as this has great potential in practice. However, to the best of our knowledge, it has not been explored before. In this paper, we identify this interesting and practical problem of behavior mining: mining peculiarity groups and defining a measurement of the degree of peculiarity. As the first attempt to tackle the problem, we present a set-value-oriented day-by-day behavioral data expression mode considering that daily behaviors with respect to an object should be recorded as a set of behaviors, and devise a peculiarity group mining algorithm in view of the set-value-oriented data expression which cannot be very well handled by existing methods. Furthermore, we show that our method is practical and efficient using real datasets.

The pairwise mining problem is to discover pairwise objects having measures greater than the user-specified minimum threshold from a collection of objects. It is essential in a large variety of database and data-mining applications. Of late, there has been increasing interest in applying a Locality-Sensitive Hashing (LSH) scheme for pairwise mining. LSH-type methods have shown themselves to be simply implementable and capable of achieving significant performance gain in running time over most exact methods. However, the present LSH-type methods still suffer from some bottlenecks, such as ¿the curse of threshold¿. In this paper, we proposed a novel LSH-based method, namely Cost-effective LSH filter (Ce-LSH for short), for pairwise mining. Compared with previous LSH-type methods, it uses a lower fixed number of LSH functions and is thus more cost-effective. Substantial experiments evidence that our method gives significant improvement in running time over existing LSH-type methods and some recently reported method based on upper-bound. Experimental results also indicate that it scales well even for a relatively low minimum threshold and for a fairly small miss ratio.