Search

A large number of servers are interconnected using a specific datacenter network to deliver the infrastructure as a service (IaaS). Multicast can jointly utilize the network resources and further reduce the consumption of network bandwidth more than individual unicast. The source of a multicast service, however, does not need to be in a specific location as long as certain constraints are satisfied. This means the multicast can have uncertain sources, which could reduce the network resource consumption more than a traditional multicast service and further improve the quality of service. In this paper, we propose a novel reliable multicast service with uncertain sources named ReMUS. The goal is to minimize the sum of the transfer cost and the recovery cost, although finding such a ReMUS is very challenging. Thus, we design a source-based multicast method to solve this problem by exploiting the flexibility of sources when no recovery nodes exist in the network. Furthermore, we design a general multicast method to jointly exploit the benefits of uncertain sources and recovery nodes to minimize the total cost of ReMUS. We conduct extensive evaluations under Internet2 and datacenter networks. The results indicate that our methods can efficiently realize the reliable and scalable multicast with uncertain sources, irrespective of the settings of networks and multicasts. To the best of our knowledge, we are the first to study the reliable multicast service under uncertain sources.

The hierarchical mobile edge computing satisfies the stringent latency requirements of data access and processing for emerging edge applications. The data location service is a basic function to provide data storage and retrieval to enable these applications. However, it still lacks research of a scalable and low-latency data location service in the environment. The existing solutions, such as DNS and DHT, fail to meet the requirement of those latency-sensitive applications. Therefore, in this article, we present a low-latency hybrid data-sharing framework, HDS. The HDS divides the data location service into two parts: intra-region and inter-region. More precisely, we design a data sharing protocol called Cuckoo Summary to achieve fast data localization in intra-region. Furthermore, for the inter-region data sharing, we develop a geographic routing based scheme to achieve efficient data localization with only one overlay hop. The advantages of HDS include short response latency, low implementation overhead, and few false positives. We implement the HDS framework based on a P4 prototype. The experimental results show that, compared to the state-of-the-art solutions, our design achieves 50.21% shorter lookup paths and 92.75% fewer false positives.

With the evolution of multi-access edge computing (MEC) and 5G communications, diverse services can be flexibly offered by multiple cache-enabled base stations (BSs) in dense deployments. This leads to an important request allocation problem, which studies how to direct user requests among multiple BSs. Efficient request allocation faces several challenges corresponding to the risk associated with the significant uncertainty in the MEC system. Thus, we formulate the request allocation in a risk-averse learning framework to minimize the expected user response delay, and also control the risk measured by the variance of uncertain return. However, solving this risk-averse optimization can be very difficult due to the high computation cost and limited computing resources in the MEC system. This further motivates us to convert the proposed model to a finite-sum composition optimization, and propose a new variant of composition stochastic variance-reduced gradient (C-SVRG) algorithm to accelerate parameter training by estimating the inner function on its linearization. Theoretical analysis proves linear convergence rate and significant complexity reduction of C-SVRG, and simulation results confirm its efficacy.

Large-scale datacenters are the key infrastructures of cloud computing. Inside a datacenter, a large number of servers are interconnected using a specific datacenter network to deliver the infrastructure as a service (IaaS) for tenants. To realize novel cloud applications like the network virtualization and network isolation among tenants, the principle of software-defined network (SDN) has been applied to datacenters. In the setting, multiple distributed controllers are deployed to offer a control plane over the entire datacenter to efficiently manage the network usage. Despite such efforts, cloud datacenters, however, still lack a scalable and resilient control plane. Consequently, this paper systematically studies the coverage problem of controllers, which means to cover all network devices using the least number of controllers. More precisely, we tackle this essential problem from three aspects, including the minimal coverage, the minimal fault-tolerant coverage, and the minimal communication overhead among controllers. After modelling and analyzing such three problems, we design efficient approaches to approximate the optimal solution, respectively. Extensive evaluation results indicate that our approaches can significantly save the number of required controllers, improve the fault-tolerant capability of the control plane and reduce the communication overhead of state synchronization among controllers. The design methodologies proposed in this paper can be applied to cloud datacenters with other networking structures after minimal modifications.

This paper provides theoretical and practical implications for the application of agent-based models (ABMs) to address the issue of coevolutionary business-IS alignment. The implications stem from ...

Basket-Sensitive Item Recommendation (BSIR) is a challenging task that aims to recommend an item to add to the current basket given a user’s historical behaviors. The recommended item is supposed to be relevant to the items in current basket. Previous works mainly produce a recommendation based on user’s current basket, ignoring the inherent preference released by user’s long-term behaviors and failing to accurately distinguish the item importance in the basket for detecting user intent. To tackle the above issues, we propose a hybrid model, i.e., Hybrid-Preference Neural Model (HPNM), where a user’s inherent preference is recognized by modeling the historical sequential baskets and the recent preference is identified by focusing on the current basket. In detail, we apply an attention mechanism for distinguishing the importance of items in a basket to generate an accurate basket representation. GRU is utilized for modeling the basket-level sequential information to obtain user’s long-term preference and the representation of the current session is regarded as user’s short-term preference. We evaluate the performance of our proposals against the state-of-the-art baselines in the field of BSIR on two public datasets, i.e., TaFeng and Foursquare. The experimental results show that HPNM can achieve obvious improvements against the baselines in terms of HLU and Recall. In addition, we find HPNM with an attention mechanism can lead to a larger improvement against the baseline for item recommendation in terms of HLU and Recall on testing baskets with relatively fewer items.

Session-aware recommendation is a special form of session-based recommendation, where users' historical interactions before the current session are available. Among the existing session-aware recommendation studies, recurrent neural network (RNN) is a popular choice to model users' current intent of the ongoing session as well as their general preference implied in previous sessions. However, these RNN-based methods present limited memory so as to have difficulty in characterizing long-term behaviour. In addition, when modeling a user's preference, existing studies mainly refer to his (her) own sessions, while ignore the collaborative information from sessions of other users that may share similar tastes. To tackle the above problems, we propose a neighbor-guided memory-based neural network (MNN) for session-aware recommendation task, which comprehensively considers users' short-term intent, long-term preference and cross-sessions information to yield the final recommendations. Specifically, we design a long-term memory generator to capture users' general preference from their historical sessions, meanwhile leverage the neighbor sessions of current session to obtain cross-session collaborative information. Furthermore, with the guidance of long-term memory and cross-session information, we employ a short-term memory generator to yield users' ongoing intent, which serves as the dominating part of recommendation. Extensive experimental results on three real world datasets show the effectiveness of our model, with improvements up to 12.0% on 30MUSIC, 26.5% on NOWPLAYING and 13.5% on TMALL in terms of Recall@20, respectively. Analysis on current session length and the number of historical interacted items comprehensively demonstrates the superiority of our proposal on processing long sessions and modeling users' long-term behaviour.

Collaborative filtering (CF) aims to make recommendations for users by detecting user’s preference from the historical user–item interactions. Existing graph neural networks (GNN) based methods achieve satisfactory performance by exploiting the high-order connectivity between users and items, however they suffer from the poor training efficiency problem and easily introduce bias for information propagation. Moreover, the widely applied Bayesian personalized ranking (BPR) loss is insufficient to provide supervision signals for training due to the extremely sparse observed interactions. To deal with the above issues, we propose the Efficient Graph Collaborative Filtering (EGCF) method. Specifically, EGCF adopts merely one-layer graph convolution to model the collaborative signal for users and items from the first-order neighbors in the user–item interactions. Moreover, we introduce contrastive learning to enhance the representation learning of users and items by deriving the self-supervisions, which is jointly trained with the supervised learning. Extensive experiments are conducted on two benchmark datasets, i.e., Yelp2018 and Amazon-book, and the experimental results demonstrate that EGCF can achieve the state-of-the-art performance in terms of Recall and normalized discounted cumulative gain (NDCG), especially on ranking the target items at right positions. In addition, EGCF shows obvious advantages in the training efficiency compared with the competitive baselines, making it practicable for potential applications.

The design of distributed control plane is an essential part of SDN. While there is an urgent need for verifying the control plane, little, however, is known about how to validate that the control plane offers assurable performance, especially across various failures. Such validation is hard due to two fundamental challenges. First, the number of potential failure scenarios could be exponential or even non-enumerable. Second, it is still an open problem to model the performance change when the control plane employs different failure recovery strategies. In this paper, we first characterize the validation of the distributed control plane as a robust optimization problem and further propose a robust validation framework to verify whether a control plane provides assurable performance across various failure scenarios and multiple failure recovery strategies. Then, we prove that identifying an optimal recovery strategy is NP-hard after developing an optimization model of failure recovery. Accordingly, we design two efficient failure recovery strategies, which can well approximate the optimal strategy and further exhibit good performance against potential failures. Furthermore, we design the capacity augmentation scheme when the control plane fails to accommodate the worst failure scenario even with the optimal failure recovery strategy. We have conducted extensive evaluations based on an SDN test bed and large-scale simulations over real network topologies. The evaluation results show the efficiency and effectiveness of the proposed validation framework.

Text representation, which converts text spans into real-valued vectors or matrices, is a crucial tool for machines to understand the semantics of text. Although most previous works employed classic methods based on statistics and neural networks, such methods might suffer from data sparsity and insensitivity to the text structure, respectively. To address the above drawbacks, we propose a general and structure-sensitive framework, i.e., the hierarchical architecture. Specifically, we incorporate the hierarchical architecture into three existing neural network models for document representation, thereby producing three new representation models for document classification, i.e., TextHFT, TextHRNN, and TextHCNN. Our comprehensive experimental results on two public datasets demonstrate the effectiveness of the hierarchical architecture. With a comparable (or substantially less) time expense, our proposals obtain significant improvements ranging from 4.65 to 35.08% in terms of accuracy against the baseline. We can conclude that the hierarchical architecture can enhance the classification performance. In addition, we find that the benefits provided by the hierarchical architecture can be strengthened as the document length increases.

Answer selection is an essential step in a question answering (QA) system. Traditional methods for this task mainly focus on developing linguistic features that are limited in practice. With the great success of deep learning method in distributed text representation, deep learning-based answer selection approaches have been well investigated, which mainly employ only one neural network, i.e., convolutional neural network (CNN) or long short term memory (LSTM), leading to failures in extracting some rich sentence features. Thus, in this paper, we propose a collaborative learning-based answer selection model (QA-CL), where we deploy a parallel training architecture to collaboratively learn the initial word vector matrix of the sentence by CNN and bidirectional LSTM (BiLSTM) at the same time. In addition, we extend our model by incorporating the sentence embedding generated by the QA-CL model into a joint distributed sentence representation using a strong unsupervised baseline weight removal (WR), i.e., the QA-CLWR model. We evaluate our proposals on a popular QA dataset, InsuranceQA. The experimental results indicate that our proposed answer selection methods can produce a better performance compared with several strong baselines. Finally, we investigate the models’ performance with respect to different question types and find that question types with a medium number of questions have a better and more stable performance than those types with too large or too small number of questions.

Answer selection is a crucial subtask in the question answering (QA) system. Conventional avenues for this task mainly concentrate on developing linguistic tools that are limited in both performance and practicability. Answer selection approaches based on deep learning have been well investigated with the tremendous success of deep learning in natural language processing. However, the traditional neural networks employed in existing answer selection models, i.e., recursive neural network or convolutional neural network, typically suffer from obtaining the global text information due to their operating mechanisms. The recent Transformer neural network is considered to be good at extracting the global information by employing only self-attention mechanism. Thus, in this paper, we design a Transformer-based neural network for answer selection, where we deploy a bidirectional long short-term memory (BiLSTM) behind the Transformer to acquire both global information and sequential features in the question or answer sentence. Different from the original Transformer, our Transformer-based network focuses on sentence embedding rather than the seq2seq task. In addition, we employ a BiLSTM rather than utilizing the position encoding to incorporate sequential features as the universal Transformer does. Furthermore, we apply three aggregated strategies to generate sentence embeddings for question and answer, i.e., the weighted mean pooling, the max pooling, and the attentive pooling, leading to three corresponding Transformer-based models, i.e., QA-TF $_{{W\!P}}$ , QA-TF $_{{M\!P}}$ , and QA-TF $_{{A\!P}}$ , respectively. Finally, we evaluate our proposals on a popular QA dataset WikiQA. The experimental results demonstrate that our proposed Transformer-based answer selection models can produce a better performance compared with several competitive baselines. In detail, our best model outperforms the state-of-the-art baseline by up to 2.37%, 2.83%, and 3.79% in terms of MAP, MRR, and accuracy, respectively.

Session-based recommendation aims to model a user’s intent and predict an item that the user may interact with in the next step based on an ongoing session. Existing session-based recommender systems mainly aim to model the sequential signals based on Recurrent Neural Network (RNN) structures or the item transition relations between items with Graph Neural Network (GNN) based frameworks to identify a user’s intent for recommendation. However, in real scenarios, there may be strong sequential signals existing in users’ adjacent behaviors or multi-step transition relations among different items. Thus, either RNN- or GNN-based methods can only capture limited information for modeling complex user behavior patterns. RNNs pay attention to the sequential relations among consecutive items, while GNNs focus on structural information, i.e., how to enrich the item embedding with its adjacent items. In this paper, we propose a Collaborative Co-attention Network for Session-based Recommendation (CCN-SR) to incorporate both sequential and structural information, as well as capture the co-relations between them for obtaining an accurate session representation. To be specific, we first model the ongoing session with an RNN structure to capture the sequential information among items. Meanwhile, we also construct a session graph to learn the item representations with a GNN structure. Then, we design a co-attention network upon these two structures to capture the mutual information between them. The designed co-attention network can enrich the representation of each node in the session with both sequential and structural information, and thus generate a more comprehensive representation for each session. Extensive experiments are conducted on two public e-commerce datasets, and the results demonstrate that our proposed model outperforms state-of-the-art baseline model for session based recommendation in terms of both Recall and MRR. We also investigate different combination strategies and the experimental results verify the effectiveness of our proposed co-attention mechanism. Besides, our CCN-SR model achieves better performance than baseline models with different session lengths.

Session-based recommendation is a challenging task. Without access to a user's historical user-item interactions, the information available in an ongoing session may be very limited. Previous work on session-based recommendation has considered sequences of items that users have interacted with sequentially. Such item sequences may not fully capture complex transition relationship between items that go beyond inspection order. Thus graph neural network (GNN) based models have been proposed to capture the transition relationship between items. However, GNNs typically propagate information from adjacent items only, thus neglecting information from items without direct connections. Importantly, GNN-based approaches often face serious overfitting problems. We propose Star Graph Neural Networks with Highway Networks (SGNN-HN) for session-based recommendation. The proposed SGNN-HN applies a star graph neural network (SGNN) to model the complex transition relationship between items in an ongoing session. To avoid overfitting, we employ highway networks (HN) to adaptively select embeddings from item representations. Finally, we aggregate the item embeddings generated by the SGNN in an ongoing session to represent a user's final preference for item prediction. Experiments on two public benchmark datasets show that SGNN-HN can outperform state-of-the-art models in terms of P@20 and MRR@20 for session-based recommendation.

Given an occurred event, human can easily predict the next event or reason the preceding event, yet which is difficult for machine to perform such event reasoning. Event representation bridges the connection and targets to model the process of event reasoning as a machine-readable format, which then can support a wide range of applications in information retrieval, e.g., question answering and information extraction. Existing work mainly resorts to a joint training to integrate all levels of training loss in event chains by a simple loss summation, which is easily trapped into a local optimum. In addition, the scenario knowledge in event chains is not well investigated for event representation. In this paper, we propose a unified fine-tuning architecture, incorporated with scenario knowledge for event representation, i.e., UniFA-S, which mainly consists of a unified fine-tuning architecture (UniFA) and a scenario-level variational auto-encoder (S-VAE). In detail, UniFA employs a multi-step fine-tuning to integrate all levels of training and S-VAE applies a stochastic variable to implicitly represent the scenario-level knowledge. We evaluate our proposal from two aspects, i.e., the representation and inference abilities. For the representation ability, our ensemble model UniFA-S can beat state-of-the-art baselines for two similarity tasks. For the inference ability, UniFA-S can outperform the best baseline, achieving 4.1%-8.2% improvements in terms of accuracy for various inference tasks.

Automatic speed control is a promising way to improve the quality of urban traffic operation. This paper proposes a fuzzy logic controller to track speed and time trajectories simultaneously for trams. Four modules are designed to improve the effectiveness of the controller. Specifically, the fuzzy inference module is to generate the recommended acceleration. Moreover, the train speed protection module and the train traction characteristic module are presents to ensure that the control laws can satisfy with the operation constraints. To verify the performance of the fuzzy logic controller, the speed and time tracking errors are compared with the PID controller based on the predefined speed profiles. Simulation results show that the proposed controller achieve the trade-off between speed and time tracking within the allowable range. This controller can be implemented by the automatic train operation systems.

Edge computing satisfies the stringent latency requirements of data access and processing for applications running on edge devices. The data location service is a key function to provide data storage and retrieval to enable these applications. However, it still lacks research of a scalable and low-latency data location service in mobile edge computing. Meanwhile, the existing solutions, such as DNS and DHT, fail to meet the low latency requirement of mobile edge computing. This paper presents a low-latency hybrid data-sharing framework, HDS, in which the data location service is divided into two parts: intra-region and inter-region. In the intra-region part, we design a data sharing protocol called Cuckoo Summary to achieve fast data localization. In the inter-region part, we develop a geographic routing based scheme to achieve efficient data localization with only one overlay hop. The advantages of HDS include short response latency, low implementation overhead, and few false positives. We implement our HDS framework based on a P4 prototype. The experimental results show that, compared to the state-of-the-art solutions, our design achieves 50.21% shorter lookup paths and 92.75% fewer false positives.

Given an incomplete event chain, script learning aims to predict the missing event, which can support a series of NLP applications. Existing work cannot well represent the heterogeneous relations and capture the discontinuous event segments that are common in the event chain. To address these issues, we introduce a heterogeneous-event (HeterEvent) graph network. In particular, we employ each unique word and individual event as nodes in the graph, and explore three kinds of edges based on realistic relations (e.g., the relations of word-and-word, word-and-event, event-and-event). We also design a message passing process to realize information interactions among homo or heterogeneous nodes. And the discontinuous event segments could be explicitly modeled by finding the specific path between corresponding nodes in the graph. The experimental results on one-step and multi-step inference tasks demonstrate that our ensemble model HeterEvent[W+E] can outperform existing baselines.

Recently, dynamic business-IT alignment (BITA) has attracted increasing attentions. However, the description and the governance of dynamic BITA are insufficient in the literature. To address this gap, a conceptual BITA evolvement model is proposed combining enterprise architecture with the BITA maturity framework, and a sustainable BITA factor on dynamic resource allocation is introduced in this evolvement model. To discuss the sustainable factor in a comprehensive way, we explored a series of possible changes and corresponding resource influences within the alignment of business functions and applications. A dynamic resource allocation algorithm is proposed, and an illustrative case study is presented to verify the algorithm’s efficiency.

Because of the dynamic environments of business and IT, achieving any alignment between the two fields has become challenging. In view of its multiple viewpoints and artifacts, the discipline of enterprise architecture (EA) is often regarded as an effective methodology to deal with business-IT alignment (BITA) issues, and thus has attracted plenty of research. This article conducts a systematic literature review of BITA research using EA. Six questions are answered through 5W1H (When, Who, What, Why, Where, How) analysis; these questions aim to acquire a thorough understanding of BITA from the perspective of EA, to discover weak points in the status quo, and to identify future research directions.

Fact verification aims to retrieve related evidence from raw text to verify the correctness of a given claim. Existing works mainly leverage the single-granularity features for the representation learning of evidences, i.e., sentence features, ignoring other features like entity-level and context-level features. In addition, they usually focus on improving the prediction accuracy while lacking the interpretability of the inference process, which leads to unreliable results. Thus, in this paper, to investigate how to utilize multi-granularity semantic units for evidence representation as well as to improve the explainability of evidence reasoning, we propose a Hierarchical Reasoning-based Heterogeneous Graph Neural Network for fact verification (HHGN). HHGN combines multiple features of entity, sentence as well as context for evidence representation, and employs a heterogeneous graph to capture their semantic relations. Inspired by the human inference process, we design a hierarchical reasoning-based node updating strategy to propagate the evidence features. Then, we extract the potential reasoning paths from the graph to predict the label, which aggregates the results of different paths weighted by their relevance to the claim. We evaluate our proposal on FEVER, a large-scale benchmark dataset for fact verification. Our experimental results demonstrate the superiority of HHGN over the competitive baselines in both single evidence and multiple evidences settings. In addition, HHGN presents reasonable interpretability in the form of aggregating the features of relevant entity units and selecting the evidence sentences with high confidence.

We propose a J-NCF method for recommender systems. The J-NCF model applies a joint neural network that couples deep feature learning and deep interaction modeling with a rating matrix. Deep feature learning extracts feature representations of users and items with a deep learning architecture based on a user-item rating matrix. Deep interaction modeling captures non-linear user-item interactions with a deep neural network using the feature representations generated by the deep feature learning process as input. J-NCF enables the deep feature learning and deep interaction modeling processes to optimize each other through joint training, which leads to improved recommendation performance. In addition, we design a new loss function for optimization, which takes both implicit and explicit feedback, point-wise and pair-wise loss into account. Experiments on several real-word datasets show significant improvements of J-NCF over state-of-the-art methods, with improvements of up to 8.24% on the MovieLens 100K dataset, 10.81% on the MovieLens 1M dataset, and 10.21% on the Amazon Movies dataset in terms of HR@10. NDCG@10 improvements are 12.42%, 14.24% and 15.06%, respectively. We also conduct experiments to evaluate the scalability and sensitivity of J-NCF. Our experiments show that the J-NCF model has a competitive recommendation performance with inactive users and different degrees of data sparsity when compared to state-of-the-art baselines., 30 pages

Business-IT alignment (BITA) is an important research problem of military information system architecture (MISA). Extant research has achieved and controlled the MISA to alleviate the BITA issue from the perspective of architecture management. This article attempts to address the problem in the architecture design phase in advance, in order to maintain a suitable alignment level at the beginning. For problems correlating with BITA in the design phase, functions-systems alignment (FSA) is selected as a study point, and portfolio decision analysis (PDA) is utilized to discuss information systems portfolio problem in it. Method feasibility is demonstrated through an anti-missile case. This paper contributes to improving the alignment quality of MISA by considering the PDA method.

Edge computing is a new paradigm in which the computing and storage resources are placed at the edge of the Internet. Data placement and retrieval are fundamental services of edge computing when a network of edge servers collaboratively provide data storage. These services require short-latency and low-overhead implementation in network devices and load balance on edge servers. However existing methods such as distributed hash tables (DHTs) are not able to achieve efficient data placement and retrieval services in the edge computing environment. This paper presents GRED, an efficient data placement and retrieval service for edge computing, which is efficient in not only the load balance but also routing path lengths and forwarding table sizes. GRED utilizes the software-defined networking paradigm to support a virtual-space based DHT with only one overlay hop. We implement GRED in a P4 prototype. Experimental results show that GRED uses

Hybrid electric trams equip with additional on-board energy storage devices to improve the performance of power sources. Both of optimal energy management and velocity control are important to maximum the energy efficiency for catenary-supercapacitors hybrid trams. An optimal control model has been developed to minimize energy consumption from traction substations with supercapacitors voltage limitations. Multiple phases model transformation and solving methods are implemented to obtain the optimal voltage trajectory and speed profile simultaneously. Finally, the effect of trip time on energy consumption is assessed.

Edge computing promises a dramatic reduction in the network latency and the traffic volume, where many edge servers are placed at the edge of the Internet. Furthermore, these edge servers cache data to provide services for edge users. The data sharing among edge servers can effectively shorten the latency to retrieve the data and further reduce the network bandwidth consumption. The key challenge is to construct an efficient data indexing mechanism no matter how the data is cached in the edge network. Although this is essential, it is still an open problem. Moreover, existing methods such as the centralized indexing and the DHT indexing in other fields fail to meet the performance demand of edge computing. This paper presents a COordinate-based INdexing (COIN) mechanism for the data sharing in edge computing. COIN maintains a virtual space where the switches and the data indexes are associated with the coordinates. Then, COIN distributes data indexes to indexing edge servers based on those coordinates. The COIN is effective because any query request from an edge server can be responded when the data has been stored in the edge network. More importantly, COIN is efficient in both routing path lengths and forwarding table sizes for publishing/querying the data indexes. We implement COIN in a P4 prototype. Experimental results show that COIN uses 59% shorter path length and 30% less forwarding table entries to retrieve the data index compared to using Chord, a well-known DHT solution.

Session-based recommendation is the task of recommending the next item a user might be interested in given partially known session information, e.g., part of a session or recent historical sessions. An effective session-based recommender should be able to exploit a user's evolving preferences, which we assume to be a mixture of her short- and long-term interests. Existing session-based recommendation methods often embed a user's long-term preference into a static representation, which plays a fixed role when dealing with her current short-term interests. This is problematic because long-term preferences may be more or less important for predicting the next conversion depending on the user's short-term interests. We propose a DCN-SR. DCN-SR applies a co-attention network to capture the dynamic interactions between the user's long- and short-term interaction behavior and generates co-dependent representations of the user's long- and short-term interests. For modeling a user's short-term interaction behavior, we design a CGRU network to take actions like "click'', "collect'' and "buy'' into account. Experiments on e-commerce datasets show significant improvements of DCN-SR over state-of-the-art session-based recommendation methods, with improvements of up to 2.58% on the Tmall dataset and 3.08% on the Tianchi dataset in terms of Recall@10. MRR@10 improvements are 3.78% and 4.05%, respectively. We also investigate the scalability and sensitivity of DCN-SR. The improvements of DCN-SR over state-of-the-art baselines are especially noticeable for short sessions and active users with many historical interactions.

Fact verification aims to retrieve relevant evidence from a knowledge base, e.g., Wikipedia, to verify the given claims. Existing methods only consider the sentence-level semantics for evidence representations, which typically neglect the importance of fine-grained features in the evidence-related sentences. In addition, the interpretability of the reasoning process has not been well studied in the field of fact verification. To address such issues, we propose an entity-graph based reasoning method for fact verification abbreviated as RoEG, which generates the fine-grained features of evidence at the entity-level and models the human reasoning paths based on an entity graph. In detail, to capture the semantic relations of retrieved evidence, RoEG introduces the entities as nodes and constructs the edges in the graph based on three linking strategies. Then, RoEG utilizes a selection gate to constrain the information propagation in the sub-graph of relevant entities and applies a graph neural network to propagate the entity-features for reasoning. Finally, RoEG employs an attention aggregator to gather the information of entities for label prediction. Experimental results on a large-scale benchmark dataset FEVER demonstrate the effectiveness of our proposal by beating the competitive baselines in terms of label accuracy and FEVER Score. In particular, for a task of multiple-evidence fact verification, RoEG produces 5.48% and 4.35% improvements in terms of label accuracy and FEVER Score against the state-of-the-art baseline. In addition, RoEG shows a better performance when more entities are involved for fact verification.

Incorporating topic information can help response generation models to produce informative responses for chat-bots. Previous work only considers the individual semantic of each topic, ignoring its specific dialog context, which may result in inaccurate topic representation and hurt response coherence. Besides, as an important feature of multi-turn conversation, dynamic topic transitions have not been well-studied. We propose a Context-Controlled Topic-Aware neural response generation model, i.e., CCTA, which makes dialog context interact with the process of topic representing and transiting to achieve balanced improvements on response informativeness and contextual coherence. CCTA focuses on capturing the semantical relations within topics as well as their corresponding contextual information in conversation, to produce context-dependent topic representations at the word-level and turn-level. Besides, CCTA introduces a context-controlled topic transition strategy, utilizing contextual topics to yield relevant transition words. Extensive experimental results on two benchmark multi-turn conversation datasets validate the superiority of our proposal on generating coherent and informative responses against the state-of-the-art baselines. We also find that topic transition modeling can work as an auxiliary learning task to boost the response generation.

A query submitted to a search engine provides limited information about the searcher’s real search intent. Alternatively, other information from user’s historical behaviors, e.g., clicks and dwell time, can provide a strong clue to identify the search purpose. In this paper, we: (1) investigate the impact of distributions of users and queries on reranking documents that are initially returned by a search engine; (2) perform tests for all users, i.e., both seen and unseen users in the training period. For unseen users, we explore the knowledge from their similar users seen in the training period. Our experiments show that integration of information from user behavior and document with an optimal weight outperforms combinations with a fixed tradeoff. On average, our model achieves near 3 % improvements than the best baseline approach in terms of metrics, e.g., MAP, P@K and NDCG@K.

In previous work on text summarization, encoder-decoder architectures and attention mechanisms have both been widely used. Attention-based encoder-decoder approaches typically focus on taking the sentences preceding a given sentence in a document into account for document representation, failing to capture the relationships between a sentence and sentences that follow it in a document in the encoder. We propose an attentive encoder-based summarization (AES) model to generate article summaries. AES can generate a rich document representation by considering both the global information of a document and the relationships of sentences in the document. A unidirectional recurrent neural network (RNN) and a bidirectional RNN are considered to construct the encoders, giving rise to unidirectional attentive encoder-based summarization (Uni-AES) and bidirectional attentive encoder-based summarization (Bi-AES), respectively. Our experimental results show that Bi-AES outperforms Uni-AES. We obtain substantial improvements over a relevant start-of-the-art baseline.

Query suggestions help users of a search engine to refine their queries. Previous work on query suggestion has mainly focused on incorporating directly observable features such as query co-occurrence and semantic similarity. The structure of such features is often set manually, as a result of which hidden dependencies between queries and users may be ignored. We propose an AHNQS model that combines a hierarchical structure with a session-level neural network and a user-level neural network to model the short- and long-term search history of a user. An attention mechanism is used to capture user preferences. We quantify the improvements of AHNQS over state-of-the-art RNN-based query suggestion baselines on the AOL query log dataset, with improvements of up to 21.86% and 22.99% in terms of MRR@10 and Recall@10, respectively, over the state-of-the-art; improvements are especially large for short sessions.

Users are strongly influenced by their friends or other users with similar interest. In this paper, we first provide a friends cluster identification approach based on the analysis of social features. Secondly, we propose our static recommendation approach based on an unbiased random walk strategy which simultaneously considers traditional recommendation approach and social relationship. Then, we further identify the change of both friendship and interest over time and propose our extended recommendation algorithm. Finally, we evaluate our approach on CiteULike and Last.fm dataset. Our experimental results demonstrate that the proposed algorithms can be very effective in recommending unknown items.

Document classification has a broad application in the field of sentiment classification, document ranking and topic labeling, etc. Previous neural network-based work has mainly focused on investigating a so-called forward implication, i.e., the preceding text segments are taken as the context of the following text segments when generating the text representation. Such a scenario typically ignores the fact that the semantics of a document are a product of the mutual implication of all text segments in a document. Thus, in this paper, we introduce a concept of interaction and propose a text representation model with Self-interaction Attention Mechanism (TextSAM) for document classification. In particular, we design three aggregated strategies to integrate the interaction into a hierarchical architecture for document classification, i.e., averaging the interaction, maximizing the interaction and adding one more attention layer on the interaction, which leads to three models, i.e., TextSAMAVE, TextSAMMAX and TextSAMATT, respectively. Our comprehensive experimental results on two public datasets, i.e., Yelp 2016 and Amazon Reviews (Electronics), show that our proposals can significantly outperform the state-of-the-art neural-based baselines for document classification, presenting a general improvement in terms of accuracy ranging from 5.97% to 14.05% against the best baseline. Furthermore, we find that our proposals with a self-interaction attention mechanism can obviously alleviate the impact brought by the increase of sentence number as the relative improvement of our proposals against the baselines are enlarged when the sentence number increases.

Document representation is widely used in practical application, for example, sentiment classification, text retrieval, and text classification. Previous work is mainly based on the statistics and the neural networks, which suffer from data sparsity and model interpretability, respectively. In this paper, we propose a general framework for document representation with a hierarchical architecture. In particular, we incorporate the hierarchical architecture into three traditional neural-network models for document representation, resulting in three hierarchical neural representation models for document classification, that is, TextHFT, TextHRNN, and TextHCNN. Our comprehensive experimental results on two public datasets, that is, Yelp 2016 and Amazon Reviews (Electronics), show that our proposals with hierarchical architecture outperform the corresponding neural-network models for document classification, resulting in a significant improvement ranging from 4.65% to 35.08% in terms of accuracy with a comparable (or substantially less) expense of time consumption. In addition, we find that the long documents benefit more from the hierarchical architecture than the short ones as the improvement in terms of accuracy on long documents is greater than that on short documents.

Query suggestions help users refine their queries after they input an initial query. We consider the task of generating query suggestions that are personalized and diversified. We propose a personalized query suggestion diversification model (PQSD), where a user's long-term search behavior is injected into a basic greedy query suggestion diversification model (G-QSD) that considers a user's search context in their current session. Query aspects are identified through clicked documents based on the Open Directory Project (ODP). We quantify the improvement of PQSD over a state-of-the-art baseline using the AOL query log and show that it beats the baseline in terms of metrics used in query suggestion ranking and diversification. The experimental results show that PQSD achieves the best performance when only queries with clicked documents are taken as search context rather than all queries.

Multicast can jointly utilize the network resources when delivering the same content to a set of destinations; hence, it can effectively reduce the consumption of network resources more than individual unicast. The source of a multicast, however, does not need to be in a specific location as long as certain constraints are satisfied. This means the multicast can have uncertain sources, which could reduce the network bandwidth consumption more than a traditional multicast. Meanwhile, a reliable multicast becomes crucial in providing reliable services for many important applications. However, a prior minimal cost forest (MCF) for such a new multicast is not designed to support reliable transmissions. In this paper, we propose a novel reliable multicast routing with uncertain sources named ReMUS. To the best of our knowledge, we are the first to study the reliable multicast under uncertain sources. The goal is to minimize the sum of the transfer cost and the recovery cost, although finding such a ReMUS is very challenging. Thus, we design a source-based multicast method to solve this problem by exploiting the flexibility of uncertain sources when no recovery nodes exist in the network. Furthermore, we design a general multicast method to jointly exploit the benefits of uncertain sources and recovery nodes to minimize the total cost of ReMUS. We conduct extensive evaluations based on the real topology of Internet2. The results indicate that our methods can efficiently realize the reliable and bandwidth-efficient multicast with uncertain sources, irrespective of the settings of networks and multicasts.

Query autocompletion (QAC) is a common interactive feature of web search engines. It aims at assisting users to formulate queries and avoiding spelling mistakes by presenting them with a list of query completions as soon as they start typing in the search box. Existing QAC models mostly rank the query completions by their past popularity collected in the query logs. For some queries, their popularity exhibits relatively stable or periodic behavior while others may experience a sudden rise in their query popularity. Current time-sensitive QAC models focus on either periodicity or recency and are unable to respond swiftly to such sudden rise, resulting in a less optimal QAC performance. In this paper, we propose a hybrid QAC model that considers two temporal patterns of query’s popularity, that is, periodicity and burst trend. In detail, we first employ the Discrete Fourier Transform (DFT) to identify the periodicity of a query’s popularity, by which we forecast its future popularity. Then the burst trend of query’s popularity is detected and incorporated into the hybrid model with its cyclic behavior. Extensive experiments on a large, real-world query log dataset infer that modeling the temporal patterns of query popularity in the form of its periodicity and its burst trend can significantly improve the effectiveness of ranking query completions.

Query suggestions help users refine their queries after they input an initial query. Previous work mainly concentrated on similarity-based and context-based query suggestion approaches. However, models that focus on adapting to a specific user (personalization) can help to improve the probability of the user being satisfied. In this paper, we propose a personalized query suggestion model based on users’ search behavior (UB model), where we inject relevance between queries and users’ search behavior into a basic probabilistic model. For the relevance between queries, we consider their semantical similarity and co-occurrence which indicates the behavior information from other users in web search. Regarding the current user’s preference to a query, we combine the user’s short-term and long-term search behavior in a linear fashion and deal with the data sparse problem with Bayesian probabilistic matrix factorization (BPMF). In particular, we also investigate the impact of different personalization strategies (the combination of the user’s short-term and long-term search behavior) on the performance of query suggestion reranking. We quantify the improvement of our proposed UB model against a state-of-the-art baseline using the public AOL query logs and show that it beats the baseline in terms of metrics used in query suggestion reranking. The experimental results show that: (i) for personalized ranking, users’ behavioral information helps to improve query suggestion effectiveness; and (ii) given a query, merging information inferred from the short-term and long-term search behavior of a particular user can result in a better performance than both plain approaches.