Showing papers by "Alibaba Group published in 2016"

Deep CTR Prediction in Display Advertising

Abstract: Click through rate (CTR) prediction of image ads is the core task of online display advertising systems, and logistic regression (LR) has been frequently applied as the prediction model. However, LR model lacks the ability of extracting complex and intrinsic nonlinear features from handcrafted high-dimensional image features, which limits its effectiveness. To solve this issue, in this paper, we introduce a novel deep neural network (DNN) based model that directly predicts the CTR of an image ad based on raw image pixels and other basic features in one step. The DNN model employs convolution layers to automatically extract representative visual features from images, and nonlinear CTR features are then learned from visual features and other contextual features by using fully-connected layers. Empirical evaluations on a real world dataset with over 50 million records demonstrate the effectiveness and efficiency of this method.

Social Friend Recommendation Based on Multiple Network Correlation

Abstract: Friend recommendation is an important recommender application in social media. Major social websites such as Twitter and Facebook are all capable of recommending friends to individuals. However, most of these websites use simple friend recommendation algorithms such as similarity, popularity, or “friend's friends are friends,” which are intuitive but consider few of the characteristics of the social network. In this paper we investigate the structure of social networks and develop an algorithm for network correlation-based social friend recommendation (NC-based SFR). To accomplish this goal, we correlate different “social role” networks, find their relationships and make friend recommendations. NC-based SFR is characterized by two key components: 1) related networks are aligned by selecting important features from each network, and 2) the network structure should be maximally preserved before and after network alignment. After important feature selection has been made, we recommend friends based on these features. We conduct experiments on the Flickr network, which contains more than ten thousand nodes and over 30 thousand tags covering half a million photos, to show that the proposed algorithm recommends friends more precisely than reference methods.

Tracking slowly moving clairvoyant: optimal dynamic regret of online learning with true and noisy gradient

Abstract: This work focuses on dynamic regret of online convex optimization that compares the performance of online learning to a clairvoyant who knows the sequence of loss functions in advance and hence selects the minimizer of the loss function at each step. By assuming that the clairvoyant moves slowly (i.e., the minimizers change slowly), we present several improved variation-based upper bounds of the dynamic regret under the true and noisy gradient feedback, which are optimal in light of the presented lower bounds. The key to our analysis is to explore a regularity metric that measures the temporal changes in the clairvoyant's minimizers, to which we refer as path variation. Firstly, we present a general lower bound in terms of the path variation, and then show that under full information or gradient feedback we are able to achieve an optimal dynamic regret. Secondly, we present a lower bound with noisy gradient feedback and then show that we can achieve optimal dynamic regrets under a stochastic gradient feedback and two-point bandit feedback. Moreover, for a sequence of smooth loss functions that admit a small variation in the gradients, our dynamic regret under the two-point bandit feedback matches what is achieved with full information.

WFID: Passive Device-free Human Identification Using WiFi Signal

Abstract: We present WFID, a passive device-free indoor human identification system with one pair of WiFi signal transmitter and receiver. WFID design is motivated by the observation that PHY layer Channel State Information (CSI) is capable of capturing the frequency diversity of wideband channel, such that the human body curve may be uniquely identified by learning the feature pattern of CSI. Different from many CSI-based techniques focusing on phase shift, we propose a novel feature of subcarrier-amplitude frequency (SAF). Based on this feature, WFID realizes human identification through a linear-kernel SVM. We have implemented a prototype of WFID with a commercial AP and a computer equipped with one Intel 5300 NIC. WFID is evaluated in two typical indoor scenarios. The results confirm that WFID achieves high classification accuracy which is permanent over several days under two typical indoor scenarios, with low computation cost. This reveals the potential for WFID to realize real-time indoor human identification.

Tracking Slowly Moving Clairvoyant: Optimal Dynamic Regret of Online Learning with True and Noisy Gradient

Abstract: This work focuses on dynamic regret of online convex optimization that compares the performance of online learning to a clairvoyant who knows the sequence of loss functions in advance and hence selects the minimizer of the loss function at each step. By assuming that the clairvoyant moves slowly (i.e., the minimizers change slowly), we present several improved variation-based upper bounds of the dynamic regret under the true and noisy gradient feedback, which are {\it optimal} in light of the presented lower bounds. The key to our analysis is to explore a regularity metric that measures the temporal changes in the clairvoyant's minimizers, to which we refer as {\it path variation}. Firstly, we present a general lower bound in terms of the path variation, and then show that under full information or gradient feedback we are able to achieve an optimal dynamic regret. Secondly, we present a lower bound with noisy gradient feedback and then show that we can achieve optimal dynamic regrets under a stochastic gradient feedback and two-point bandit feedback. Moreover, for a sequence of smooth loss functions that admit a small variation in the gradients, our dynamic regret under the two-point bandit feedback matches what is achieved with full information.

A Domain Robust Approach For Image Dataset Construction

Abstract: There have been increasing research interests in automatically constructing image dataset by collecting images from the Internet. However, existing methods tend to have a weak domain adaptation ability, known as the "dataset bias problem". To address this issue, in this work, we propose a novel image dataset construction framework which can generalize well to unseen target domains. In specific, the given queries are first expanded by searching in the Google Books Ngrams Corpora (GBNC) to obtain a richer semantic description, from which the noisy query expansions are then filtered out. By treating each expansion as a "bag" and the retrieved images therein as "instances", we formulate image filtering as a multi-instance learning (MIL) problem with constrained positive bags. By this approach, images from different data distributions will be kept while with noisy images filtered out. Comprehensive experiments on two challenging tasks demonstrate the effectiveness of our proposed approach.

Reporting and summarizing metrics in sparse relationships on an OLTP database

Abstract: Storing and/or accessing data in a transactional database, with use of the following technique: (i) selecting a set of base tables in a transactional database; and (ii) creating a factless materialized query table, having maximum sparsity, for the set of base tables. The set of base tables includes at least two base tables. The set of base tables includes a set of keys including at least two distinct primary keys. The factless materialized query table includes one record associated with each record in the set of base tables. Each record in the materialized query table includes a value for every primary key in the set of base tables.

Online stochastic linear optimization under one-bit feedback

Abstract: In this paper, we study a special bandit setting of online stochastic linear optimization, where only one-bit of information is revealed to the learner at each round. This problem has found many applications including online advertisement and online recommendation. We assume the binary feedback is a random variable generated from the logit model, and aim to minimize the regret defined by the unknown linear function. Although the existing method for generalized linear bandit can be applied to our problem, the high computational cost makes it impractical for real-world applications. To address this challenge, we develop an efficient online learning algorithm by exploiting particular structures of the observation model. Specifically, we adopt online Newton step to estimate the unknown parameter and derive a tight confidence region based on the exponential concavity of the logistic loss. Our analysis shows that the proposed algorithm achieves a regret bound of O(d√T), which matches the optimal result of stochastic linear bandits.

Semi-supervised Convolutional Networks for Translation Adaptation with Tiny Amount of In-domain Data

Abstract: In this paper, we propose a method which uses semi-supervised convolutional neural networks (CNNs) to select in-domain training data for statistical machine translation. This approach is particularly effective when only tiny amounts of in-domain data are available. The in-domain data and randomly sampled general-domain data are used to train a data selection model with semi-supervised CNN, then this model computes domain relevance scores for all the sentences in the generaldomain data set. The sentence pairs with top scores are selected to train the system. We carry out experiments on 4 language directions with three test domains. Compared with strong baseline systems trained with large amount of data, this method can improve the performance up to 3.1 BLEU. Its performances are significant better than three state-of-the-art language model based data selection methods. We also show that the in-domain data used to train the selection model could be as fewas 100sentences, whichmakesfinegrained topic-dependent translation adaptation possible.

SunForum: Understanding Depression in a Chinese Online Community

Abstract: More than 350 million people worldwide suffer from depression. Major depressive disorder has a hugely negative impact on psychological well-being, work, and family life. Yet culture may shape how depressed patients interpret their symptoms, choose treatments, and behave. This paper reports a case study, including participant observations and interviews, of the Chinese online depression community, SunForum. Our findings reveal that Chinese cultural beliefs (e.g., the power of inner self-control) and Chinese beliefs about traditional medicine (e.g., the integrated body-mind relationship) significantly affect patients' understandings of depression, illness management, and social interactions. These beliefs create problems of understanding depression in society - including family members, friends, co-workers, and others - and present various challenges for depressed patients who can become marginalized, suffer discrimination, and lose their jobs. We draw implications for how Chinese society as a whole may respond to the misunderstanding of mental illness and the raising of public awareness. We also propose specific social media design to support depressed patients as they seek online information and social support.

Information Theoretic Subspace Clustering

Abstract: This paper addresses the problem of grouping the data points sampled from a union of multiple subspaces in the presence of outliers. Information theoretic objective functions are proposed to combine structured low-rank representations (LRRs) to capture the global structure of data and information theoretic measures to handle outliers. In theoretical part, we point out that group sparsity-induced measures ( $\ell _{2,1}$ -norm, $\ell _{\alpha }$ -norm, and correntropy) can be justified from the viewpoint of half-quadratic (HQ) optimization, which facilitates both convergence study and algorithmic development. In particular, a general formulation is accordingly proposed to unify HQ-based group sparsity methods into a common framework. In algorithmic part, we develop information theoretic subspace clustering methods via correntropy. With the help of Parzen window estimation, correntropy is used to handle either outliers under any distributions or sample-specific errors in data. Pairwise link constraints are further treated as a prior structure of LRRs. Based on the HQ framework, iterative algorithms are developed to solve the nonconvex information theoretic loss functions. Experimental results on three benchmark databases show that our methods can further improve the robustness of LRR subspace clustering and outperform other state-of-the-art subspace clustering methods.

System and method for processing task resources

Abstract: A device and method for automatically allocating computing resources is disclosed herein. The method includes receiving a task from a client, the task including a plurality of instances and a resource description manifest representing resource needs of the plurality of instances; determining an initial computing resource allocation of a cluster of machines based on the resource description manifest, wherein the initial computing resource allocation is determined based on the resource needs included in the resource description manifest; determining that the resource description manifest indicates a request to utilize an actual computing resource allocation in excess of the initial computing resource allocation; configuring a plurality of actual computing resources to process the plurality of instances, wherein the plurality of actual computing resources are configured to utilize resources in excess of the initial computing resource allocation; and executing the plurality of instances using the plurality of actual computing resources.

Express route transmissions between virtual machines and cloud service computing devices

Abstract: An apparatus includes a virtual machine (VM) that is accessible via a cloud service The VM has a virtual switch via which data is routed to and from the VM The virtual switch is configured to: access an express route flow table which indicates whether a particular front end computing device has previously received a first data packet from the VM, and determine, based on the express route flow table, whether to encapsulate a second data packet from the VM destined for the particular front end computing device for an express route so as to bypass a load balancer of the cloud service

Automatic image dataset construction with multiple textual metadata

Abstract: The goal of this work is to automatically collect a large number of highly relevant images from the Internet for given queries. A novel image dataset construction framework is proposed by employing multiple textual metadata. In specific, the given queries are first expanded by searching in the Google Books Ngrams Corpora to obtain a richer semantic description, from which the visually non-salient and less relevant expansions are then filtered. After retrieving images from the Internet with filtered expansions, we further filter noisy images by clustering and progressively Convolutional Neural Networks (CNN). To verify the effectiveness of our proposed method, we construct a dataset with 10 categories, which is not only much larger than but also have comparable cross-dataset generalization ability with manually labeled dataset STL-10 and CIFAR-10.

Statistics-based machine translation method, apparatus and electronic device

Abstract: A statistics-based machine translation method is disclosed. The method generates probabilities of translation from a sentence to be translated to candidate translated texts based on features of the candidate translated texts that affect the probabilities of translation and a pre-generated translation probability prediction model. The features that affect probabilities of translation include at least degrees of semantic similarity between the sentence to be translated and the candidate translated texts. A preset number of candidate translated texts with highly ranked probabilities of translation are selected to serve as translated texts of the sentence to be translated. The method is able to go deep into a semantic level of a natural language when a machine translation model is constructed to avoid a semantic deviation of a translated text from an original text, thereby achieving the effect of improving the quality of translation.

Mobile App Tagging

Abstract: Mobile app tagging aims to assign a list of keywords indicating core functionalities, main contents, key features or concepts of a mobile app. Mobile app tags can be potentially useful for app ecosystem stakeholders or other parties to improve app search, browsing, categorization, and advertising, etc. However, most mainstream app markets, e.g., Google Play, Apple App Store, etc., currently do not explicitly support such tags for apps. To address this problem, we propose a novel auto mobile app tagging framework for annotating a given mobile app automatically, which is based on a search-based annotation paradigm powered by machine learning techniques. Specifically, given a novel query app without tags, our proposed framework (i) first explores online kernel learning techniques to retrieve a set of top-N similar apps that are semantically most similar to the query app from a large app repository; and (ii) then mines the text data of both the query app and the top-N similar apps to discover the most relevant tags for annotating the query app. To evaluate the efficacy of our proposed framework, we conduct an extensive set of experiments on a large real-world dataset crawled from Google Play. The encouraging results demonstrate that our technique is effective and promising.

Method, apparatus, and system for quantum key distribution

Abstract: A quantum key distribution system includes a quantum security key management (QSKM) device, a plurality of quantum security key distribution (QSKD) devices, and a quantum security key service (QSKS) device. The QSKD device splits an identity-based system private key into a plurality of system sub-private keys, and distributes the plurality of system sub-private keys to a corresponding number of the QSKD devices. The QSKS device forwards a request for acquiring an authorized private key from a first QSKD device to a predetermined number of second QSKD devices. The predetermined number of second QSKD devices each generate an identity-based authorized sub-private key from the system sub-private key. The first QSKD device acquires, from the predetermined number of second QSKD devices, the identity-based authorized sub-private keys, and reconstructs an identity-based authorized private key based on the identity-based authorized sub-private keys.

Bilingual Methods for Adaptive Training Data Selection for Machine Translation

Abstract: In this paper, we propose a new data selection method which uses semi-supervised convolutional neural networks based on bitokens (Bi-SSCNNs) for training machine translation systems from a large bilingual corpus. In earlier work, we devised a data selection method based on semi-supervised convolutional neural networks (SSCNNs). The new method, Bi-SSCNN, is based on bitokens, which use bilingual information. When the new methods are tested on two translation tasks (Chinese-to-English and Arabic-to-English), they significantly outperform the other three data selection methods in the experiments. We also show that the BiSSCNN method is much more effective than other methods in preventing noisy sentence pairs from being chosen for training. More interestingly, this method only needs a tiny amount of in-domain data to train the selection model, which makes fine-grained topic-dependent translation adaptation possible. In the follow-up experiments, we find that neural machine translation (NMT) is more sensitive to noisy data than statistical machine translation (SMT). Therefore, Bi-SSCNN which can effectively screen out noisy sentence pairs, can benefit NMT much more than SMT.We observed a BLEU improvement over 3 points on an English-to-French WMT task when Bi-SSCNNs were used.

Deep CTR Prediction in Display Advertising

Abstract: Click through rate (CTR) prediction of image ads is the core task of online display advertising systems, and logistic regression (LR) has been frequently applied as the prediction model. However, LR model lacks the ability of extracting complex and intrinsic nonlinear features from handcrafted high-dimensional image features, which limits its effectiveness. To solve this issue, in this paper, we introduce a novel deep neural network (DNN) based model that directly predicts the CTR of an image ad based on raw image pixels and other basic features in one step. The DNN model employs convolution layers to automatically extract representative visual features from images, and nonlinear CTR features are then learned from visual features and other contextual features by using fully-connected layers. Empirical evaluations on a real world dataset with over 50 million records demonstrate the effectiveness and efficiency of this method.

Network system and method for connecting a private network with a virtual private network

Abstract: A networking method including receiving, at an edge router of a cloud data center, a virtual private cloud (“VPC”) network communication from a private network via a dedicated physical connection line to the edge router. The VPC network communication is forwarded to a gateway hardware group via a first connection using Virtual Extensible Local Area Network (“VXLAN”) technology. The VPC network communication is then forwarded from the gateway hardware group to VPC of a user of the private network via a second connection using VXLAN technology to access a virtual machine (“VM”).

PRIMT: A Pick-Revise Framework for Interactive Machine Translation

Abstract: Interactive machine translation (IMT) is a method which uses human-computer interactions to improve the quality of MT. Traditional IMT methods employ a left-to-right order for the interactions, which is difficult to directly modify critical errors at the end of the sentence. In this paper, we propose an IMT framework in which the interaction is decomposed into two simple human actions: picking a critical translation error (Pick) and revising the translation (Revise). The picked phrase could be at any position of the sentence, which improves the efficiency of human computer interaction. We also propose automatic suggestion models for the two actions to further reduce the cost of human interaction. Experiment results demonstrate that by interactions through either one of the actions, the translation quality could be significantly improved. Greater gains could be achieved by iteratively performing both actions.

Quantum key distribution system, method and apparatus based on trusted relay

Abstract: A quantum key distribution system is provided. The quantum key distribution system includes a plurality of routing devices configured to relay keys and a quantum key distribution device connected with the routing devices and configured to use two or more different paths to perform corresponding quantum key negotiations with another quantum key distribution device to obtain shared keys. The two or more different paths each include one or more of the routing devices.

Video eCommerce: Towards Online Video Advertising

Abstract: The prevalence of online videos provides an opportunity for e-commerce companies to exhibit their product ads in videos by recommendation. In this paper, we propose an advertising system named Video eCommerce to exhibit appropriate product ads to particular users at proper time stamps of videos, which takes into account video semantics, user shopping preference and viewing behavior feedback by a two-level strategy. At the first level, Co-Relation Regression (CRR) model is novelly proposed to construct the semantic association between keyframes and products. Heterogeneous information network (HIN) is adopted to build the user shopping preference from two different e-commerce platforms, Tmall and MagicBox, which alleviates the problems of data sparsity and cold start. In addition, Video Scene Importance Model (VSIM) utilizes the viewing behavior of users to embed ads at the most attractive position within the video stream. At the second level, taking the results of CRR, HIN and VSIM as the input, Heterogeneous Relation Matrix Factorization (HRMF) is applied for product advertising. Extensive evaluation on a variety of online videos from Tmall MagicBox demonstrates that Video eCommerce achieves promising performance, which significantly outperforms the state-of-the-art advertising methods.

An Empirical Study of Language CNN for Image Captioning

Abstract: Language Models based on recurrent neural networks have dominated recent image caption generation tasks. In this paper, we introduce a Language CNN model which is suitable for statistical language modeling tasks and shows competitive performance in image captioning. In contrast to previous models which predict next word based on one previous word and hidden state, our language CNN is fed with all the previous words and can model the long-range dependencies of history words, which are critical for image captioning. The effectiveness of our approach is validated on two datasets MS COCO and Flickr30K. Our extensive experimental results show that our method outperforms the vanilla recurrent neural network based language models and is competitive with the state-of-the-art methods.

Video playing method and device

Abstract: The present application provides a video playing method and apparatus The method includes: establishing a video message communication channel between a webpage main document including a webpage iframe and the webpage iframe; searching, by the webpage iframe, for a video marker included in the webpage iframe and acquiring video information according to the video marker; and receiving, by the webpage main document, the video information returned by the webpage iframe by using the communication channel, and performing video playing according to the video information By means of the present application, a player built in a current browser can be called to perform playing during video playing, to improve user experience in playing of cross-domain video services

Speech control method and device

Abstract: An embodiment of the invention discloses a speech control method and device so as to improve speech control efficiency. The method is characterized in that a speech system receives a speech command input by a user and judges whether the speech command is associated with a foreground application, if so, sends a control instruction corresponding the speech command to the foreground application to enable the foreground application to execute the control instruction, and if not, determines an application associated with the speech command in a background application and sends the control instruction corresponding the speech command to the background application to enable the background application to execute the control instruction, so that speech control efficiency can be improved effectively.

Data processing method and apparatus

Abstract: A method and an apparatus of data processing are disclosed The method includes obtaining a message including a non-triggered event; determining an event type and a triggering time of the non-triggered event according to the message; and providing a reminder for the non-triggered event according to the event type and the triggering time, and/or controlling a state of the mobile terminal according to the event type and the triggering time By determining a triggering time and an event type of a non-triggered event, a prompt may also be made before the triggering time even if a user has read the information Alternatively, the user may also be helped to adjust the state of the mobile terminal to a state related to the event type, so that the mobile terminal is in a state required by the user during the execution of the non-triggered event, thus improving the user experience

Exploiting Web Images for Dataset Construction: A Domain Robust Approach

Abstract: Labelled image datasets have played a critical role in high-level image understanding. However, the process of manual labelling is both time-consuming and labor intensive. To reduce the cost of manual labelling, there has been increased research interest in automatically constructing image datasets by exploiting web images. Datasets constructed by existing methods tend to have a weak domain adaptation ability, which is known as the "dataset bias problem". To address this issue, we present a novel image dataset construction framework that can be generalized well to unseen target domains. Specifically, the given queries are first expanded by searching the Google Books Ngrams Corpus to obtain a rich semantic description, from which the visually non-salient and less relevant expansions are filtered out. By treating each selected expansion as a "bag" and the retrieved images as "instances", image selection can be formulated as a multi-instance learning problem with constrained positive bags. We propose to solve the employed problems by the cutting-plane and concave-convex procedure (CCCP) algorithm. By using this approach, images from different distributions can be kept while noisy images are filtered out. To verify the effectiveness of our proposed approach, we build an image dataset with 20 categories. Extensive experiments on image classification, cross-dataset generalization, diversity comparison and object detection demonstrate the domain robustness of our dataset.

Method, apparatus, and system for identity authentication

Abstract: An identity authentication method for a quantum key distribution process includes selecting, by a sender, preparation bases of an identity authentication bit string in accordance with a preset basis vector selection rule; sending, by a sender, quantum states of the identity authentication bit string and quantum states of a randomly generated key bit string by using different wavelengths. The identity authentication bit string is interleaved in the key bit string at a random position and with a random length. The method further includes measuring, by a receiver, the received quantum states in the quantum state information in accordance with the different wavelengths and measurement bases selected according to the preset basis vector selection rule to obtain identity authentication information from the measurement of the identity authentication bit string; and determining, by the receiver, whether the identity authentication information obtained through the measurement corresponds with the preset basis vector selection rule.