Search

Due to the high flexibility of the spectral shape generated by a spectrally tunable multichannel LED light source, the traditional spectral estimation by a digital camera has difficulty in the training procedure and cannot reach a satisfactory accuracy. Considering the characteristics of the linear system for both lighting and imaging equipment, a method based on matrix factorization was proposed by capturing a standard color chart with its known reflectance illuminated under the individual channels of the multichannel LED light source. The experimental results indicate that the proposed method outperforms the traditional one, in terms of spectral accuracy and colorimetric difference, with the upper bound performance being higher than 0.999 goodness of fit coefficient (GFC) and less than 1.0 CIELAB unit, respectively.

Correlated color temperature (CCT) is one of the most important lighting factors to influence the subjective perceptions, while there currently are few researches involving a wide range of CCT to study the comfort and preference within specific indoor scenarios. Therefore a psychophysical experiment with the method of forced choice was carried out in a light-emitting diode (LED) light booth with a panel of twelve observers to evaluate these estimations with different combinations of CCT and illuminance. The results show that Kruithof’s rule is just a qualitative conclusion rather than an accurate expression about pleasing combination of CCT and illuminance, and that illuminance is not a significant factor to influence the estimations if it is limited within a moderate range. The detailed analysis indicates that CCT has significant impact on the subjective comfort and preference, and a set of models were proposed to predict these subjective perceptions for the scenarios of relaxing and working.

Objective To investigate the effects of microRNA-18a (miR-18a) on migration and invasion of hepatocellular carcinoma (HCC) cells, and its possible mechanism associated with Dicer l.Methods HepG2 and HepG2.2.15 cells were transfected with miR-18a inhibitor using Lipofectamine. Cell invasion was evaluated by transwell invasion assay, and cell migration was detected by transwell migration and wound-healing assays. Moreover, luciferase reporter assay was used to identify whether Dicer expression was regulated by miR-18a. Real-time RT-PCR and western blot were performed to analyze Dicer 1 expression. In addition, a functional restoration assay was performed to investigate whether miR-18a promotes HCC cell migration and invasion by directly targeting Dicer 1.Results miR-18a inhibitor can suppress the migration and invasion of HCC cells. Furthermore, suppression of Dicer l expression by small interfering RNA essentially abolished the inhibition of cell migration and invasion induced by miR-18a inhibitor, restorating these activities to levels similar to the parental HCC cells. Interestingly, suppression of miR-18a in HCC cells resulted in enhanced expression of Dicer l. In addition, the results of a luciferase assay demonstrated targeted regulation of Dicer l by miR-18a.Conclusion Our findings suggest that miR-18a promotes migration and invasion of HCC cells by inhibiting Dicer l expression.

Aspect-based sentiment analysis (ABSA) aims to predict fine-grained sentiments of comments with respect to given aspect terms or categories. In previous ABSA methods, the importance of aspect has been realized and verified. Most existing LSTM-based models take aspect into account via the attention mechanism, where the attention weights are calculated after the context is modeled in the form of contextual vectors. However, aspect-related information may be already discarded and aspect-irrelevant information may be retained in classic LSTM cells in the context modeling process, which can be improved to generate more effective context representations. This paper proposes a novel variant of LSTM, termed as aspect-aware LSTM (AA-LSTM), which incorporates aspect information into LSTM cells in the context modeling stage before the attention mechanism. Therefore, our AA-LSTM can dynamically produce aspect-aware contextual representations. We experiment with several representative LSTM-based models by replacing the classic LSTM cells with the AA-LSTM cells. Experimental results on SemEval-2014 Datasets demonstrate the effectiveness of AA-LSTM., Comment: accepted by IJCAI 2019

The influence of surface topography on adhesion hysteresis of thin film was studied. The Cu/Ti bilayer thin film was fabricated using dc magnetron sputtering technology and its surface adhesion force was measured using force-displacement curves of AFM. The experiment results show that the unload curve exhibits obvious adhesion hysteresis. The theoretical analysis on the influence of thin film roughness to the adhesion hysteresis was carried out using JKR model. The results show that the load curve and unload curve constitute a closed curve and that the smoother the surface, the more obvious adhesion hysteresis (that is, the rougher the surface, the less obvious adhesion hysteresis).

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 cross&compress 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 cross&compress 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 a decent performance even if user-item interactions are sparse., In Proceedings of The 2019 Web Conference (WWW 2019)

Human conversations are often embedded with emotions. To simulate human conversations, the response generated by a chatbot not only has to be topically relevant to the post, but should also carry an appropriate emotion. In this paper, we conduct analysis based on social media data to investigate how emotions influence conversation generation. Based on observation, we propose methods to determine the appropriate emotions to be included in a response and to generate responses with the emotions. The encoder-decoder architecture is extended to incorporate emotions. We propose two implementations which train the two steps separately or jointly. An empirical study on a public dataset from STC at NTCIR-12 shows that our models outperform both a retrieval-based method and a generation model without emotion, indicating the importance of emotions in short text conversation generation and the effectiveness of our approach.

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.

// Jian-feng Huang 1,2,* , Fu-zheng Zhang 2,* , Qin-zhou Zou 2 , Le-yuan Zhou 2 , Bo Yang 2 , Jian-jun Chu 2 , Jia-hua Yu 1 , Hao-wen Zhang 1 , Xiao-peng Yuan 3 , Guo-mei Tai 3 , Fen-ju Liu 1 and C-M Charlie Ma 4 1 Department of Radiobiology, School of Radiation Medicine and Protection and Jiangsu Provincial Key Laboratory of Radiation Medicine and Protection, Medical College of Soochow University, Collaborative Innovation Center of Radiation Medicine of Jiangsu Higher Education Institutions and School for Radiological and Interdisciplinary Sciences (RAD-X), Soochow University, Suzhou, China 2 Department of Radiation Oncology, Affiliated Hospital of Jiangnan University, Wuxi, China 3 Department of Radiation Oncology, Nantong Tumor Hospital, Affiliated Tumor Hospital of Nantong University, Nantong, China 4 Department of Radiation Oncology, Fox Chase Cancer Center, Philadelphia, PA, USA * These authors have contributed equally to this work Correspondence to: Fen-ju Liu, email: // Keywords : nasopharyngeal carcinoma, nimotuzumab, chemoradiotherapy, induction chemotherapy Received : October 13, 2016 Accepted : November 24, 2016 Published : December 10, 2016 Abstract Overexpression of epidermal growth factor receptor can be found in more than 80% of patients with locoregionally advanced nasopharyngeal carcinoma and is associated with shorter survival. In this work, we evaluated the feasibility of adding nimotuzumab to chemoradiation in locoregionally advanced nasopharyngeal carcinoma. Twenty-three patients with clinically staged T3-4 or any node-positive disease were enrolled. They were scheduled to receive one cycle of induction chemotherapy followed by intensity-modulated radiotherapy, weekly administration of nimotuzumab and concurrent chemotherapy. Results showed that all patients received a full course of radiotherapy, 19(82.6%)patients completed the scheduled neoadjuvant and concurrent chemotherapy, and 22(95.7%) patients received ≥6 weeks of nimotuzumab. During the period of concurrent chemoradiation and nimotuzumab, grade 3-4 toxicities occurred in 14(60.9%) patients: 8 (34.8%) had grade 3-4 oral mucositis, 6(26.1%) had grade 3 neutropenia, and 1(4.3%) had grade 3 dermatitis. No acne-like rash was observed. With a median follow-up of 24.1 months, the 2-year progression-free survival and overall survival were 83.5% and 95.0%, respectively. In conclusion, concurrent administration of chemoradiation and nimotuzumab was well-tolerated with good compliance. Preliminary clinical outcome data appear encouraging with favorable normal tissue toxicity results comparing with historical data of concurrent chemoradiation plus cetuximab.

More and more geo-tagged social media data is generated, nowadays, from the geo-tagged tweets, geo-tagged photos to check-ins. Analyzing this flourish data enables the possibility for us to discover users daily mobility patterns, profiles and preferences. As a result, based on the analyzed results, new types of location-based services emerge. In this article, we first introduce the recent advances in location-based user preferences modeling, which includes: 1) inferring users demographics, 2) identifying users novelty-seeking characteristics and 3) discovering users shopping impulsiveness. After that, we present a comprehensive summary on the state-of-arts of the location-based services, which take advantage of the geo-social media, including: 1) location-based recommendations, 2) location-based predication.

An optimization method was proposed to select the peak wavelengths of LED chips for correlated color temperature (CCT) tunable three-chip LED light sources based on the Commission Internationale de l’Eclairage (CIE) 2017 color fidelity index (Rf). The results indicate that the optimized peak wavelengths of three LED chips, i.e., 460 nm, 536 nm, and 609 nm, can achieve high color fidelity across the CCT range from 2700 K to 6500 K. Accordingly, three practical LED chips were chosen to realize the three-chip LED light source. The three-chip LED source can achieve the CIE Rf scores of around 80 in a wide range of CCTs, validating the effectiveness of the optimized peak wavelengths. The proposed optimization method can apply to other white-light LEDs for higher color fidelity.

In this paper, a novel micro-displacement sensor with double optical fiber probes is proposed and designed, which can realize the highly sensitive sensing of longitudinal or lateral micro-displacements. The optical fiber probes are made through photopolymer formulation, and the effects of reaction time and optical power on the growth length of the probe are illustrated. The relationship between light intensity and longitudinal micro-displacement is a power function in the range of 0–100 μm at room temperature with a correlation coefficient of 98.92%. For lateral micro-displacement, the sensitivity is −2.9697 dBm/μm in the range of 0–6 μm with a linear fit of 99.61%. In addition, the linear correlation coefficient decreases as the initial longitudinal distance increases, and the function of these correlation coefficients is also linear with a linearity of 96.14%. This sensor has a simple manufacturing process, low cost, high sensitivity, and fast response speed. It is suitable for harsh environments such as strong electromagnetic interference and corrosivity, and has a broad application prospect in the field of micro-displacement sensing.

Various improvement methods have been proposed to cope with the deficiencies of the Commission Internationale de l’Eclairage (CIE) color rendering index (CRI). However, standardizing these methods lack quantitative evidences. Accordingly, with numerous light sources having rich spectral properties, the changes in CRI scores before and after improvement were quantitatively evaluated in terms of six critical components, including reference illuminant, standard observer, chromatic adaptation transform, uniform color space, color difference combination, and the rescaling operation, involved in the CIE CRI calculation. The results indicate that the largest change in CRI scores occurs in improving uniform color space, especially with CAM02-UCS, and that the CRI scores of the sources with highly-structured spectra tend to be changed considerably by most improvement methods. Compared with the other CRIs, the change in the special CRI R9 is frequently much larger for almost all the improvement methods. The least change in CRI scores is caused by partly updating reference illuminants as the mixed illuminants of the Planckian radiation and CIE daylight, which should be the most acceptable improvement. Additionally, adopting a rescaling function with the original scaling factor could be a candidate for avoiding negative CRI scores without leading to significantly large change in positive scores.

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 Ripple Network, an end-to-end framework that naturally incorporates the knowledge graph into recommender systems. Similar to actual ripples propagating on the surface of water, Ripple Network 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 Ripple Network achieves substantial gains in a variety of scenarios, including movie, book and news recommendation, over several state-of-the-art baselines., Comment: CIKM 2018

At present, chemotherapy and radiotherapy represent the primary modalities of treatment for patients with unresectable esophageal squamous cell carcinoma (ESCC). However, the outcome of patients remains poor owing to radioresistance. The present study aimed to determine the radiosensitizing effect of liriodenine, an aporphine alkaloid derived from Enicosanthellum pulchrum, and investigating the underlying mechanisms in ESCC, using the esophageal cancer ECA-109 cell line. Cellular proliferation was evaluated using the Cell Counting kit-8 assay. Colony formation assay was performed to characterize the radiosensitive effects of liriodenine on ECA-109 cells, and flow cytometry was used to detect the percentage of cells undergoing apoptosis. An immunofluorescence assay was utilized to evaluate the DNA damage repair ability. Western blotting was used to assess the protein levels of caspase-3, B cell lymphoma-2 (Bcl-2) and Bcl-2-associated X (Bax). Liriodenine dose-dependently inhibited ECA-109 cell viability. The clonogenic survival assay demonstrated that liriodenine increased the radiosensitivity of ESCC cells, with a sensitization enhancement ratio of 1.11-1.69. The results of flow cytometry demonstrated that liriodenine induced apoptosis and G2/M arrest. The immunofluorescence assay revealed that liriodenine delays DNA damage repair. The upregulation of Bax and Caspase-3, and the suppression of Bcl-2 confirmed that apoptosis was occurring. Liriodenine radiosensitizes ECA-109 cells by inducing apoptosis and G2/M arrest. The findings of the present study indicated that liriodenine may represent an anticancer agent with promising potential for the treatment of ESCC.

Transcribing content from structural images, e.g., writing notes from music scores, is a challenging task as not only the content objects should be recognized, but the internal structure should also be preserved. Existing image recognition methods mainly work on images with simple content (e.g., text lines with characters), but are not capable to identify ones with more complex content (e.g., structured symbols), which often follow a fine-grained grammar. To this end, in this paper, we propose a hierarchical Spotlight Transcribing Network (STN) framework followed by a two-stage "where-to-what" solution. Specifically, we first decide "where-to-look" through a novel spotlight mechanism to focus on different areas of the original image following its structure. Then, we decide "what-to-write" by developing a GRU based network with the spotlight areas for transcribing the content accordingly. Moreover, we propose two implementations on the basis of STN, i.e., STNM and STNR, where the spotlight movement follows the Markov property and Recurrent modeling, respectively. We also design a reinforcement method to refine the framework by self-improving the spotlight mechanism. We conduct extensive experiments on many structural image datasets, where the results clearly demonstrate the effectiveness of STN framework., Accepted by KDD2018 Research Track. In proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD'18)

Jianfeng Huang,1,* Xiaopeng Yuan,2,* Qingfeng Pang,3 Haowen Zhang,4 Jiahua Yu,4 Bo Yang,1 Leyuan Zhou,1 Fuzheng Zhang,1 Fenju Liu4 1Department of Radiation Oncology, Affiliated Hospital of Jiangnan University, Wuxi, People’s Republic of China; 2Department of Radiation Oncology, Nantong Tumor Hospital, Affiliated Tumor Hospital of Nantong University, Nantong, People’s Republic of China; 3Department of Physiopathology, Wuxi School of Medicine, Jiangnan University, Wuxi, People’s Republic of China; 4State Key Laboratory of Radiation Medicine and Protection, School of Radiation Medicine and Protection, Medical College of Soochow University, Collaborative Innovation Center of Radiological Medicine of Jiangsu Higher Education Institutions, Jiangsu Provincial Key Laboratory of Radiation Medicine and Protection, Suzhou, People’s Republic of China *These authors contributed equally to this work Introduction: In this study, the radiation-enhancing effects of combined treatment with nimotuzumab, a humanized EGFR-blocking antibody, and celecoxib, a COX-2 selective inhibitor, in human nasopharyngeal carcinoma (NPC) cells were investigated. Materials and methods: 3-(4,5-Dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide and clonogenic survival assays were done to evaluate the combined cytotoxic and radiosensitizing effects of nimotuzumab or celecoxib or the combination on CNE1 and CNE2 cells. Western blot analysis was performed to identify the effect of nimotuzumab and/or celecoxib with or without irradiation on the cytoplasmic and nuclear EGFR signaling pathways in CNE2 cells. Results: Our results demonstrated that concurrent administration of nimotuzumab and celecoxib cooperatively enhanced the cytotoxicity and radiosensitivity of CNE2 cells but not CNE1 cells. The combination of both drugs with or without irradiation also cooperatively inhibited cytoplasmic and nuclear EGFR signaling pathways in CNE2 cells. Conclusion: Our results suggest a promising approach for the treatment of poorly differentiated NPC. Keywords: celecoxib, cyclooxygenase-2, epidermal growth factor receptor, nasopharyngeal carcinoma, nimotuzumab, radiosensitivity

With a large amount of user activity data accumulated, it is crucial to exploit user sequential behavior for sequential recommendations. Conventionally, user general taste and recent demand are combined to promote recommendation performances. However, existing methods often neglect that user long-term preference keep evolving over time, and building a static representation for user general taste may not adequately reflect the dynamic characters. Moreover, they integrate user-item or item-item interactions through a linear way which limits the capability of model. To this end, in this paper, we propose a novel two-layer hierarchical attention network, which takes the above properties into account, to recommend the next item user might be interested. Specifically, the first attention layer learns user long-term preferences based on the historical purchased item representation, while the second one outputs final user representation through coupling user long-term and short-term preferences. The experimental study demonstrates the superiority of our method compared with other state-of-the-art ones.

Online news recommender systems aim to address the information explosion of news and make personalized recommendation for users. In general, news language is highly condensed, full of knowledge entities and common sense. However, existing methods are unaware of such external knowledge and cannot fully discover latent knowledge-level connections among news. The recommended results for a user are consequently limited to simple patterns and cannot be extended reasonably. Moreover, news recommendation also faces the challenges of high time-sensitivity of news and dynamic diversity of users' interests. To solve the above problems, in this paper, we propose a deep knowledge-aware network (DKN) that incorporates knowledge graph representation into news recommendation. DKN is a content-based deep recommendation framework for click-through rate prediction. The key component of DKN is a multi-channel and word-entity-aligned knowledge-aware convolutional neural network (KCNN) that fuses semantic-level and knowledge-level representations of news. KCNN treats words and entities as multiple channels, and explicitly keeps their alignment relationship during convolution. In addition, to address users' diverse interests, we also design an attention module in DKN to dynamically aggregate a user's history with respect to current candidate news. Through extensive experiments on a real online news platform, we demonstrate that DKN achieves substantial gains over state-of-the-art deep recommendation models. We also validate the efficacy of the usage of knowledge in DKN., Comment: The 27th International Conference on World Wide Web (WWW'18)

In this paper, we give the overview of the social media user modeling shared task in the NLPCC-ICCPOL 2017. We first review the background of social media user modeling, and then describe two social media user modeling tasks in this year’s NLPCC-ICCPOL, including the construction of the benchmark datasets and the evaluation metrics. The evaluation results of submissions from participating teams are presented in the experimental part.

In this paper, we propose a novel Deep Reinforcement Learning framework for news recommendation. Online personalized news recommendation is a highly challenging problem due to the dynamic nature of news features and user preferences. Although some online recommendation models have been proposed to address the dynamic nature of news recommendation, these methods have three major issues. First, they only try to model current reward (e.g., Click Through Rate). Second, very few studies consider to use user feedback other than click / no click labels (e.g., how frequent user returns) to help improve recommendation. Third, these methods tend to keep recommending similar news to users, which may cause users to get bored. Therefore, to address the aforementioned challenges, we propose a Deep Q-Learning based recommendation framework, which can model future reward explicitly. We further consider user return pattern as a supplement to click / no click label in order to capture more user feedback information. In addition, an effective exploration strategy is incorporated to find new attractive news for users. Extensive experiments are conducted on the offline dataset and online production environment of a commercial news recommendation application and have shown the superior performance of our methods.

Combinatorial features are essential for the success of many commercial models. Manually crafting these features usually comes with high cost due to the variety, volume and velocity of raw data in web-scale systems. Factorization based models, which measure interactions in terms of vector product, can learn patterns of combinatorial features automatically and generalize to unseen features as well. With the great success of deep neural networks (DNNs) in various fields, recently researchers have proposed several DNN-based factorization model to learn both low- and high-order feature interactions. Despite the powerful ability of learning an arbitrary function from data, plain DNNs generate feature interactions implicitly and at the bit-wise level. In this paper, we propose a novel Compressed Interaction Network (CIN), which aims to generate feature interactions in an explicit fashion and at the vector-wise level. We show that the CIN share some functionalities with convolutional neural networks (CNNs) and recurrent neural networks (RNNs). We further combine a CIN and a classical DNN into one unified model, and named this new model eXtreme Deep Factorization Machine (xDeepFM). On one hand, the xDeepFM is able to learn certain bounded-degree feature interactions explicitly; on the other hand, it can learn arbitrary low- and high-order feature interactions implicitly. We conduct comprehensive experiments on three real-world datasets. Our results demonstrate that xDeepFM outperforms state-of-the-art models. We have released the source code of xDeepFM at \url{https://github.com/Leavingseason/xDeepFM}., Comment: 10 pages

Urban transportation is an important factor in energy consumption and pollution, and is of increasing concern due to its complexity and economic significance. Its importance will only increase as urbanization continues around the world. In this article, we explore drivers’ refueling behavior in urban areas. Compared to questionnaire-based methods of the past, we propose a complete data-driven system that pushes towards real-time sensing of individual refueling behavior and citywide petrol consumption. Our system provides the following: detection of individual refueling events (REs) from which refueling preference can be analyzed; estimates of gas station wait times from which recommendations can be made; an indication of overall fuel demand from which macroscale economic decisions can be made, and a spatial, temporal, and economic view of urban refueling characteristics. For individual behavior, we use reported trajectories from a fleet of GPS-equipped taxicabs to detect gas station visits. For time spent estimates, to solve the sparsity issue along time and stations, we propose context-aware tensor factorization (CATF), a factorization model that considers a variety of contextual factors (e.g., price, brand, and weather condition) that affect consumers’ refueling decision. For fuel demand estimates, we apply a queue model to calculate the overall visits based on the time spent inside the station. We evaluated our system on large-scale and real-world datasets, which contain 4-month trajectories of 32,476 taxicabs, 689 gas stations, and the self-reported refueling details of 8,326 online users. The results show that our system can determine REs with an accuracy of more than 90%, estimate time spent with less than 2 minutes of error, and measure overall visits in the same order of magnitude with the records in the field study.

With the growing popularity of location-based social networks, numerous location visiting records (e.g., check-ins) continue to accumulate over time. The more these records are collected, the better we can understand users’ mobility patterns and the more accurately we can predict their future locations. However, due to the personality trait of neophilia, people also show propensities of novelty seeking in human mobility, that is, exploring unvisited but tailored locations for them to visit. As such, the existing prediction algorithms, mainly relying on regular mobility patterns, face severe challenges because such behavior is beyond the reach of regularity. As a matter of fact, the prediction of this behavior not only relies on the forecast of novelty-seeking tendency but also depends on how to determine unvisited candidate locations. To this end, we put forward a Collaborative Exploration and Periodically Returning model (CEPR), based on a novel problem, Exploration Prediction (EP), which forecasts whether people will seek unvisited locations to visit, in the following. When people are predicted to do exploration, a state-of-the-art recommendation algorithm, armed with collaborative social knowledge and assisted by geographical influence, will be applied for seeking the suitable candidates; otherwise, a traditional prediction algorithm, incorporating both regularity and the Markov model, will be put into use for figuring out the most possible locations to visit. We then perform case studies on check-ins and evaluate them on two large-scale check-in datasets with 6M and 36M records, respectively. The evaluation results show that EP achieves a roughly 20% classification error rate on both datasets, greatly outperforming the baselines, and that CEPR improves performances by as much as 30% compared to the traditional location prediction algorithms.

In online social networks people often express attitudes towards others, which forms massive sentiment links among users. Predicting the sign of sentiment links is a fundamental task in many areas such as personal advertising and public opinion analysis. Previous works mainly focus on textual sentiment classification, however, text information can only disclose the "tip of the iceberg" about users' true opinions, of which the most are unobserved but implied by other sources of information such as social relation and users' profile. To address this problem, in this paper we investigate how to predict possibly existing sentiment links in the presence of heterogeneous information. First, due to the lack of explicit sentiment links in mainstream social networks, we establish a labeled heterogeneous sentiment dataset which consists of users' sentiment relation, social relation and profile knowledge by entity-level sentiment extraction method. Then we propose a novel and flexible end-to-end Signed Heterogeneous Information Network Embedding (SHINE) framework to extract users' latent representations from heterogeneous networks and predict the sign of unobserved sentiment links. SHINE utilizes multiple deep autoencoders to map each user into a low-dimension feature space while preserving the network structure. We demonstrate the superiority of SHINE over state-of-the-art baselines on link prediction and node recommendation in two real-world datasets. The experimental results also prove the efficacy of SHINE in cold start scenario., The 11th ACM International Conference on Web Search and Data Mining (WSDM 2018)

Considering that the existing color quality (CQ) metrics for light sources cannot correlate well with the subjective evaluation, in an immersive environment equipped with a multichannel LED light source, a psychophysical experiment by categorical judgment method was carried out to assess the three perception-related CQ attributes of light sources in terms of naturalness, colorfulness, and preference. The experiment collected the subjective responses to these attributes of up to 41 metameric spectra at each of four test correlated color temperatures (CCTs) ranging from 2800 to 6500 K, which covers the usual white-light range for general lighting. The results indicate that preference exhibits relatively high correlation with naturalness and colorfulness, and naturalness is weakly related to colorfulness. Besides, 20 typical CQ metrics were adopted to examine their validity in characterizing the subjective data, confirming their limited performance. Meanwhile, the underlying relationship of these metrics and the subjective data was also analyzed by the multidimensional scaling, revealing that almost all metrics can correspond to one attribute of naturalness, colorfulness, and preference, and that the saturation level is identified as a critical factor affecting these attributes. Based on these results, a unified CQ model was developed with a multiple nonlinear regression equation combining the Illuminating Engineering Society of North America color rendition method. The model accords satisfactorily with the subjective evaluation, while being applicable to a wide range of CCTs.

The 2017 ACM RecSys Challenge focuses on the problem of job recommendations on XING in a cold-start scenario. In this paper we describe our solution as well as some practical lessons learned from the competition. We model this task as a binary classification problem. Negative candidate selection is the first key phase in our solution. We design a negative sampling strategy which performs significantly better than taking users' deleted or unclicked items as negative candidates. We then extract comprehensive features to model the relationship between a user-job candidate, including the direct profile similarity between the user and the job, and the profile similarity between the user's historical interested jobs and the target job. To make the whole pipeline scalable and easy to deploy online, we decide to use a single boosting tree model as the final discriminative model, instead of using a stacking ensemble of multiple models. Overall our model ranked 5th on the challenge leaderboard, and our last model has remained in 2nd place during the last two online weeks. We have open-sourced our implementation on https://github.com/Leavingseason/RecsysChallenge2017.

This note points out additional funding information that was not added to Appl. Opt.56, 1962 (2017).APOPAI1559-128X10.1364/AO.56.001962.

The trade-off between the color fidelity index (Rsubf/sub) released recently by the Illuminating Engineering Society of North America (IES) and luminous efficacy of radiation (LER) was investigated by adjusting the peak wavelengths, spectral widths, and intensities of four-channel LEDs utilizing a multiobjective optimization algorithm based on differential evolution in the correlated color temperature (CCT) ranging from 2800 to 6500 K for general lighting. The results indicate that Rsubf/subat a specific LER value decreases with the increasing CCT, and vice versa, and that Rsubf/subhas significant improvements over the Commission Internationale de l'Eclairage (CIE) color rendering index (CRI) in avoiding spectral gaming and evaluating the light sources even with negative CIE general CRI (Rsuba/sub). Further, the optimal peak wavelengths with regard to Rsubf/subwere identified as 629 nm, 568 nm, 504 nm, and 447 nm, yielding high color rendering in terms of Rsubf/sub(93∼94) and Rsuba/sub(95∼97) and relatively excellent LER (299 lm/W∼339 lm/W) over a wide range of CCTs from 2800 to 6500 K. This suggests that Rsubf/subcan be compatible with Rsuba/sub, making it possible to obtain a common set of optimal peak wavelengths for Rsubf/suband Rsuba/sub. Besides, the IES method could assess saturated red and skin tones more fairly than the CIE CRI. With a practical 17-channel LED array covering the resulted four optimal peak wavelengths, the improvements of the IES method over the CIE CRI were validated further.

An incisive understanding of user personality is not only essential to many scientific disciplines, but also has a profound business impact on practical applications such as digital marketing, personalized recommendation, mental diagnosis, and human resources management. Previous studies have demonstrated that language usage in social media is effective in personality prediction. However, except for single language features, a less researched direction is how to leverage the heterogeneous information on social media to have a better understanding of user personality. In this paper, we propose a Heterogeneous Information Ensemble framework, called HIE, to predict users' personality traits by integrating heterogeneous information including self-language usage, avatar, emoticon, and responsive patterns. In our framework, to improve the performance of personality prediction, we have designed different strategies extracting semantic representations to fully leverage heterogeneous information on social media. We evaluate our methods with extensive experiments based on a real-world data covering both personality survey results and social media usage from thousands of volunteers. The results reveal that our approaches significantly outperform several widely adopted state-of-the-art baseline methods. To figure out the utility of HIE in a real-world interactive setting, we also present DiPsy, a personalized chatbot to predict user personality through heterogeneous information in digital traces and conversation logs.

Recent studies in psychology suggest that novelty-seeking trait is highly related to consumer behavior, which has a profound business impact on online recommendation. This paper studies the problem of mining novelty seeking trait across domains to improve the recommendation performance in target domain. We propose an efficient model, CDNST, which significantly improves the recommendation performance by transferring the knowledge from auxiliary source domain. We conduct extensive experiments on three domain datasets crawled from Douban (www.douban.com) to demonstrate the effectiveness of the proposed model. Moreover, we find that the property of sequential data affects the performance of CDNST.

Recommendation systems (RS) play an important role in directing customers to their favorite items. Data sparsity, which usually leads to overfitting, is a major bottleneck for making precise recommendations. Several cross-domain RSs have been proposed in the past decade in order to reduce the sparsity issues via transferring knowledge. However, existing works only focus on either nearest neighbor model or latent factor model for cross domain scenario. In this paper, we introduce a Multifaceted Cross-Domain Recommendation System (MCDRS) which incorporates two different types of collaborative filtering for cross domain RSs. The first part is a latent factor model. In order to utilize as much knowledge as possible, we propose a unified factorization framework to combine both CF and content-based filtering for cross domain learning. On the other hand, to overcome the potential inconsistency problem between different domains, we equip the neighbor model with a selective learning mechanism so that domain-independent items gain more weight in the transfer process. We conduct extensive experiments on two real-world datasets. The results demonstrate that our MCDRS model consistently outperforms several state-of-the-art models.

For shopkeepers, one of their biggest common concerns is whether their business will thrive or fail in the future. With the development of new ways to collect business data, it is possible to leverage multiple domains' knowledge to build an intelligent model for business assessment. In this paper, we discuss what the potential indicators are for the long-term survival of a physical store. To this end, we study factors from four pillars: geography, user mobility, user rating, and review text. We start by exploring the impact of geographic features, which describe the location environment of the retailer store. The location and nearby places play an important role in the popularity of the shop, and usually less competitiveness and more heterogeneity is better. Then we study user mobility. It can be viewed as supplementary to the geographical placement, showing how the location can attract users from anywhere. Another important factor is how the shop can serve and satisfy users. We find that restaurant survival prediction is a hard task that can not be solved simply using consumers' ratings or sentiment metrics. Compared with conclusive and well-formatted ratings, the various review words provide more insight of the shop and deserve in-depth mining. We adopt several language models to fully explore the textual message. Comprehensive experiments demonstrate that review text indeed have the strongest predictive power. We further compare different cities' models and find the conclusions are highly consistent. Although we focus on the class of restaurant in this paper, the method can be easily extended to other shop categories.

To overcome data sparsity problem, we propose a cross domain recommendation system named CCCFNet which can combine collaborative filtering and content-based filtering in a unified framework. We first introduce a factorization framework to tie CF and content-based filtering together. Then we find that the MAP estimation of this framework can be embedded into a multi-view neural network. Through this neural network embedding the framework can be further extended by advanced deep learning techniques.

Smart card transactions capture rich information of human mobility and urban dynamics and therefore are of particular interest to urban planners and location-based service providers. However, since most transaction systems are only designated for billing purpose, typically, fine-grained location information, such as the exact boarding and alighting stops of a bus trip, is only partially or not available at all, which blocks deep exploitation of this rich and valuable data at individual level. This paper presents a collaborative space alignment framework to reconstruct individual mobility history from a metropolitan-scale smart card transaction dataset. Specifically, we show that by delicately aligning the monetary space and geospatial space with the temporal space, we are able to extrapolate a series of critical domain-specific constraints. Later, these constraints are naturally incorporated into a semi-supervised conditional random field (CRF) to infer the exact boarding and alighting stops of all transit routes, where the features of the CRF model consist of not only pre-defined indicator features extracted from individual trips but also latent features crafted from different users' trips using collaborative filtering. Here, we consider two types of collaborative features: (1) the similarity in terms of users' choices of bus lines and (2) latent temporal patterns of users' commuting behaviors. Extensive experimental results show that our approach achieves a high accuracy, e.g., given only 10 % trips with known alighting/boarding stops, and we successfully inferred more than 79 % alighting and boarding stops from all unlabeled trips. In particular, we validated that the extracted collaborative features significantly contribute to the accuracy of our model. In addition, we have demonstrated that by applying our approach to enrich the data, the performance of a conventional method for identifying users' home and work places can be dramatically improved (with 83 % improvement on home detection and 38 % improvement on work place detection). The proposed method offers the possibility to mine individual mobility from common public transit transactions, and showcases how uncertain data can be leveraged with domain knowledge and constraints, to support cross-application data-mining tasks.

Wireless sensor nodes (WSNs) for temperature and humidity monitoring are commonly used in a cold chain logistics container. Energy harvesting technology is expected to realize the sustainable self-power supply for the WSN. Low amplitude and broadband vibration energy harvesting performance are the key points in train application. In this study, two piezoelectric vibration energy harvesters (PVEHs) are designed and simulated via COMSOL. Their low resonant frequencies and high electromechanical sensitivities are realized by big L-shaped mass blocks with different material densities. Their broadband vibration energy harvesting performance is achieved by the stopper and series connection. Experimental data are shown at an acceleration of 0.5 m/s2; PVEH-1 and PVEH-2 have maximum powers of 0.24 mW and 0.1 mW when excited at the resonant frequencies of 13.1 Hz and 18.8 Hz, respectively, and they both have the optimal load resistance of 40 kΩ. Two circuit design styles of two PVEHs, independent and series styles, are proposed for broadband vibration energy harvesting. Experimental results show that the series style has a wider operating frequency bandwidth and shorter charging time. Two PVEHs in series style can be effectively used for power supply of the temperature and humidity WSN in the broadband frequency range of 8.7-22.0 Hz above charging root mean square voltage of 5 V at the acceleration of 3.0 m/s2. This scheme is promised to be applied to the cold chain logistics train container.

A high-accuracy method for calculating the correlated color temperature (CCT) of a light source was proposed through constructing a small-scale lookup table (LUT) based on the golden section search algorithm. The LUT, which only needs to be constructed once, is employed to narrow the search range of CCT. And it is easily extended to apply to a desired CCT range. The golden section search algorithm suitable for finding the extremum of a unimodal function is adopted to determine the CCT value of a test light source in a reduced range. The proposed method does not need to store a large number of precomputed data without utilizing any fitting operation. The method was tested in a wide range of chromaticity with CCT from 1000 K to 20,000 K and Duv from −0.03 to 0.03. The results show that relative to the combined solution, the proposed method can achieve significantly high accuracy, yielding an average error of 0.008 K and the maximum error less than 0.32 K. Furthermore, the calculation accuracy of the method is independent on the Duv value of a test light source.

Fiber Bragg gratings (FBGs) have become a promising sensor element for measuring temperature, strain and other parameters. However, there exists a cross-sensitivity between temperature and strain. An all-fiber sensing system composed of gold-plated FBGs and acrylic FBGs has been proposed and fabricated, which can achieve simultaneous measurement of axial strain and temperature. Using a spectrometer to monitor the shift in the transmission spectrum in real time, the strain and ambient temperature applied to the sensor can be measured conveniently. The experimental results show that the temperature sensitivity of a gold-plated FBG is 3.2 times of that of an acrylate-plated FBG, which can distinguish the temperature changes efficiently. In other words, its sensitivity is increased 2.2 times compared with ordinary fiber grating. The strain sensitivity of gold-plated FBG and acrylate-plated FBG is the same, which is 1.19 pm/µe. The sensor has the advantages of having a simple structure and high sensitivity and can be used to monitor the running state of rail traffic.

Although weighted tensor factorization tailored to implicit feedback has shown its superior performance in temporal-aware location recommendation, it suffers from three critical challenges. First, it doesn't distinguish the confidence of negative preference for time-dependent unvisited locations from that for fully unvisited ones. Second, discontinuity arises from time discretization, and thus an infinitely large margin may exist between different bins of time. Third, geographical constraints of neighbor locations are not taken into account. To address these challenges, we propose a regularized content-aware tensor factorization (RCTF) algorithm, which exploits three strategies to address the corresponding challenges. First, it introduces a novel interaction regularization, second, it represents each bin of time by a derived feature vector from eigen decomposition of a time-bin similarity matrix, to capture the proximity of neighbor bins of time, third, it encodes geographical information of locations by discrete spatial distributions, so that spatial proximity constraints can be satisfied by simply feeding them into location content. The proposed algorithm is then evaluated for time-aware location recommendation on two large scale location-based social network datasets. The experimental results show the superiority of the proposed algorithm to several competing time-aware recommendation baselines, and verify the significant benefit of three strategies in the proposed algorithm.

The methods were proposed to design the spectra of light sources with multi-channel LEDs based on a differential evolution (DE) algorithm. Specifically, to produce a desired illuminance level while matching a specific spectrum, a matching method combining the DE algorithm and photometry was verified by matching the Commission Internationale de l’Eclairage (CIE) standard illuminants A and D65 at the two desired illuminance levels (500 and 1000 lx) with a multi-channel LED array. Furthermore, considering that correlated color temperature (CCT), illuminance, and color rendering are of practical importance to general lighting, on the basis of photometry and colorimetry, the DE algorithm integrating a mechanism dealing with constraints was developed to optimize the CIE general color rendering index of light sources under a joint target of CCT and illuminance. Sixteen selected targets (eight CCTs from 2800 to 6500 K at 500 and 1000 lx, respectively) were realized with the LED array. The results indicate that the relative errors of CCT and illuminance are less than 0.7% and 1.5%, respectively, while achieving excellent or even perfect color rendering. These methods also enable a practical tool for smart and dynamic lighting.

At present, chemotherapy and radiotherapy represent the primary modalities of treatment for patients with unresectable esophageal squamous cell carcinoma (ESCC). However, the outcome of patients remains poor owing to radioresistance. The present study aimed to determine the radiosensitizing effect of liriodenine, an aporphine alkaloid derived from Enicosanthellum pulchrum, and investigating the underlying mechanisms in ESCC, using the esophageal cancer ECA-109 cell line. Cellular proliferation was evaluated using the Cell Counting kit-8 assay. Colony formation assay was performed to characterize the radiosensitive effects of liriodenine on ECA-109 cells, and flow cytometry was used to detect the percentage of cells undergoing apoptosis. An immunofluorescence assay was utilized to evaluate the DNA damage repair ability. Western blotting was used to assess the protein levels of caspase-3, B cell lymphoma-2 (Bcl-2) and Bcl-2-associated X (Bax). Liriodenine dose-dependently inhibited ECA-109 cell viability. The clonogenic survival assay demonstrated that liriodenine increased the radiosensitivity of ESCC cells, with a sensitization enhancement ratio of 1.11–1.69. The results of flow cytometry demonstrated that liriodenine induced apoptosis and G2/M arrest. The immunofluorescence assay revealed that liriodenine delays DNA damage repair. The upregulation of Bax and Caspase-3, and the suppression of Bcl-2 confirmed that apoptosis was occurring. Liriodenine radiosensitizes ECA-109 cells by inducing apoptosis and G2/M arrest. The findings of the present study indicated that liriodenine may represent an anticancer agent with promising potential for the treatment of ESCC.

The wide adoption of location-based services provide the potential to understand people's mobility pattern at an unprecedented level, which can also enable food-service industry to accurately predict consumers' dining behavior. In this paper, based on users' dining implicit feedbacks (restaurant visit via check-ins), explicit feedbacks (restaurant reviews) as well as some meta data (e.g., location, user demographics, restaurant attributes), we aim at recommending each user a list of restaurants for his next dining. Implicit and Explicit feedbacks of dining behavior exhibit different characteristics of user preference. Therefore, in our work, user's dining preference mainly contains two parts: implicit preference coming from check-in data (implicit feedbacks) and explicit preference coming from rating and review data (explicit feedbacks). For implicit preference, we first apply a probabilistic tensor factorization model (PTF) to capture preference in a latent subspace. Then, in order to incorporate contextual signals from meta data, we extend PTF by proposing an Implicit Preference Model (IPM), which can simultaneously capture users'/restaurants'/time' preference in the collaborative filtering and dining preference in a specific context (e.g., spatial distance preference, environmental preference). For explicit preference, we propose Explicit Preference Model (EPM) by combining matrix factorization with topic modeling to discover the user preference embedded both in rating score and text content. Finally, we design a unified model termed as Collective Implicit Explicit Preference Model (CIEPM) to combine implicit and explicit preference together for restaurant recommendation. To evaluate the performance of our system, we conduct extensive experiments with large-scale datasets covering hundreds of thousands of users and restaurants. The results reveal that our system is effective for restaurant recommendation.

Friendships are dynamic. Previous studies have converged to suggest that social interactions, in both online and offline social networks, are diagnostic reflections of friendship relations (also called social ties). However, most existing approaches consider a social tie as either a binary relation, or a fixed value (named tie strength). In this paper, we investigate the dynamics of dyadic friend relationships through online social interactions, in terms of a variety of aspects, such as reciprocity, temporality, and contextuality. In turn, we propose a model to predict repliers and retweeters given a particular tweet posted at a certain time in a microblog-based social network. More specifically, we have devised a learning-to-rank approach to train a ranker that considers elaborate user-level and tweet-level features (like sentiment, self-disclosure, and responsiveness) to address these dynamics. In the prediction phase, a tweet posted by a user is deemed a query and the predicted repliers/retweeters are retrieved using the learned ranker. We have collected a large dataset containing 73.3 million dyadic relationships with their interactions (replies and retweets). Extensive experimental results based on this dataset show that by incorporating the dynamics of friendship relations, our approach significantly outperforms state-of-the-art models in terms of multiple evaluation metrics, such as MAP, NDCG and Topmost Accuracy. In particular, the advantage of our model is even more promising in predicting the exact sequence of repliers/retweeters considering their orders. Furthermore, the proposed approach provides emerging implications for many high-value applications in online social networks.

Color perception of an object is tremendously influenced by the spectral power distribution (SPD) of a light source. Therefore, it is of great significance to seek for an optimal SPD to enhance the color contrast of a biological specimen in medical lighting. Meanwhile, considering that color rendering property and illuminance are of importance to lighting, with a multichannel LED light source and a differential evolution algorithm dealing with the constraints of color rendering and illuminance, a contrast enhancement method based on maximizing CIEDE2000 color difference (ΔE00*) was proposed in this study. An illuminance level of 500 lx was first verified as the desired illuminance. Afterward, the color contrasts of six groups of color patches were optimized under four different constraints of the Commission Internationale de l’Eclairage (CIE) general color rendering index (Ra), i.e., Ra ≥ 0, Ra ≥ 60, Ra ≥ 80, and Ra ≥ 90, manifesting their significant color contrast enhancement under the optimized light sources compared with those under the simulated CIE A, D50, D65, and D75 light sources. Further, two biological samples, pork tissues and porcine heart, were optimized under the four constraints of Ra. The results show that even in the case of Ra ≥ 90, the optimized ΔE00* values for the two specimens are, respectively, 4.13 and 3.20 higher than those under the D75 light source. More importantly, the developed method also provides a flexible lighting solution to color contrast enhancement.

Consumer impulsivity is a psychological feature characterizing the impulsive buying tendency. In this paper, by bridging consumer behavior with perceived stimuli on social networks, we present a computational framework, termed Consumer Impulsivity Model (CIM), for exploring a consumer's impulsivity in both offline and online context: consumption-related location visit indicating consumption patterns in the physical realm, and online shopping behavior indicating economic activities on the Internet. To demonstrate the effectiveness of CIM, we conduct extensive experiments, with a large dataset we have collected from thousands of consumers. The results show that 1) for 103 subjects, the inferred consumer impulsivity has a positive Pearson correlation with survey results in the situation of product and product category, respectively. 2) females inferred impulsivity is higher than males on average in the situation of product and product category, respectively. Age has a negative Pearson correlation with inferred impulsivity in the situation of POI, POI category and product category, respectively. 3) for next behavior prediction, our model defeats several presented baselines. These results suggest that our framework CIM offers a powerful paradigm for 1) presenting an effective measurement for consumer impulsivity. 2) uncovering the correlation between consumer impulsivity and demographic factors and 3) revealing that the introduction of impulsivity is effective in predicting consumer behavior.