Alignment of Large Language Models (LLMs) aims to align outputs with human preferences, and personalized alignment further adapts models to individual users. This relies on personalized reward models that capture user-specific preferences and automatically provide individualized feedback. However, developing these models faces two critical challenges: the scarcity of feedback from individual users and the need for efficient adaptation to unseen users. We argue that addressing these constraints requires a paradigm shift from fitting static user models to ''learning to learn'' adaptation. To realize this, we propose Meta Reward Modeling (MRM), which reformulates personalized reward modeling as a meta-learning problem. Specifically, we represent each user's reward model as a weighted combination of base reward functions, and optimize the initialization of these weights using a Model-Agnostic Meta-Learning (MAML)-style framework to support fast adaptation under limited feedback. To ensure robustness, we introduce the Robust Personalization Objective (RPO), which places greater emphasis on hard-to-learn users during meta optimization. Extensive experiments on personalized preference datasets validate that MRM enhances few-shot personalization, improves user robustness, and consistently outperforms baselines. We release code at https://github.com/ModalityDance/MRM.

Text-to-image retrieval is a fundamental task in multimedia retrieval.Traditional studies have typically approached this task as a discriminative problem, matching the text and image via the cross-attention mechanism (one-tower framework) or in a common embedding space (two-tower framework).The one-tower framework excels in effectiveness but falls short in efficiency, whereas the two-tower framework is efficient but struggles to maintain competitive effectiveness.In this study, we aim to enhance both effectiveness and efficiency by transforming the text-to-image retrieval task into a token-to-voken generation problem, where fine-grained interactions are incorporated to improve effectiveness while maintaining high efficiency.Despite its potential advantages, this paradigm shift presents significant challenges: 1) misalignment with high-level semantics and 2) learning gap towards the retrieval target.To address the challenges, we propose AVG, which discretizes images into vokens while aligning with both the visual information and high-level semantics.Additionally, to bridge the learning gap between generative training and the retrieval target, AVG incorporates discriminative training to modify the learning direction during token-to-voken training.Experiments demonstrate that the benefits of paradigm innovation are realized: compared with the classical two-tower method, CLIP, AVG achieves the 7.53% relative effectiveness improvement and also 4× efficiency improvement.We release code at the GitHub repository.

Generative retrieval reformulates retrieval as an autoregressive generation task, where large language models (LLMs) generate target documents directly from a query. As a novel paradigm, the mechanisms that underpin its performance and scalability remain largely unexplored. We systematically investigate training and inference scaling laws in generative retrieval, exploring how model size, training data scale, and inference-time compute jointly influence performance. We propose a novel evaluation metric inspired by contrastive entropy and generation loss, providing a continuous performance signal that enables robust comparisons across diverse generative retrieval methods. Our experiments show that n-gram-based methods align strongly with training and inference scaling laws. We find that increasing model size, training data scale, and inference-time compute all contribute to improved performance, highlighting the complementary roles of these factors in enhancing generative retrieval. Across these settings, LLaMA models consistently outperform T5 models, suggesting a particular advantage for larger decoder-only models in generative retrieval. Our findings underscore that model sizes, data availability, and inference computation interact to unlock the full potential of generative retrieval, offering new insights for designing and optimizing future systems. We release code at SLGR GitHub repository.

Modality incompleteness is a critical yet underexplored challenge in multimodal sentiment analysis (MSA). Existing efforts, trained and evaluated under fixed missing rates, struggle to adapt to real-world scenarios with varying missing rates. To address this, we propose the Missing Modality Adaptation Framework (M2AF), leveraging model-agnostic meta-learning to enhance robustness against different levels of modality incompleteness. M2AF operates in two stages: meta-training and meta-testing. In the meta-training stage, a pre-trained MSA model, initially optimized for fixed missing rates, is further adapted to different levels of missing rates-low, moderate, and high. In the meta-testing stage, the model rapidly updates its parameters using minimal training data to handle target missing scenarios. Experiments on two popular datasets demonstrate that M2AF significantly improves the performance and generalization of various MSA models, ensuring more robust sentiment analysis in real-world settings with different modality incompleteness.

Web agents have emerged as a promising direction to automate Web task completion based on user instructions, significantly enhancing user experience. Recently, Web agents have evolved from traditional agents to Large Language Models (LLMs)-based Web agents. Despite their success, existing LLM-based Web agents overlook the importance of personalized data (e.g., user profiles and historical Web behaviors) in assisting the understanding of users' personalized instructions and executing customized actions.

Job recommendation aims to provide potential talents with suitable job descriptions (JDs) consistent with their career trajectory, which plays an essential role in proactive talent recruitment. In real-world management scenarios, the available JD-user records always consist of JDs, user profiles, and click data, in which the user profiles are typically summarized as the user's skill distribution for privacy reasons. Although existing sophisticated recommendation methods can be directly employed, effective recommendation still has challenges considering the information deficit of JD itself and the natural heterogeneous gap between JD and user profile. To address these challenges, we proposed a novel skill-aware recommendation model based on the designed semantic-enhanced Transformer to parse JDs and complete personalized job recommendation. Specifically, we first model the relative items of each JD and then adopt an encoder with the local-global attention mechanism to better mine the intra-job and inter-job dependencies from JD tuples. Moreover, we adopt a two-stage learning strategy for skill-aware recommendation, in which we utilize the skill distribution to guide JD representation learning in the recall stage and then combine the user profiles for final prediction in the ranking stage. Consequently, we can embed rich contextual semantic representations for learning JDs, while skill-aware recommendation provides effective JD-user joint representation for click-through rate (CTR) prediction. To validate the superior performance of our method for job recommendation, we present a thorough empirical analysis of large-scale real-world and public datasets to demonstrate its effectiveness and interpretability.

Target-oriented proactive dialogue systems aim at leading conversations from a dialogue context toward a pre-determined target, such as making recommendations on designated items or introducing new specific topics. To this end, it is critical for such dialogue systems to plan reasonable actions to drive the conversation proactively, and meanwhile, to plan appropriate topics to move the conversation forward to the target topic smoothly. In this work, we mainly focus on effective dialogue planning for target-oriented dialogue generation. Inspired by decision-making theories in cognitive science, we propose a novel target-constrained bidirectional planning (TRIP) approach, which plans an appropriate dialogue path by looking ahead and looking back. By formulating the planning as a generation task, our TRIP bidirectionally generates a dialogue path consisting of a sequence of pairs using two Transformer decoders. They are expected to supervise each other and converge on consistent actions and topics by minimizing the decision gap and contrastive generation of targets. Moreover, we propose a target-constrained decoding algorithm with a bidirectional agreement to better control the planning process. Subsequently, we adopt the planned dialogue paths to guide dialogue generation in a pipeline manner, where we explore two variants: prompt-based generation and plan-controlled generation. Extensive experiments are conducted on two challenging dialogue datasets, which are re-purposed for exploring target-oriented dialogue. Our automatic and human evaluations demonstrate that the proposed methods significantly outperform various baseline models.

Learning and development, or L&D, plays an important role in talent management, which aims to improve the knowledge and capabilities of employees through a variety of performance-oriented training activities. Recently, with the rapid development of enterprise management information systems, many research efforts and industrial practices have been devoted to building personalized employee training course recommender systems. Nevertheless, a widespread challenge is how to provide explainable recommendations with the consideration of different learning motivations from talents. To this end, we propose CKGE, a contextualized knowledge graph (KG) embedding approach for developing an explainable training course recommender system. A novel perspective of CKGE is to integrate both the contextualized neighbor semantics and high-order connections as motivation-aware information for learning effective representations of talents and courses. Specifically, in CKGE, for each entity pair (i.e., the talent-course pair), we first construct a meta-graph, including the neighbors of each entity and the meta-paths between entities as motivation-aware information. Then, we develop a novel KG-based Transformer, which can serialize entities and paths in the meta-graph as a sequential input, with the specially designed relational attention and structural encoding mechanisms to better model the global dependence of KG structured data. Meanwhile, the local path mask prediction can effectively reveal the importance of different paths. As a result, CKGE not only can make precise predictions but also can discriminate the saliencies of meta-paths in characterizing corresponding preferences. Extensive experiments on real-world and public datasets clearly validate the effectiveness and interpretability of CKGE compared with state-of-the-art baselines.

Recommender systems aim to accurately and actively provide users with potentially interesting items (products, information or services). Deep reinforcement learning has been successfully applied to recommender systems, but still heavily suffer from data sparsity and cold-start in real-world tasks. In this work, we propose an effective way to address such issues by leveraging the pervasive social networks among users in the estimation of action-values (Q). Specifically, we develop a Social Attentive Deep Q-network (SADQN) to approximate the optimal action-value function based on the preferences of both individual users and social neighbors, by successfully utilizing a social attention layer to model the influence between them. Further, we propose an enhanced variant of SADQN, termed SADQN++, to model the complicated and diverse trade-offs between personal preferences and social influence for all involved users, making the agent more powerful and flexible in learning the optimal policies. The experimental results on real-world datasets demonstrate that the proposed SADQNs remarkably outperform the state-of-the-art deep reinforcement learning agents, with reasonable computation cost.

There are a variety of underlying factors influencing what and how people communicate in their daily life. The ability to capture and utilize these factors enables the conversational systems to generate favorable responses and set up amicable connections with users. In this work, we investigate two major factors in response generation, i.e., emotion and intention. To explore the dependency between them, we develop a hierarchical variational model that predicts in sequence the emotion and intention to be conveyed in a response. The response can then be generated word-by-word based on the predictions. We also apply a novel adversarial-augmented inference network to facilitate model training. The experimental results demonstrate the effectiveness of the proposed model as well as the novel adversarial objective. The hypothesis that emotion shapes human communication behavior is also validated.

Information sharing behavior and social link building behavior have shown strong correlation on social media. The aim of this paper is to explore this correlation for simultaneously modeling and predicting sharing behavior in information diffusion sequences and linking behavior in social network, which correspond to information diffusion prediction and social link prediction problems. To achieve this goal, we propose a joint user representation learning model to characterize the two correlated behaviors in a shared latent space. The proposed model learns user representations via two maximum likelihood estimation objectives defined on observed information diffusion sequences and social network structure respectively and incorporates them in a unified framework. A multi-task learning algorithm is designed for efficient model optimization. Based on the learned representations, the model can be directly applied to predicting diffusion processes and inferring unobserved social links at the same time. We evaluate the proposed model on two real social media datasets with extensive experiments. The model consistently achieves significant improvements over the state-of-the-art approaches on diffusion prediction and link prediction tasks. The better robustness of our model in further ablation studies demonstrates that capturing the behavior correlation in the shared representation space is beneficial.

In recent years, conversational agents have provided a natural and convenient access to useful information in people’s daily life, along with a broad and new research topic, conversational question answering (QA). On the shoulders of conversational QA, we study the conversational open-domain QA problem, where users’ information needs are presented in a conversation and exact answers are required to extract from the Web. Despite its significance and value, building an effective conversational open-domain QA system is non-trivial due to the following challenges: (1) precisely understand conversational questions based on the conversation context; (2) extract exact answers by capturing the answer dependency and transition flow in a conversation; and (3) deeply integrate question understanding and answer extraction. To address the aforementioned issues, we propose an end-to-end Dynamic Graph Reasoning approach to Conversational open-domain QA (DGRCoQA for short). DGRCoQA comprises three components, i.e., a dynamic question interpreter (DQI), a graph reasoning enhanced retriever (GRR), and a typical Reader, where the first one is developed to understand and formulate conversational questions while the other two are responsible to extract an exact answer from the Web. In particular, DQI understands conversational questions by utilizing the QA context, sourcing from predicted answers returned by the Reader, to dynamically attend to the most relevant information in the conversation context. Afterwards, GRR attempts to capture the answer flow and select the most possible passage that contains the answer by reasoning answer paths over a dynamically constructed context graph . Finally, the Reader, a reading comprehension model, predicts a text span from the selected passage as the answer. DGRCoQA demonstrates its strength in the extensive experiments conducted on a benchmark dataset. It significantly outperforms the existing methods and achieves the state-of-the-art performance.

Clarification has attracted much attention because of its many potential applications especially in Web search. Since search queries are very short, the underlying user intents are often ambiguous. This makes it challenging for search engines to return the appropriate results that pertain to the users' actual information needs. To address this issue, asking clarifying questions has been recognized as a critical technique. Although previous studies have analyzed the importance of asking to clarify, generating clarifying questions for Web search remains under-explored. In this paper, we tackle this problem in a template-guided manner. Our objective is jointly learning to select question templates and fill question slots, using Transformer-based networks. We conduct experiments on MIMICS, a collection of datasets containing real Web search queries sampled from Bing's search logs. Our method is demonstrated to achieve significant improvements over various competitive baselines.

In this work, we establish a context graph from both conversation utterances and external knowledge, and develop a novel graph-based encoder to better understand the conversation context. Specifically, the encoder fuses the information in the context graph stage-by-stage and provides global context-graph-aware representations of each node in the graph to facilitate knowledge-grounded response generation. On a large-scale conversation corpus, we validate the effectiveness of the proposed approach and demonstrate the benefit of knowledge in conversation understanding.

Abstract Machine learning methods, especially deep learning models, have achieved impressive performance in various natural language processing tasks including sentiment analysis. However, deep learning models are more demanding for training data. Data augmentation techniques are widely used to generate new instances based on modifications to existing data or relying on external knowledge bases to address annotated data scarcity, which hinders the full potential of machine learning techniques. This paper presents our work using part‐of‐speech (POS) focused lexical substitution for data augmentation (PLSDA) to enhance the performance of machine learning algorithms in sentiment analysis. We exploit POS information to identify words to be replaced and investigate different augmentation strategies to find semantically related substitutions when generating new instances. The choice of POS tags as well as a variety of strategies such as semantic‐based substitution methods and sampling methods are discussed in detail. Performance evaluation focuses on the comparison between PLSDA and two previous lexical substitution‐based data augmentation methods, one of which is thesaurus‐based, and the other is lexicon manipulation based. Our approach is tested on five English sentiment analysis benchmarks: SST‐2, MR, IMDB, Twitter, and AirRecord. Hyperparameters such as the candidate similarity threshold and number of newly generated instances are optimized. Results show that six classifiers (SVM, LSTM, BiLSTM‐AT, bidirectional encoder representations from transformers [BERT], XLNet, and RoBERTa) trained with PLSDA achieve accuracy improvement of more than 0.6% comparing to two previous lexical substitution methods averaged on five benchmarks. Introducing POS constraint and well‐designed augmentation strategies can improve the reliability of lexical data augmentation methods. Consequently, PLSDA significantly improves the performance of sentiment analysis algorithms.

Graph representation learning aims to embed each vertex in a graph into a low-dimensional vector space. Existing graph representation learning methods can be classified into two categories: generative models that learn the underlying connectivity distribution in a graph, and discriminative models that predict the probability of edge between a pair of vertices. In this paper, we propose GraphGAN, an innovative graph representation learning framework unifying the above two classes of methods, in which the generative and the discriminative model play a game-theoretical minimax game. Specifically, for a given vertex, the generative model tries to fit its underlying true connectivity distribution over all other vertices and produces “fake” samples to fool the discriminative model, while the discriminative model tries to detect whether the sampled vertex is from ground truth or generated by the generative model. With the competition between these two models, both of them can alternately and iteratively boost their performance. Moreover, we propose a novel graph softmax as the implementation of the generative model to overcome the limitations of traditional softmax function, which can be proven satisfying desirable properties of normalization, graph structure awareness, and computational efficiency. Through extensive experiments on real-world datasets, we demonstrate that GraphGAN achieves substantial gains in a variety of applications, including graph reconstruction, link prediction, node classification, recommendation, and visualization, over state-of-the-art baselines.

Reinforcement learning (RL) has been successfully applied to recommender systems. However, the existing RL-based recommendation methods are limited by their unstructured state/action representations. To address this limitation, we propose a novel way that builds high-quality graph-structured states/actions according to the user-item bipartite graph. More specifically, we develop an end-to-end RL agent, termed Graph Convolutional Q-network (GCQN), which is able to learn effective recommendation policies based on the inputs of the proposed graph-structured representations. We show that GCQN achieves significant performance margins over the existing methods, across different datasets and task settings.

Interactions between neighborhoods of two target nodes are often regarded as important clues for link prediction. In this paper, we propose a novel link prediction neural model named Neighborhood Interaction Attention Network (NIAN), which is able to automatically learn comprehensive neighborhood interaction features and predict links in an end-to-end way. The proposed model mainly consists of two attention layers. A node-level attention is designed to extract latent structure features of nodes in target neighborhoods. Based on the latent node features, a neighborhood-level attention is proposed to learn neighborhood interaction features by considering different importance of pair-wise interactions. The superiority of NIAN is demonstrated by extensive experiments on 6 benchmark datasets against 12 popular and state-of-the-art approaches.

Knowledge graphs capture structured information and relations between a set of entities or items. As such knowledge graphs represent an attractive source of information that could help improve recommender systems. However, existing approaches in this domain rely on manual feature engineering and do not allow for an end-to-end training. Here we propose Knowledge-aware Graph Neural Networks with Label Smoothness regularization (KGNN-LS) to provide better recommendations. Conceptually, our approach computes user-specific item embeddings by first applying a trainable function that identifies important knowledge graph relationships for a given user. This way we transform the knowledge graph into a user-specific weighted graph and then apply a graph neural network to compute personalized item embeddings. To provide better inductive bias, we rely on label smoothness assumption, which posits that adjacent items in the knowledge graph are likely to have similar user relevance labels/scores. Label smoothness provides regularization over the edge weights and we prove that it is equivalent to a label propagation scheme on a graph. We also develop an efficient implementation that shows strong scalability with respect to the knowledge graph size. Experiments on four datasets show that our method outperforms state of the art baselines. KGNN-LS also achieves strong performance in cold-start scenarios where user-item interactions are sparse.

While deep reinforcement learning has been successfully applied to recommender systems, it is challenging and unexplored to improve the performance of deep reinforcement learning recommenders by effectively utilizing the pervasive social networks. In this work, we develop a Social Attentive Deep Q-network (SADQN) agent, which is able to provide high-quality recommendations during user-agent interactions by leveraging social influence among users. Specifically, SADQN is able to estimate action-values not only based on the users' personal preferences, but also based on their social neighbors' preferences by employing a particular social attention layer. The experimental results on three real-world datasets demonstrate that SADQN significantly improves the performance of deep reinforcement learning agents that overlook social influence.

To alleviate sparsity and cold start problem of collaborative filtering based recommender systems, researchers and engineers usually collect attributes of users and items, and design delicate algorithms to exploit these additional information. In general, the attributes are not isolated but connected with each other, which forms a knowledge graph (KG). In this paper, we propose Knowledge Graph Convolutional Networks (KGCN), an end-to-end framework that captures inter-item relatedness effectively by mining their associated attributes on the KG. To automatically discover both high-order structure information and semantic information of the KG, we sample from the neighbors for each entity in the KG as their receptive field, then combine neighborhood information with bias when calculating the representation of a given entity. The receptive field can be extended to multiple hops away to model high-order proximity information and capture users' potential long-distance interests. Moreover, we implement the proposed KGCN in a minibatch fashion, which enables our model to operate on large datasets and KGs. We apply the proposed model to three datasets about movie, book, and music recommendation, and experiment results demonstrate that our approach outperforms strong recommender baselines.

Collaborative filtering often suffers from sparsity and cold start problems in real recommendation scenarios, therefore, researchers and engineers usually use side information to address the issues and improve the performance of recommender systems. In this paper, we consider knowledge graphs as the source of side information. We propose MKR, a Multi-task feature learning approach for Knowledge graph enhanced Recommendation. MKR is a deep end-to-end framework that utilizes knowledge graph embedding task to assist recommendation task. The two tasks are associated by crosscompress units, which automatically share latent features and learn high-order interactions between items in recommender systems and entities in the knowledge graph. We prove that crosscompress units have sufficient capability of polynomial approximation, and show that MKR is a generalized framework over several representative methods of recommender systems and multi-task learning. Through extensive experiments on real-world datasets, we demonstrate that MKR achieves substantial gains in movie, book, music, and news recommendation, over state-of-the-art baselines. MKR is also shown to be able to maintain satisfactory performance even if user-item interactions are sparse.

In this article, we study a multi-step interactive recommendation problem for explicit-feedback recommender systems. Different from the existing works, we propose a novel user-specific deep reinforcement learning approach to the problem. Specifically, we first formulate the problem of interactive recommendation for each target user as a Markov decision process (MDP). We then derive a multi-MDP reinforcement learning task for all involved users. To model the possible relationships (including similarities and differences) between different users’ MDPs, we construct user-specific latent states by using matrix factorization. After that, we propose a user-specific deep Q-learning (UDQN) method to estimate optimal policies based on the constructed user-specific latent states. Furthermore, we propose Biased UDQN (BUDQN) to explicitly model user-specific information by employing an additional bias parameter when estimating the Q-values for different users. Finally, we validate the effectiveness of our approach by comprehensive experimental results and analysis.

In this article, we aim at developing a user representation learning model to solve the information diffusion prediction problem in social media. The main idea is to project the diffusion users into a continuous latent space as the role-based (sender and receiver) representations, which capture unique diffusion characteristics of users. The model learns the role-based representations based on a cascade modeling objective that aims at maximizing the likelihood of observed cascades, and employs the matrix factorization objective of reconstructing structural proximities as a regularization on representations. By jointly embedding the information of cascades and network, the learned representations are robust on different diffusion data. We evaluate the proposed model on three real-world datasets. The experimental results demonstrate the better performance of the proposed model than state-of-the-art diffusion embedding and network embedding models and other popular graph-based methods.

To address the sparsity and cold-start problem of collaborative filtering, researchers usually make use of side information, such as social networks or item attributes, to improve the performance of recommendation. In this article, we consider the knowledge graph (KG) as the source of side information. To address the limitations of existing embedding-based and path-based methods for KG-aware recommendation, we propose RippleNet , an end-to-end framework that naturally incorporates the KG into recommender systems. RippleNet has two versions: (1) The outward propagation version, which is analogous to the actual ripples on water, stimulates the propagation of user preferences over the set of knowledge entities by automatically and iteratively extending a user’s potential interests along links in the KG. The multiple “ripples” activated by a user’s historically clicked items are thus superposed to form the preference distribution of the user with respect to a candidate item. (2) The inward aggregation version aggregates and incorporates the neighborhood information biasedly when computing the representation of a given entity. The neighborhood can be extended to multiple hops away to model high-order proximity and capture users’ long-distance interests. In addition, we intuitively demonstrate how a KG assists with recommender systems in RippleNet, and we also find that RippleNet provides a new perspective of explainability for the recommended results in terms of the KG. Through extensive experiments on real-world datasets, we demonstrate that both versions of RippleNet achieve substantial gains in a variety of scenarios, including movie, book, and news recommendations, over several state-of-the-art baselines.

In this paper, we propose a novel sequential neural network with structure attention to model information diffusion. The proposed model explores both sequential nature of an information diffusion process and structural characteristics of user connection graph. The recurrent neural network framework is employed to model the sequential information. The attention mechanism is incorporated to capture the structural dependency among users, which is defined as the diffusion context of a user. A gating mechanism is further developed to effectively integrate the sequential and structural information. The proposed model is evaluated on the diffusion prediction task. The performances on both synthetic and real datasets demonstrate its superiority over popular baselines and state-of-the-art sequence-based models.

Session-based recommendation performance has been significantly improved by Recurrent Neural Networks (RNN). However, existing RNN-based models do not expose the global knowledge of frequent click patterns or consider variability of sequential behaviors in sessions. In this paper, we propose a novel Variational Recurrent Model (VRM), which employs the stochastic latent variable to capture the knowledge of frequent click patterns and impose variability for the sequential behavior modeling. A stochastic generative process of session sequence is specified, where the latent variable modulates the generation of session sequences in RNN. We further extend VRM to a Conditional Variational Recurrent Model (CVRM) by considering additional information (e.g., focused category in sessions) as the generative condition. When evaluated on a public benchmark dataset, VRM and its extension clearly demonstrate their superiority over popular baselines and state-of-the-art models.

To address the sparsity and cold start problem of collaborative filtering, researchers usually make use of side information, such as social networks or item attributes, to improve recommendation performance. This paper considers the knowledge graph as the source of side information. To address the limitations of existing embedding-based and path-based methods for knowledge-graph-aware recommendation, we propose RippleNet, an end-to-end framework that naturally incorporates the knowledge graph into recommender systems. Similar to actual ripples propagating on the water, RippleNet stimulates the propagation of user preferences over the set of knowledge entities by automatically and iteratively extending a user's potential interests along links in the knowledge graph. The multiple "ripples" activated by a user's historically clicked items are thus superposed to form the preference distribution of the user with respect to a candidate item, which could be used for predicting the final clicking probability. Through extensive experiments on real-world datasets, we demonstrate that RippleNet achieves substantial gains in a variety of scenarios, including movie, book and news recommendation, over several state-of-the-art baselines.

Language discrepancy is inherent and be part of human languages. Thereby, the same sentiment would be expressed in different patterns across different languages. Unfortunately, the language discrepancy is overlooked by existing works of cross-lingual sentiment analysis. How to accommodate the inherent language discrepancy in sentiment for better cross-lingual sentiment analysis is still an open question. In this paper, we aim to model the language discrepancy in sentiment expressions as intrinsic bilingual polarity correlations (IBPCs) for better cross-lingual sentiment analysis. Specifically, given a document of source language and its translated counterpart, we firstly devise a sentiment representation learning phase to extract monolingual sentiment representation for each document in this pair separately. Then, the two sentiment representations are transferred to be the points in a shared latent space, named hybrid sentiment space. The language discrepancy is then modeled as a fixed transfer vector under each particular polarity between the source and target languages in this hybrid sentiment space. Two relation-based bilingual sentiment transfer models (i.e., RBST-s, RBST-hp) are proposed to learn the fixped transfer vectors. The sentiment of a target-language document is then determined based on the transfer vector between it and its translated counterpart in the hybrid sentiment space. Extensive experiments over a real-world benchmark dataset demonstrate the superiority of the proposed models against several state-of-the-art alternatives.

Exploring the mechanism that explains how a user's opinion changes under the influence of his/her neighbors is of practical importance (e.g., for predicting the sentiment of his/her future opinion) and has attracted wide attention from both enterprises and academics.Though various opinion influence models have been proposed for opinion prediction, they only consider users' personal identities, but ignore their social identities with which people behave to fit the expectations of the others in the same group. In this work, we explore users' dual identities, including both personal identities and social identities to build a more comprehensive opinion influence model for a better understanding of opinion behaviors. A novel joint learning framework is proposed to simultaneously model opinion dynamics and detect social identity in a unified model. The effectiveness of the proposed approach is demonstrated through the experiments conducted on Twitter datasets

Pointwise and pairwise collaborative ranking are two major classes of algorithms for personalized item ranking. This paper proposes a novel joint learning method named alternating pointwise-pairwise learning (APPL) to improve ranking performance. APPL combines the ideas of both pointwise and pairwise learning, and is able to produce a more effective prediction model. The extensive experiments with both explicit and implicit feedback settings on four real-world datasets demonstrate that APPL performs significantly better than the state-of-the-art methods.

The massive amount of noisy and redundant information in text streams makes it a challenge for users to acquire timely and relevant information in social media. Real-time notification pushing on text stream is of practical importance. In this paper, we formulate the real-time pushing on text stream as a sequential decision making problem and propose a Neural Network based Reinforcement Learning (NNRL) algorithm for real-time decision making, e.g., push or skip the incoming text, with considering both history dependencies and future uncertainty. A novel Q-Network which contains a Long Short Term Memory (LSTM) layer and three fully connected neural network layers is designed to maximize the long-term rewards. Experiment results on the real data from TREC 2016 Real-time Summarization track show that our algorithm significantly outperforms state-of-the-art methods.

In this paper, we propose a predictive network representation learning (PNRL) model to solve the structural link prediction problem. The proposed model defines two learning objectives, i.e., observed structure preservation and hidden link prediction. To integrate the two objectives in a unified model, we develop an effective sampling strategy to select certain edges in a given network as assumed hidden links and regard the rest network structure as observed when training the model. By jointly optimizing the two objectives, the model can not only enhance the predictive ability of node representations but also learn additional link prediction knowledge in the representation space. Experiments on four real-world datasets demonstrate the superiority of the proposed model over the other popular and state-of-the-art approaches.

Micro-blog has been increasingly used for the public to express their opinions, and for organizations to detect public sentiment about social events or public policies. In this article, we examine and identify the key problems of this field, focusing particularly on the characteristics of innovative words, multi-media elements, and hierarchical structure of Chinese “Weibo.” Based on the analysis, we propose a novel approach and develop associated theoretical and technological methods to address these problems. These include a new sentiment word mining method based on three wording metrics and point-wise information, a rule set model for analyzing sentiment features of different linguistic components, and the corresponding methodology for calculating sentiment on multi-granularity considering emoticon elements as auxiliary affective factors. We evaluate our new word discovery and sentiment detection methods on a real-life Chinese micro-blog dataset. Initial results show that our new diction can improve sentiment detection, and they demonstrate that our multi-level rule set method is more effective, with the average accuracy being 10.2% and 1.5% higher than two existing methods for Chinese micro-blog sentiment analysis. In addition, we exploit visualization techniques to study the relationships between online sentiment and real life. The visualization of detected sentiment can help depict temporal patterns and spatial discrepancy.

Twitter, as one of the most popular social media platforms, provides a convenient way for people to communicate and interact with each other. It has been well recognized that influence exists during users' interactions. Some pioneer studies on finding influential users have been reported in the literature, but they do not distinguish different influence roles, which are of great value for various marketing purposes. In this paper, we move a step forward trying to further distinguish influence roles of Twitter users in a certain topic. By defining three views of features relating to topic, sentiment and popularity respectively, we propose a Multi-view Influence Role Clustering (MIRC) algorithm to group Twitter users into five categories. Experimental results show the effectiveness of the proposed approach in inferring influence roles.

With the tremendous amount of research publications, it has become increasingly important to provide a researcher with a rapid and accurate recommendation of a list of reference papers about a research field or topic. In this paper, we propose a unified graph model that can easily incorporate various types of useful information (e.g., content, authorship, citation and collaboration networks etc.) for efficient recommendation. The proposed model not only allows to thoroughly explore how these types of information can be better combined, but also makes personalized query-oriented reference paper recommendation possible, which as far as we know is a new issue that has not been explicitly addressed in the past. The experiments have demonstrated the clear advantages of personalized recommendation over non-personalized recommendation.

Automatic term extraction (ATE) aims at extracting domain-specific terms from a corpus of a certain domain. Termhood is one essential measure for judging whether a phrase is a term. Previous researches on termhood mainly depend on the word frequency information. In this paper, we propose to compute termhood based on semantic representation of words. A novel topic model, namely i-SWB, is developed to map the domain corpus into a latent semantic space, which is composed of some general topics, a background topic and a documents-specific topic. Experiments on four domains demonstrate that our approach outperforms the state-of-the-art ATE approaches.

Many real-world data mining problems contain hidden information (e.g., unobservable latent dependencies). We propose a perceptron-style method, latent structured perceptron, for fast discriminative learning of structured classification with hidden information. We also give theoretical analysis and demonstrate good convergence properties of the proposed method. Our method extends the perceptron algorithm for the learning task with hidden information, which can be hardly captured by traditional models. It relies on Viterbi decoding over latent variables, combined with simple additive updates. We perform experiments on one synthetic data set and two real-world structured classification tasks. Compared to conventional nonlatent models (e.g., conditional random fields, structured perceptrons), our method is more accurate on real-world tasks. Compared to existing heavy probabilistic models of latent variables (e.g., latent conditional random fields), our method lowers the training cost significantly (almost one order magnitude faster) yet with comparable or even superior classification accuracy. In addition, experiments demonstrate that the proposed method has good scalability on large-scale problems.

The classical bag-of-word models for information retrieval (IR) fail to capture contextual associations between words. In this article, we propose to investigate pure high-order dependence among a number of words forming an unseparable semantic entity, that is, the high-order dependence that cannot be reduced to the random coincidence of lower-order dependencies. We believe that identifying these pure high-order dependence patterns would lead to a better representation of documents and novel retrieval models. Specifically, two formal definitions of pure dependence—unconditional pure dependence (UPD) and conditional pure dependence (CPD)—are defined. The exact decision on UPD and CPD, however, is NP-hard in general. We hence derive and prove the sufficient criteria that entail UPD and CPD, within the well-principled information geometry (IG) framework, leading to a more feasible UPD/CPD identification procedure. We further develop novel methods for extracting word patterns with pure high-order dependence. Our methods are applied to and extensively evaluated on three typical IR tasks: text classification and text retrieval without and with query expansion.

Twitter, the most famous micro-blogging service and online social network, collects millions of tweets every day. Due to the length limitation, users usually need to explore other ways to enrich the content of their tweets. Some studies have provided findings to suggest that users can benefit from added hyperlinks in tweets. In this paper, we focus on the hyperlinks in Twitter and propose a new application, called hyperlink recommendation in Twitter. We expect that the recommended hyperlinks can be used to enrich the information of user tweets. A three-way tensor is used to model the user-tweet-hyperlink collaborative relations. Two tensor-based clustering approaches, tensor decomposition-based clustering (TDC) and tensor approximation-based clustering (TAC) are developed to group the users, tweets and hyperlinks with similar interests, or similar contexts. Recommendation is then made based on the reconstructed tensor using cluster information. The evaluation results in terms of Mean Absolute Error (MAE) shows the advantages of both the TDC and TAC approaches over a baseline recommendation approach, i.e., memory-based collaborative filtering. Comparatively, the TAC approach achieves better performance than the TDC approach.

Though numerous research has been devoted to social network discovery and analysis, relatively little research has been conducted on business network discovery. The main contribution of our research is the development of a novel probabilistic generative model for latent business networks mining. Our experimental results confirm that the proposed method outperforms the well-known vector space based model by 24% in terms of AUC value.

The paper is focused on blogosphere research based on the TREC blog distillation task, and aims to explore unbiased and significant features automatically and efficiently. Feedback from faceted feeds is introduced to harvest relevant features and information gain is used to select discriminative features. The evaluation result shows that the selected feedback features can greatly improve the performance and adapt well to the terabyte data.

The key issue of person name disambiguation is to discover different namesakes in massive web documents rather than simply cluster documents by using textual features. In this paper, we describe a novel person name disambiguation method based on social networks to effectively identify namesakes. The social network snippets in each document are extracted. Then, the namesakes are identified via splicing the social networks of each namesake by using the snippets as a bipartite graph. Experimental results show that our method achieves better result than the top performance of WePS-2 in identifying different namesakes.

Abstract Sentence clustering plays a pivotal role in theme‐based summarization, which discovers topic themes defined as the clusters of highly related sentences to avoid redundancy and cover more diverse information. As the length of sentences is short and the content it contains is limited, the bag‐of‐words cosine similarity traditionally used for document clustering is no longer suitable. Special treatment for measuring sentence similarity is necessary. In this article, we study the sentence‐level clustering problem. After exploiting concept‐ and context‐enriched sentence vector representations, we develop two co‐clustering frameworks to enhance sentence‐level clustering for theme‐based summarization—integrated clustering and interactive clustering—both allowing word and document to play an explicit role in sentence clustering as independent text objects rather than using word or concept as features of a sentence in a document set. In each framework, we experiment with two‐level co‐clustering (i.e., sentence‐word co‐clustering or sentence‐document co‐clustering) and three‐level co‐clustering (i.e., document‐sentence‐word co‐clustering). Compared against concept‐ and context‐oriented sentence‐representation reformation, co‐clustering shows a clear advantage in both intrinsic clustering quality evaluation and extrinsic summarization evaluation conducted on the Document Understanding Conferences (DUC) datasets.

Abstract The authors address the problem of unsupervised ensemble ranking. Traditional approaches either combine multiple ranking criteria into a unified representation to obtain an overall ranking score or to utilize certain rank fusion or aggregation techniques to combine the ranking results. Beyond the aforementioned “combine‐then‐rank” and “rank‐then‐combine” approaches, the authors propose a novel “rank‐learn‐combine” ranking framework, called Interactive Ranking (iRANK), which allows two base rankers to “teach” each other before combination during the ranking process by providing their own ranking results as feedback to the others to boost the ranking performance. This mutual ranking refinement process continues until the two base rankers cannot learn from each other any more. The overall performance is improved by the enhancement of the base rankers through the mutual learning mechanism. The authors further design two ranking refinement strategies to efficiently and effectively use the feedback based on reasonable assumptions and rational analysis. Although iRANK is applicable to many applications, as a case study, they apply this framework to the sentence ranking problem in query‐focused summarization and evaluate its effectiveness on the DUC 2005 and 2006 data sets. The results are encouraging with consistent and promising improvements.

Abstract In this article, the authors address the problem of sentence ranking in summarization. Although most existing summarization approaches are concerned with the information embodied in a particular topic (including a set of documents and an associated query) for sentence ranking, they propose a novel ranking approach that incorporates intertopic information mining. Intertopic information, in contrast to intratopic information, is able to reveal pairwise topic relationships and thus can be considered as the bridge across different topics. In this article, the intertopic information is used for transferring word importance learned from known topics to unknown topics under a learning‐based summarization framework. To mine this information, the authors model the topic relationship by clustering all the words in both known and unknown topics according to various kinds of word conceptual labels, which indicate the roles of the words in the topic. Based on the mined relationships, we develop a probabilistic model using manually generated summaries provided for known topics to predict ranking scores for sentences in unknown topics. A series of experiments have been conducted on the Document Understanding Conference (DUC) 2006 data set. The evaluation results show that intertopic information is indeed effective for sentence ranking and the resultant summarization system performs comparably well to the best‐performing DUC participating systems on the same data set.

A fundamental goal of unsupervised feature selection is denoising, which aims to identify and reduce noisy features that are not discriminative. Due to the lack of information about real classes, denoising is a challenging task. The noisy features can disturb the reasonable distance metric and result in unreasonable feature spaces, i.e., the feature spaces in which common clustering algorithms cannot effectively find real classes. To overcome the problem, we make a primary observation that the relevance of features is intrinsic and independent of any metric scaling on the feature space. This observation implies that feature selection should be invariant, at least to some extent, with respect to metric scaling. In this paper, we clarify the necessity of considering the metric invariance in unsupervised feature selection and propose a novel model incorporating metric invariance. Our proposed method is motivated by the following observations: if the statistic that guides the unsupervised feature selection process is invariant with respect to possible metric scaling, the solution of this model will also be invariant. Hence, if a metric-invariant model can distinguish discriminative features from noisy ones in a reasonable feature space, it will also work on the unreasonable counterpart transformed from the reasonable one by metric scaling. A theoretical justification of the metric invariance of our proposed model is given and the empirical evaluation demonstrates its promising performance.

Graph based sentence ranking algorithms such as PageRank and HITS have been successfully used in query-oriented summarization. With these algorithms, the documents to be summarized are often modeled as a text graph where nodes represent sentences and edges represent pairwise similarity relationships between two sentences. A deficiency of conventional graph modeling is its incapability of naturally and effectively representing complex group relationships shared among multiple objects. Simply squeezing complex relationships into pairwise ones will inevitably lead to loss of information which can be useful for ranking and learning. In this paper, we propose to take advantage of hypergraph, i.e. a generalization of graph, to remedy this defect. In a text hypergraph, nodes still represent sentences, yet hyperedges are allowed to connect more than two sentences. With a text hypergraph, we are thus able to integrate both group relationships formulated among multiple sentences and pairwise relationships formulated between two sentences in a unified framework. As essential work, it is first addressed in the paper that how a text hypergraph can be built for summarization by applying clustering techniques. Then, a hypergraph based semi-supervised sentence ranking algorithm is developed for query-oriented extractive summarization, where the influence of query is propagated to sentences through the structure of the constructed text hypergraph. When evaluated on DUC data sets, performance of the proposed approach is remarkable.

We address the problem of unsupervised ensemble ranking in this paper. Traditional approaches either combine multiple ranking criteria into a unified representation to obtain an overall ranking score or to utilize certain rank fusion or aggregation techniques to combine the ranking results. Beyond the aforementioned combine-then-rank and rank-then-combine approaches, we propose a novel rank-learn-combine ranking framework, called Interactive Ranking (iRANK), which allows two base rankers to "teach" each other before combination during the ranking process by providing their own ranking results as feedback to the others so as to boost the ranking performance. This mutual ranking refinement process continues until the two base rankers cannot learn from each other any more. The overall performance is improved by the enhancement of the base rankers through the mutual learning mechanism. We apply this framework to the sentence ranking problem in query-focused summarization and evaluate its effectiveness on the DUC 2005 data set. The results are encouraging with consistent and promising improvements.

Abstract Sentence ranking is the issue of most concern in document summarization today. While traditional feature‐based approaches evaluate sentence significance and rank the sentences relying on the features that are particularly designed to characterize the different aspects of the individual sentences, the newly emerging graph‐based ranking algorithms (such as the PageRank‐like algorithms) recursively compute sentence significance using the global information in a text graph that links sentences together. In general, the existing PageRank‐like algorithms can model well the phenomena that a sentence is important if it is linked by many other important sentences. Or they are capable of modeling the mutual reinforcement among the sentences in the text graph. However, when dealing with multidocument summarization these algorithms often assemble a set of documents into one large file. The document dimension is totally ignored. In this article we present a framework to model the two‐level mutual reinforcement among sentences as well as documents. Under this framework we design and develop a novel ranking algorithm such that the document reinforcement is taken into account in the process of sentence ranking. The convergence issue is examined. We also explore an interesting and important property of the proposed algorithm. When evaluated on the DUC 2005 and 2006 query‐oriented multidocument summarization datasets, significant results are achieved.

Sentence ranking is the issue of most concern in document summarization. Early researchers have presented the mutual reinforcement principle (MR) between sentence and term for simultaneous key phrase and salient sentence extraction in generic single-document summarization. In this work, we extend the MR to the mutual reinforcement chain (MRC) of three different text granularities, i.e., document, sentence and terms. The aim is to provide a general reinforcement framework and a formal mathematical modeling for the MRC. Going one step further, we incorporate the query influence into the MRC to cope with the need for query-oriented multi-document summarization. While the previous summarization approaches often calculate the similarity regardless of the query, we develop a query-sensitive similarity to measure the affinity between the pair of texts. When evaluated on the DUC 2005 dataset, the experimental results suggest that the proposed query-sensitive MRC (Qs-MRC) is a promising approach for summarization.

This paper presents an opinion analysis system based on linguistic knowledge which is acquired from small-scale annotated text and raw topic-relevant Web page. Based on the observation on the annotated opinion corpus, some word-, collocation- and sentence-level linguistic features for opinion analysis are discovered. Supervised and unsupervised learning techniques are developed to learn these features from annotated text and raw relevant Web page, respectively. These features are then incorporated into a classifier based on support vector machine (SVM) to identify opinionated sentences and determine their polarities. Evaluations show that the proposed opinion analysis system, namely OA, achieved promising performance, which shows the effectiveness of linguistic knowledge learning from relevant Web page.

Most up-to-date well-behaved topic-based summarization systems are built upon the extractive framework. They score the sentences based on the associated features by manually assigning or experimentally tuning the weights of the features. In this paper, we discuss how to develop learning strategies in order to obtain the optimal feature weights automatically, which can be used for assigning a sound score to a sentence characterized with a set of features. The two fundamental issues are about training data and learning models. To save the costly manual annotation time and effort, we construct the training data by labeling the sentence with a "true" score calculated according to human summaries. The Support Vector Regression (SVR) model is then used to learn how to relate the "true" score of the sentence to its features. Once the relations have been mathematically modeled, SVR is able to predict the "estimated" score for any given sentence. The evaluations by ROUGE-2 criterion on DUC 2006 and DUC 2005 document sets demonstrate the competitiveness and the adaptability of the proposed approaches.

The work presented in this paper is motivated by the practical need for content extraction, and the available data source and evaluation benchmark from the ACE program. The Chinese Entity Detection and Recognition (EDR) task is of particular interest to us. This task presents us several language-independent and language-dependent challenges, e.g. rising from the complication of extraction targets and the problem of word segmentation, etc. In this paper, we propose a novel solution to alleviate the problems special in the task. Mention detection takes advantages of machine learning approaches and character-based models. It manipulates different types of entities being mentioned and different constitution units (i.e. extents and heads) separately. Mentions referring to the same entity are linked together by integrating most-specific-first and closest-first rule based pairwise clustering algorithms. Types of mentions and entities are determined by head-driven classification approaches. The implemented system achieves ACE value of 66.1 when evaluated on the EDR 2005 Chinese corpus, which has been one of the top-tier results. Alternative approaches to mention detection and clustering are also discussed and analyzed.

In this paper, we propose a new strategy with time granularity reasoning for utilizing temporal information in topic tracking. Compared with previous ones, our work has four distinguished characteristics. Firstly, we try to determine a set of topic times for a target topic from the given on-topic stories. It helps to avoid the negative influence from other irrelevant times. Secondly, we take into account time granularity variance when deciding whether a coreference relationship exists between two times. Thirdly, both publication time and times presented in texts are considered. Finally, as time is only one attribute of a topic, we increase the similarity between a story and a target topic only when they are related not only temporally but also semantically. Experiments on two TDT corpora show that our method makes good use of temporal information in news stories.

In this research, we focus on tracking topics that originate and evolve from a specific event. Intuitively, a few key elements of a target event, such as date, location, and persons involved, would be enough for making a decision on whether a test story is on-topic. Consequently, a profile-based event tracking method is proposed. We attempt to build an event profile from the given on-topic stories by robust information retrieval technologies. A feature selection metric and a recognized event clause are utilized to determine most (if not all) key semantic elements of the target event. Preliminary experiments on the TDT2 mandarin corpus show that this profile-based event tracking method is promising.

Abstract Over the past few years, temporal information processing and temporal database management have increasingly become hot topics. Nevertheless, only a few researchers have investigated these areas in the Chinese language. This lays down the objective of our research: to exploit Chinese language processing techniques for temporal information extraction and concept reasoning. In this article, we first study the mechanism for expressing time in Chinese. On the basis of the study, we then design a general frame structure for maintaining the extracted temporal concepts and propose a system for extracting time‐dependent information from Hong Kong financial news. In the system, temporal knowledge is represented by different types of temporal concepts (TTC) and different temporal relations, including absolute and relative relations, which are used to correlate between action times and reference times. In analyzing a sentence, the algorithm first determines the situation related to the verb. This in turn will identify the type of temporal concept associated with the verb. After that, the relevant temporal information is extracted and the temporal relations are derived. These relations link relevant concept frames together in chronological order, which in turn provide the knowledge to fulfill users' queries, e.g., for question‐answering (i.e., Q&A) applications.