Today’s artificial intelligence still faces two major challenges. One is that, in most industries, data exists the form of isolated islands. The other strengthening privacy and security. We propose a possible solution to these challenges: secure federated learning. Beyond federated-learning framework first proposed by Google 2016, we introduce comprehensive framework, which includes horizontal learning, vertical transfer provide definitions, architectures, applications for survey existing...

Today's AI still faces two major challenges. One is that in most industries, data exists the form of isolated islands. The other strengthening privacy and security. We propose a possible solution to these challenges: secure federated learning. Beyond learning framework first proposed by Google 2016, we introduce comprehensive framework, which includes horizontal learning, vertical transfer provide definitions, architectures applications for survey existing works on this subject. In addition,...

With the rapid development of smartphones, spatial crowdsourcing platforms are getting popular. A foundational research is to allocate micro-tasks suitable crowd workers. Most existing studies focus on offline scenarios, where all spatiotemporal information and workers given. However, they impractical since in real applications appear dynamically their cannot be known advance. In this paper, address shortcomings approaches, we first identify a more practical micro-task allocation problem,...

Taxi-calling apps are gaining increasing popularity for their efficiency in dispatching idle taxis to passengers need. To precisely balance the supply and demand of taxis, online taxicab platforms need predict Unit Original Taxi Demand (UOTD), which refers number taxi-calling requirements submitted per unit time (e.g., every hour) region each POI). Predicting UOTD is non-trivial large-scale industrial because both accuracy flexibility essential. Complex non-linear models such as GBRT deep...

Recently, with the development of mobile Internet and smartphones, <u>o</u>nline <u>m</u>inimum <u>b</u>ipartite <u>m</u>atching in real time spatial data (OMBM) problem becomes popular. Specifically, given a set service providers specific locations users who dynamically appear one by one, OMBM is to find maximum-cardinality matching minimum total distance following that once user appears, s/he must be immediately matched an unmatched provider,...

Due to stricter data management regulations such as General Data Protection Regulation (GDPR), traditional production mode of machine learning services is shifting federated learning, a paradigm that allows multiple providers train joint model collaboratively with their kept locally. A key enabler for practical adoption how allocate the prolit earned by each provider. For fair allocation, metric quantify contribution provider essential. Shapley value classical concept in cooperative game...

In recent years, due to the wide applications of uncertain data, mining frequent itemsets over databases has attracted much attention. databases, support an itemset is a random variable instead fixed occurrence counting this itemset. Thus, unlike corresponding problem in deterministic where unique definition, under environments two different definitions so far. The first referred as expected support-based itemset, employs expectation measure whether frequent. second probabilistic uses...

There has been a dramatic growth of shared mobility applications such as ride-sharing, food delivery and crowdsourced parcel delivery. Shared refers to transportation services that are among users, where central issue is route planning . Given set workers requests, finds for each worker route, i.e. , sequence locations pick up drop off passengers/parcels arrive from time time, with different optimization objectives. Previous studies lack practicability due their conflicted objectives...

The popularity of Online To Offline (O2O) service platforms has spurred the need for online task assignment in real-time spatial data, where streams spatially distributed tasks and workers are matched real time such that total number assigned pairs is maximized. Existing models assume each worker either a immediately or waits subsequent at fixed location once she/he appears on platform. Yet practice may actively move around rather than passively wait place if no assigned. In this paper, we...

It is universal to see people obtain knowledge on micro-blog services by asking others decision making questions. In this paper, we study the Jury Selection Problem(JSP) utilizing crowdsourcing for tasks services. Specifically, problem enroll a subset of crowd under limited budget, whose aggregated wisdom via Majority Voting scheme has lowest probability drawing wrong answer(Jury Error Rate-JER). Due various individual error-rates crowd, calculation JER non-trivial. Firstly, explicitly state...

As one of the successful forms using Wisdom Crowd, crowdsourcing, has been widely used for many human intrinsic tasks, such as image labeling, natural language understanding, market predication and opinion mining. Meanwhile, with advances in pervasive technology, mobile devices, phones tablets, have become extremely popular. These devices can work sensors to collect multimedia data(audios, images videos) location information. This power makes it possible implement new crowdsourcing mode:...

The prevalence of mobile Internet techniques and Online-To-Offline (O2O) business models has led the emergence various spatial crowdsourcing (SC) platforms in our daily life. A core issue SC is to assign real-time tasks suitable crowd workers. Existing approaches usually focus on matching two types objects, workers, or assume static offline scenarios, where spatio-temporal information all workers known advance. Recently, some new emerging O2O applications incur challenges: need three tasks,...

In spatial crowdsourcing, requesters submit their task-related locations and increase the demand of a local area. The platform prices these tasks assigns workers to serve if are accepted by requesters. There exist mature pricing strategies which specialize in tackling imbalance between supply market. However, global optimization, should consider mobility workers; that is, any single worker can be potential for several areas, while it only true one area when assigned platform. hardness lies...

Crowdsourcing is a new computing paradigm where humans are actively enrolled to participate in the procedure of computing, especially for tasks that intrinsically easier than computers. The popularity mobile and sharing economy has extended conventional web-based crowdsourcing spatial (SC), data such as location, mobility associated contextual information, plays central role. In fact, stimulated series recent industrial successes including Citizen Sensing (Waze), P2P ride-sharing (Uber)...

Crowdsourcing has been shown to be effective in a wide range of applications, and is seeing increasing use. A large-scale crowdsourcing task often consists thousands or millions atomic tasks, each which usually simple such as binary choice voting. To distribute limited crowd workers, common practice pack set tasks into bin send worker batch. It challenging decompose execute batches ensures reliable answers at minimal total cost. Large lead unreliable while small incur unnecessary In this...

Online event-based social network (EBSN) platforms are becoming popular these days. An important task of managing EBSNs is to arrange proper events interested users. Existing approaches usually assume that each user only attends one event or ignore location information. The overall utility such strategy limited in real world: 1) may attend multiple events; 2) attending will incur spatio-temporal conflicts and travel expenses. Thus, a more intelligent EBSN platform provides personalized...

Latent Dirichlet Allocation (LDA) is a widely adopted topic model for industrial-grade text mining applications. However, its performance heavily relies on the collection of large amount data from users' everyday life training. Such risks severe privacy leakage if collector untrustworthy. To protect while allowing accurate training, we investigate federated learning LDA models. That is, collaboratively trained between an untrustworthy and multiple users, where raw each user are stored...

Online bipartite graph matching is attracting growing research attention due to the development of dynamic task assignment in sharing economy applications, where tasks need be assigned dynamically workers. Past studies lack practicability terms both problem formulation and solution framework. On one hand, some settings prior online are impractical for real-world applications. other existing solutions inefficient unnecessary real-time decision making. In this paper, we propose (DBGM) better...

A central issue in on-demand taxi dispatching platforms is task assignment, which designs matching policies among dynamically arrived drivers (workers) and passengers (tasks). Previous maximize the profit of platform without considering preferences workers tasks (e.g., may prefer high-rewarding while nearby workers). Such ignorance impairs user experience will decrease long run. To address this problem, we propose preference-aware assignment using online stable matching. Specifically, define...

With the rapid development of Web 2.0 and Online To Offline (O2O) marketing model, various <i>online <u> e</u>vent-<u>b</u>ased <u>s</u>ocial <u>n</u>etwork<u>s </u></i> (EBSNs) are getting popular. An important task EBSNs is to facilitate most satisfactory event-participant arrangement for both sides, i.e., events enroll more participants arranged with personally interesting events. Existing approaches usually focus on each single event a set potential users, or ignore conflicts between...

To benefit from the Cloud platform's unlimited resources, managing and evaluating huge volume of RDF data in a scalable manner has attracted intensive research efforts recently. Progresses have been made on SPARQL queries with either high-level declarative programming languages, like Pig [1], or sequence sophisticated designed MapReduce jobs, both which tend to answer query multiple join operations. However, due simplicity storage coarse organization existing solutions, operations easily...

With the rapid development of mobile internet and online to offline marketing model, various spatial crowdsourcing platforms, such as Gigwalk Gmission, are getting popular. Most existing studies assume that crowdsourced tasks simple trivial. However, many real complex need be collaboratively finished by a team crowd workers with different skills. Therefore, an important issue platforms is recommend some suitable teams satisfy requirements skills in task. In this paper, address issue, we...

With the rapid development of smartphones, spatial crowdsourcing platforms are getting popular. A foundational research is to allocate micro-tasks suitable crowd workers. Many existing studies focus on offline scenario, where all spatiotemporal information and workers given. In this paper, we online scenario identify a more practical micro-task allocation problem, called <i><u>G</u>lobal <u>O</u>nline <u>M</u>icro-task <u>A</u>llocation in crowdsourcing</i> (GOMA) problem. We first extend...