Umair ul Hassan

A Multi-armed Bandit Approach to Online Spatial Task Assignment

Spatial crowd sourcing uses workers for performing tasks that require travel to different locations in the physical world. This paper considers the online spatial task assignment problem. In this problem, spatial tasks arrive in an online manner and an appropriate worker must be assigned to each task. However, outcome of an assignment is stochastic since the worker can choose to accept or reject the task. Primary goal of the assignment algorithm is to maximize the number of successful assignments over all tasks. This presents an exploration-exploitation challenge, the algorithm must learn the task acceptance behavior of workers while selecting the best worker based on the previous learning. We address this challenge by defining a framework for online spatial task assignment based on the multi-armed bandit formalization of the problem. Furthermore, we adapt a contextual bandit algorithm to assign a worker based on the spatial features of tasks and workers. The algorithm simultaneously adapts the worker assignment strategy based on the observed task acceptance behavior of workers. Finally, we present an evaluation methodology based on a real world dataset, and evaluate the performance of the proposed algorithm against the baseline algorithms. The results demonstrate that the proposed algorithm performs better in terms of the number of successful assignments.

Leveraging matching dependencies for guided user feedback in linked data applications

This paper presents a new approach for managing integration quality and user feedback, for entity consolidation, within applications consuming Linked Open Data. The quality of a dataspace containing multiple linked datasets is defined in term of a utility measure, based on domain specific matching dependencies. Furthermore, the user is involved in the consolidation process through soliciting feedback about identity resolution links, where each candidate link is ranked according to its benefit to the dataspace; calculated by approximating the improvement in the utility of dataspace utility. The approach evaluated on real world and synthetic datasets demonstrates the effectiveness of utility measure; through dataspace integration quality improvement that requires less overall user feedback iterations.

ACRyLIQ: Leveraging DBpedia for Adaptive Crowdsourcing in Linked Data Quality Assessment

Crowdsourcing has emerged as a powerful paradigm for quality assessment and improvement of Linked Data. A major challenge of employing crowdsourcing, for quality assessment in Linked Data, is the cold-start problem: how to estimate the reliability of crowd workers and assign the most reliable workers to tasks? We address this challenge by proposing a novel approach for generating test questions from DBpedia based on the topics associated with quality assessment tasks. These test questions are used to estimate the reliability of the new workers. Subsequently, the tasks are dynamically assigned to reliable workers to help improve the accuracy of collected responses. Our proposed approach, ACRyLIQ, is evaluated using workers hired from Amazon Mechanical Turk, on two real-world Linked Data datasets. We validate the proposed approach in terms of accuracy and compare it against the baseline approach of reliability estimate using gold-standard task. The results demonstrate that our proposed approach achieves high accuracy without using gold-standard task.

Flag-verify-fix: adaptive spatial crowdsourcing leveraging location-based social networks

This paper introduces the flag-verify-fix pattern that employs spatial crowdsourcing for city maintenance. The patterns motivates the need for appropriate assignment of dynamically arriving spatial tasks to a pool for workers on the ground. The assignment is aimed at maximizing the coverage of tasks spread over spatial locations; however, the coverage depends of willingness of workers to perform tasks assigned to them. We introduce the maximum coverage assignment problem that formulates two design issues of dynamic assignment. The quantity issue determines the number of worker required for a task and selection issue determines the set of workers. We propose an adaptive algorithm that uses location diversity based on a location-based social network to address the quantity issue and employs Thompson sampling for selecting the workers by learning their willingness. We evaluate the performance of the proposed algorithm in terms of coverage and number of assignments using real world datasets. The results show that our proposed algorithm achieves 30%--50% more coverage than the baseline algorithms, while requiring less workers per task.

A Real-time Linked Dataspace for the Internet of Things: Enabling “Pay-As-You-Go” Data Management in Smart Environments

Abstract As smart environments move from a research vision to concrete manifestations in real-world enabled by the Internet of Things, they are encountering a number of very practical challenges in data management in terms of the flexibility needed to bring together contextual and real-time data, the interface between new digital infrastructures and existing information systems, and how to easily share data between stakeholders in the environment. Therefore, data management approaches for smart environments need to support flexibility, dynamicity, incremental change, while keeping costs to a minimum. A Dataspace is an emerging approach to data management that has proved fruitful for personal information and scientific data management. However, their use within smart environments and for real-time data remains largely unexplored. This paper introduces a Real-time Linked Dataspace (RLD) as an enabling platform for data management within smart environments. This paper identifies common data management requirements for smart energy and water environments, details the RLD architecture and the key support services and their tiered support levels, and a principled approach to “Pay-As-You-Go” data management. The paper presents a dataspace query service for real-time data streams and entities to enable unified entity-centric queries across live and historical stream data. The RLD was validated in 5 real-world pilot smart environments following the OODA (Observe, Orient, Decide, and Act) Loop to build real-time analytics, decisions support, and smart apps for energy and water management. The pilots demonstrate that the RLD enables incremental pay-as-you-go data management with support services that simplify the development of applications and analytics for smart environments. Finally, the paper discusses experiences, lessons learnt, and future directions.

Automatic Anomaly Detection over Sliding Windows: Grand Challenge

With the advances in the Internet of Things and rapid generation of vast amounts of data, there is an ever growing need for leveraging and evaluating event-based systems as a basis for building realtime data analytics applications. The ability to detect, analyze, and respond to abnormal patterns of events in a timely manner is as challenging as it is important. For instance, distributed processing environment might affect the required order of events, time-consuming computations might fail to scale, or delays of alarms might lead to unpredicted system behavior. The ACM DEBS Grand Challenge 2017 focuses on real-time anomaly detection for manufacturing equipments based on the observation of a stream of measurements generated by embedded digital and analogue sensors. In this paper, we present our solution to the challenge leveraging the Apache Flink stream processing framework and anomaly ordering based on sliding windows, and evaluate the performance in terms of event latency and throughput.

Water Analytics and Management with Real-Time Linked Dataspaces

The research leading to these results has received funding under the European Commission’s Seventh Framework Programme from ICT grant agreement WATERNOMICS no. 619660. It is supported in part by Science Foundation Ireland (SFI) under Grant Number SFI/12/RC/2289.