Change search
Link to record
Permanent link

Direct link
BETA
Publications (10 of 100) Show all publications
Mrazovic, P., Larriba-Pey, J. L. & Matskin, M. (2018). A Deep Learning Approach for Estimating Inventory Rebalancing Demand in Bicycle Sharing Systems. In: Proceedings - International Computer Software and Applications Conference: . Paper presented at 42nd IEEE Computer Software and Applications Conference, COMPSAC 2018, 23 July 2018 through 27 July 2018 (pp. 110-115). IEEE Computer Society
Open this publication in new window or tab >>A Deep Learning Approach for Estimating Inventory Rebalancing Demand in Bicycle Sharing Systems
2018 (English)In: Proceedings - International Computer Software and Applications Conference, IEEE Computer Society , 2018, p. 110-115Conference paper, Published paper (Refereed)
Abstract [en]

Meeting user demand is one of the most challenging problems arising in public bicycle sharing systems. Various factors, such as daily commuting patterns or topographical conditions, can lead to an unbalanced state where the numbers of rented and returned bicycles differ significantly among the stations. This can cause spatial imbalance of the bicycle inventory which becomes critical when stations run completely empty or full, and thus prevent users from renting or returning bicycles. To prevent such service disruptions, we propose to forecast user demand in terms of expected number of bicycle rentals and returns and accordingly to estimate number of bicycles that need to be manually redistributed among the stations by maintenance vehicles. As opposed to traditional solutions to this problem, which rely on short-term demand forecasts, we aim to maximise the time within which the stations remain balanced by forecasting user demand multiple steps ahead of time. We propose a multi-input multi-output deep learning model based on Long Short-Term Memory networks to forecast user demand over long future horizons. Conducted experimental study over real-world dataset confirms the efficiency and accuracy of our approach.

Place, publisher, year, edition, pages
IEEE Computer Society, 2018
Keywords
Deep learning, Demand prediction, Smart cities, Time series forecasting, Application programs, Bicycles, Forecasting, MIMO systems, Smart city, Sporting goods, Learning approach, Learning models, Multi input multi output, Public bicycle sharing systems, Service disruptions, Short term memory
National Category
Computer and Information Sciences
Identifiers
urn:nbn:se:kth:diva-247214 (URN)10.1109/COMPSAC.2018.10213 (DOI)2-s2.0-85055538509 (Scopus ID)9781538626665 (ISBN)
Conference
42nd IEEE Computer Software and Applications Conference, COMPSAC 2018, 23 July 2018 through 27 July 2018
Note

QC 20190415

Available from: 2019-04-15 Created: 2019-04-15 Last updated: 2019-04-15Bibliographically approved
Hammar, K., Jaradat, S., Dokoohaki, N. & Matskin, M. (2018). Deep Text Mining of Instagram Data Without Strong Supervision. In: Proceedings - 2018 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2018: . Paper presented at 18th IEEE/WIC/ACM International Conference on Web Intelligence, WI 2018; Santiago; Chile; 3 December 2018 through 6 December 2018 (pp. 158-165). IEEE
Open this publication in new window or tab >>Deep Text Mining of Instagram Data Without Strong Supervision
2018 (English)In: Proceedings - 2018 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2018, IEEE, 2018, p. 158-165Conference paper, Published paper (Refereed)
Abstract [en]

With the advent of social media, our online feeds increasingly consist of short, informal, and unstructured text. This textual data can be analyzed for the purpose of improving user recommendations and detecting trends. Instagram is one of the largest social media platforms, containing both text and images. However, most of the prior research on text processing in social media is focused on analyzing Twitter data, and little attention has been paid to text mining of Instagram data. Moreover, many text mining methods rely on annotated training data, which in practice is both difficult and expensive to obtain. In this paper, we present methods for unsupervised mining of fashion attributes from Instagram text, which can enable a new kind of user recommendation in the fashion domain. In this context, we analyze a corpora of Instagram posts from the fashion domain, introduce a system for extracting fashion attributes from Instagram, and train a deep clothing classifier with weak supervision to classify Instagram posts based on the associated text. With our experiments, we confirm that word embeddings are a useful asset for information extraction. Experimental results show that information extraction using word embeddings outperforms a baseline that uses Levenshtein distance. The results also show the benefit of combining weak supervision signals using generative models instead of majority voting. Using weak supervision and generative modeling, an F1 score of 0.61 is achieved on the task of classifying the image contents of Instagram posts based solely on the associated text, which is on level with human performance. Finally, our empirical study provides one of the few available studies on Instagram text and shows that the text is noisy, that the text distribution exhibits the long-tail phenomenon, and that comment sections on Instagram are multi-lingual.

Place, publisher, year, edition, pages
IEEE, 2018
Keywords
Information extraction, Instagram, Weak Supervision, Word Embeddings
National Category
Language Technology (Computational Linguistics)
Identifiers
urn:nbn:se:kth:diva-245979 (URN)10.1109/WI.2018.00-94 (DOI)000458968200021 ()2-s2.0-85061892408 (Scopus ID)9781538673256 (ISBN)
Conference
18th IEEE/WIC/ACM International Conference on Web Intelligence, WI 2018; Santiago; Chile; 3 December 2018 through 6 December 2018
Note

QC 20190313

Available from: 2019-03-13 Created: 2019-03-13 Last updated: 2019-03-13Bibliographically approved
Fernando, T., Gureev, N., Matskin, M., Zwick, M. & Natschlager, T. (2018). WorkflowDSL: Scalable Workflow Execution with Provenance for Data Analysis Applications. In: Proceedings - International Computer Software and Applications Conference: . Paper presented at 42nd IEEE Computer Software and Applications Conference, COMPSAC 2018, 23 July 2018 through 27 July 2018 (pp. 774-779). IEEE Computer Society
Open this publication in new window or tab >>WorkflowDSL: Scalable Workflow Execution with Provenance for Data Analysis Applications
Show others...
2018 (English)In: Proceedings - International Computer Software and Applications Conference, IEEE Computer Society , 2018, p. 774-779Conference paper, Published paper (Refereed)
Abstract [en]

Data analysis projects typically use different programming languages (from Python for prototyping to C++ for support of runtime constraints) at their different stages by different experts. This creates a need for a data processing framework that is re-usable across multiple programming languages and supports collaboration of experts. In this work, we discuss implementation of a framework which uses a Domain Specific Language (DSL), called WorkflowDSL, that enables domain experts to collaborate on fine-tuning workflows. The framework includes support for parallel execution without any specialized code. It also provides a provenance capturing framework that enables users to analyse past executions and retrieve complete lineage of any data item generated. Graph database is used for storing provenance data. Advantages of usage of a graph database compare to relational databases are demonstrated. Experiments which were performed using a real-world scientific workflow from the bioinformatics domain and industrial data analysis models show that users were able to execute workflows efficiently when using WorkflowDSL for workflow composition and Python for task implementations. Moreover, we show that capturing provenance data can be useful for analysing past workflow executions.

Place, publisher, year, edition, pages
IEEE Computer Society, 2018
Keywords
Data analysis workflows, Linage, Parallel execution, Provenance, Application programs, C++ (programming language), Graph Databases, Information analysis, Problem oriented languages, Software prototyping, Domain specific language (DSL), Parallel executions, Relational Database, Scientific workflows, Work-flows, Workflow composition, Data handling
National Category
Computer and Information Sciences
Identifiers
urn:nbn:se:kth:diva-247215 (URN)10.1109/COMPSAC.2018.00115 (DOI)2-s2.0-85055419742 (Scopus ID)9781538626665 (ISBN)
Conference
42nd IEEE Computer Software and Applications Conference, COMPSAC 2018, 23 July 2018 through 27 July 2018
Note

QC 20190415

Available from: 2019-04-15 Created: 2019-04-15 Last updated: 2019-04-15Bibliographically approved
Mrazovic, P., De La Rubia, I., Urmeneta, J., Balufo, C., Tapias, R., Matskin, M. & Larriba-Pey, J. L. (2016). CIGO! Mobility Management Platform for Growing Efficient and Balanced Smart City Ecosystem. In: IEEE SECOND INTERNATIONAL SMART CITIES CONFERENCE (ISC2 2016): . Paper presented at 2nd IEEE International Smart Cities Conference (ISC2), SEP 12-15, 2016, Trento, ITALY (pp. 106-109). IEEE
Open this publication in new window or tab >>CIGO! Mobility Management Platform for Growing Efficient and Balanced Smart City Ecosystem
Show others...
2016 (English)In: IEEE SECOND INTERNATIONAL SMART CITIES CONFERENCE (ISC2 2016), IEEE, 2016, p. 106-109Conference paper, Published paper (Refereed)
Abstract [en]

The massive amount of tourists, citizens and traffic in big cities usually collapse busy areas causing transport inefficiency, unbalanced economic growth, crime, and nuisance among citizens and visitors. Therefore, the Smart City strategies such as Smart Mobility and Smart Governance naturally arise as means to improve mobility in urban areas. In this paper we propose a novel mobility management platform and business model that can attract numerous actors and still be orchestrated by the city government. The proposed platform integrates mobility data from various sources such as Open Data, mobile applications, sensors and government data, allowing for its visualisation and analysis while making it actionable through associated third party mobile applications. We propose to inject the city mobility policies to the third party mobile applications which provide services related to the city resources. In this way we form a value chain which connects different actors (city governments, mobile application providers, POI owners, companies that require logistics in cities, and final users) who both take a part in improving the mobility in urban areas, and benefit from the way mobility policies being executed. In this paper we discuss the business model and logical architecture of the proposed platform which has been already deployed in the city of Barcelona.

Place, publisher, year, edition, pages
IEEE, 2016
Keywords
smart cities, smart mobility, mobility policies
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
urn:nbn:se:kth:diva-202502 (URN)10.1109/ISC2.2016.7580750 (DOI)000392263700020 ()2-s2.0-84994127946 (Scopus ID)978-1-5090-1845-1 (ISBN)
Conference
2nd IEEE International Smart Cities Conference (ISC2), SEP 12-15, 2016, Trento, ITALY
Note

QC 20170228

Available from: 2017-02-28 Created: 2017-02-28 Last updated: 2017-03-07Bibliographically approved
Sato, H., Matskin, M. & Claycomb, W. (2016). Message from the Program Chairs-in-Chief. Paper presented at 10 June 2016 through 14 June 2016. Computer Software and Applications Conference, 1, Article ID 7551982.
Open this publication in new window or tab >>Message from the Program Chairs-in-Chief
2016 (English)In: Computer Software and Applications Conference, ISSN 0730-3157, Vol. 1, article id 7551982Article in journal (Refereed) Published
Place, publisher, year, edition, pages
IEEE Computer Society, 2016
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
urn:nbn:se:kth:diva-194940 (URN)10.1109/COMPSAC.2016.14 (DOI)2-s2.0-84988039968 (Scopus ID)
Conference
10 June 2016 through 14 June 2016
Note

QC 20161124

Available from: 2016-11-24 Created: 2016-11-01 Last updated: 2017-11-29Bibliographically approved
Sato, H., Matskin, M. & Claycomb, W. (2016). Message from the Program Chairs-in-Chief - Volume 2. Paper presented at 10 June 2016 through 14 June 2016. Computer Software and Applications Conference, 2, Article ID 7551963.
Open this publication in new window or tab >>Message from the Program Chairs-in-Chief - Volume 2
2016 (English)In: Computer Software and Applications Conference, ISSN 0730-3157, Vol. 2, article id 7551963Article in journal (Refereed) Published
Place, publisher, year, edition, pages
IEEE Computer Society, 2016
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
urn:nbn:se:kth:diva-194941 (URN)10.1109/COMPSAC.2016.241 (DOI)2-s2.0-84988037177 (Scopus ID)
Conference
10 June 2016 through 14 June 2016
Note

QC 20161124

Available from: 2016-11-24 Created: 2016-11-01 Last updated: 2017-11-29Bibliographically approved
Jaradat, S., Dokoohaki, N., Matskin, M. & Ferrari, E. (2016). Trust And Privacy Correlations in Social Networks: A Deep Learning Framework. In: PROCEEDINGS OF THE 2016 IEEE/ACM INTERNATIONAL CONFERENCE ON ADVANCES IN SOCIAL NETWORKS ANALYSIS AND MINING ASONAM 2016: . Paper presented at 8th IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), AUG 18-21, 2016, San Francisco, CA (pp. 203-206). IEEE conference proceedings
Open this publication in new window or tab >>Trust And Privacy Correlations in Social Networks: A Deep Learning Framework
2016 (English)In: PROCEEDINGS OF THE 2016 IEEE/ACM INTERNATIONAL CONFERENCE ON ADVANCES IN SOCIAL NETWORKS ANALYSIS AND MINING ASONAM 2016, IEEE conference proceedings, 2016, p. 203-206Conference paper, Published paper (Refereed)
Abstract [en]

Online Social Networks (OSNs) remain the focal point of Internet usage. Since the beginning, networking sites tried best to have right privacy mechanisms in place for users, enabling them to share the right content with the right audience. With all these efforts, privacy customizations remain hard for users across the sites. Existing research that address this problem mainly focus on semi-supervised strategies that introduce extra complexity by requiring the user to manually specify initial privacy preferences for their friends. In this work, we suggest an adaptive solution that can dynamically generate privacy labels for users in OSNs. To this end, we introduce a deep reinforcement learning framework that targets two key problems in OSNs like Facebook: the exposure of users' interactions through the network to less trusted direct friends, and the possibility of propagating user updates through direct friends' interactions to indirect friends. By implementing this framework, we aim at understanding how social trust and privacy could be correlated, specifically in a dynamic fashion. We report the ranked dependence between the generated privacy labels and the estimated user trust values, which indicate the ability of the framework to identify the highly trusted users and share with them higher percentages of data.

Place, publisher, year, edition, pages
IEEE conference proceedings, 2016
National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-200264 (URN)000390760100031 ()2-s2.0-85006765626 (Scopus ID)978-1-5090-2846-7 (ISBN)
Conference
8th IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), AUG 18-21, 2016, San Francisco, CA
Note

QC 20170130

Available from: 2017-01-30 Created: 2017-01-23 Last updated: 2018-01-13Bibliographically approved
Ozyagci, O. Z. & Matskin, M. (2016). Truthful Incentive Mechanism for Mobile Crowdsensing with Smart Consumer Devices. In: Proceedings - International Computer Software and Applications Conference: . Paper presented at 2016 IEEE 40th Annual Computer Software and Applications Conference, COMPSAC 2016, 10 June 2016 through 14 June 2016 (pp. 282-287). IEEE Computer Society
Open this publication in new window or tab >>Truthful Incentive Mechanism for Mobile Crowdsensing with Smart Consumer Devices
2016 (English)In: Proceedings - International Computer Software and Applications Conference, IEEE Computer Society, 2016, p. 282-287Conference paper, Published paper (Refereed)
Abstract [en]

Smart consumer devices have become one of the fundamental communication and computing devices in people's everyday lives over the past decade. Their various sensors and wireless connectivity have paved the way for a new application area called mobile crowdsensing (MCS) where sensing services are provided by using the sensor outputs collected from smart consumer devices. MCS system's service quality heavily depends on the participation of smart device users who probably expect to be compensated in return for their participation. Therefore, MCS applications need incentive mechanisms to motivate such people into participating. In this work, we first defined a reverse auction based incentive mechanism for a representative MCS system. Then, we integrated the Vickrey-Clarke-Groves (VCG) mechanism into the initial incentive mechanism so that truthful bidding would become the dominant strategy in the resulting incentive mechanism. Finally, we conducted simulations of both incentive mechanisms in order to measure the fairness of service prices and the fairness of cumulative participant earnings using Jain's fairness index. We observed that both the fairness of service prices and the fairness of cumulative participant earnings were generally better in the derived incentive mechanism when the VCG mechanism was applicable. We also found that at least 70% of service requests had fair prices, while between 5% and 85% of participants had fair cumulative earnings in both incentive mechanisms.

Place, publisher, year, edition, pages
IEEE Computer Society, 2016
Keywords
fairness, mobile crowdsensing, social network services for consumer devices, truthful incentive mechanism, Vickrey-Clarke-Groves mechanism, Computer software, Costs, Quality of service, Consumer devices, Incentive mechanism, Application programs
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
urn:nbn:se:kth:diva-194938 (URN)10.1109/COMPSAC.2016.214 (DOI)000389532200042 ()2-s2.0-84987984846 (Scopus ID)9781467388450 (ISBN)
Conference
2016 IEEE 40th Annual Computer Software and Applications Conference, COMPSAC 2016, 10 June 2016 through 14 June 2016
Note

QC 20161124

Available from: 2016-11-24 Created: 2016-11-01 Last updated: 2017-01-23Bibliographically approved
Claycomb, W., Matskin, M. & Nakamura, M. (2015). Message from the Workshop Chairs - Part III. In: Proceedings - International Computer Software and Applications Conference: . Paper presented at 39th IEEE Annual Computer Software and Applications Conference Workshops, COMPSACW 2015; Taichung; Taiwan. IEEE Communications Society, 3
Open this publication in new window or tab >>Message from the Workshop Chairs - Part III
2015 (English)In: Proceedings - International Computer Software and Applications Conference, IEEE Communications Society, 2015, Vol. 3Conference paper, Published paper (Refereed)
Place, publisher, year, edition, pages
IEEE Communications Society, 2015
National Category
Computer Sciences
Identifiers
urn:nbn:se:kth:diva-187109 (URN)10.1109/COMPSAC.2015.322 (DOI)2-s2.0-84962091477 (Scopus ID)
Conference
39th IEEE Annual Computer Software and Applications Conference Workshops, COMPSACW 2015; Taichung; Taiwan
Note

QC 20160519

Available from: 2016-05-19 Created: 2016-05-17 Last updated: 2018-01-10Bibliographically approved
Mrazovic, P. & Matskin, M. (2015). MobiCS: Mobile Platform for Combining Crowdsourcing and Participatory Sensing. In: : . Paper presented at Computer Software and Applications Conference (COMPSAC), 2015 IEEE 39th Annual (pp. 553-562). IEEE, 2
Open this publication in new window or tab >>MobiCS: Mobile Platform for Combining Crowdsourcing and Participatory Sensing
2015 (English)Conference paper, Published paper (Refereed)
Abstract [en]

Current participatory sensing approaches usuallydo not consider device carriers as intelligent participants insensing processes. However, modern mobile communicationdevices allow users express their opinions and judgementswhich can complement to captured sensor data. In this paperwe bring together different modes of mobile crowdsourcinginto a general sensing platform which treats device carriersas intelligent problem solvers. We propose a conceptual archi-tecture for versatile context-aware mobile crowdsourcing, andaddress issues related to data representation, quality control,trust and reputation management, and task allocation. Toprove the potential advantages of the proposed conceptualarchitecture we developedMobiCS, a prototype platform whichallows crowdsourcers formulate and distribute both sensingand human intelligence tasks to Android-powered mobilecommunication devices.

Place, publisher, year, edition, pages
IEEE, 2015
Keywords
mobile crowdsourcing, participatory sensing, mobile crowd sensing
National Category
Communication Systems
Research subject
Information and Communication Technology
Identifiers
urn:nbn:se:kth:diva-183157 (URN)10.1109/COMPSAC.2015.26 (DOI)000380584300075 ()2-s2.0-84962142430 (Scopus ID)
Conference
Computer Software and Applications Conference (COMPSAC), 2015 IEEE 39th Annual
Note

QC 20160415

Available from: 2016-03-02 Created: 2016-03-02 Last updated: 2017-01-20Bibliographically approved
Organisations
Identifiers
ORCID iD: ORCID iD iconorcid.org/0000-0002-4722-0823

Search in DiVA

Show all publications