A5001443156- Peer-to-Peer Network Technologies
- Semantic Web and Ontologies
- Caching and Content Delivery
- Advanced Database Systems and Queries
- Data Management and Algorithms
- Distributed and Parallel Computing Systems
- Service-Oriented Architecture and Web Services
- Topic Modeling
- Web Data Mining and Analysis
- Access Control and Trust
- Mobile Crowdsensing and Crowdsourcing
- Complex Network Analysis Techniques
- Privacy-Preserving Technologies in Data
- Human Mobility and Location-Based Analysis
- Time Series Analysis and Forecasting
- Natural Language Processing Techniques
- Misinformation and Its Impacts
- Data Stream Mining Techniques
- Scientific Computing and Data Management
- Context-Aware Activity Recognition Systems
- Energy Efficient Wireless Sensor Networks
- Spam and Phishing Detection
- Advanced Data Storage Technologies
- Multimedia Communication and Technology
- Business Process Modeling and Analysis
École Polytechnique Fédérale de Lausanne
2015-2025
Laboratoire d'Informatique Fondamentale de Lille
2008-2023
The University of Queensland
2021
Humboldt-Universität zu Berlin
2021
University of Pittsburgh
2019-2021
Purdue University System
2019-2021
Drexel University
2019-2021
Swinburne University of Technology
2019-2021
IBM (United States)
2019-2021
Georgia Institute of Technology
2019-2021
Managing trust is a problem of particular importance in peer-to-peer environments where one frequently encounters unknown agents. Existing methods for management, that are based on reputation, focus the semantic properties model. They do not scale as they either rely central database or require to maintain global knowledge at each agent provide data earlier interactions. In this paper we present an approach addresses reputation-based management both and level. We employ levels scalable...
With the price of wireless sensor technologies diminishing rapidly we can expect large numbers autonomous networks being deployed in near future. These will typically not remain isolated but need interconnecting them on network level to enable integrated data processing arise, thus realizing vision a global "sensor Internet." This requires flexible middleware layer which abstracts from underlying, heterogeneous and supports fast simple deployment addition new platforms, facilitates efficient...
Location-based social networks (LBSNs) offer researchers rich data to study people's online activities and mobility patterns. One important application of such studies is provide personalized point-of-interest (POI) recommendations enhance user experience in LBSNs. Previous solutions directly predict users' preference on locations but fail insights about transitions among locations. In this work, we propose a novel category-aware POI recommendation model, which exploits the transition...
Power consumption on mobile phones is a painful obstacle towards adoption of continuous sensing driven applications, e.g., continuously inferring individual's locomotive activities (such as 'sit', 'stand' or 'walk') using the embedded accelerometer sensor. To reduce energy overhead such activity sensing, we first investigate how choice sampling frequency & classification features affects, separately for each activity, "energy overhead" vs. "classification accuracy" tradeoff. We find that...
With the large-scale adoption of GPS equipped mobile sensing devices, positional data generated by moving objects (e.g., vehicles, people, animals) are being easily collected. Such typically modeled as streams spatio-temporal (x,y,t) points, called trajectories . In recent years trajectory management research has progressed significantly towards efficient storage and indexing techniques, well suitable knowledge discovery. These works focused on geometric aspect raw mobility data. We now...
Background With increased specialization of health care services and high levels patient mobility, accessing across multiple hospitals or clinics has become very common for diagnosis treatment, particularly patients with chronic diseases such as cancer. informed knowledge a patient’s history, physicians can make prompt clinical decisions smarter, safer, more efficient care. However, due to the privacy sensitivity electronic records (EHR), most EHR data sharing still happens through fax mail...
article Share on P-Grid: a self-organizing structured P2P system Authors: Karl Aberer Distributed Information Systems Laboratory, École Polytechnique Fédérale de Lausanne (EPFL) (EPFL)View Profile , Philippe Cudré-Mauroux Anwitaman Datta Zoran Despotovic Manfred Hauswirth Magdalena Punceva Roman Schmidt Authors Info & Claims ACM SIGMOD RecordVolume 32Issue 3September 2003 pp 29–33https://doi.org/10.1145/945721.945729Published:01 September 2003Publication History...
A key problem in current sensor network technology is the heterogeneity of available software and hardware platforms which makes deployment application development a tedious time consuming task. To minimize unnecessary repetitive implementation identical functionalities for different platforms, we present our Global Sensor Networks (GSN) middleware supports flexible integration discovery networks data, enables fast addition new provides distributed querying, filtering, combination dynamic...
The authors present Gridella, a Gnutella-compatible P2P system. Gridella is based on the Peer-Grid (P-Grid) approach, which draws research in distributed and cooperative information systems to provide decentralized, scalable data access structure. improves highly chaotic inefficient Gnutella infrastructure with directed search advanced concepts, thus enhancing efficiency providing model for further analysis research.
This paper studies the problem of updates in decentralised and self-organising P2P systems which peers have low online probabilities only local knowledge. The update strategy we propose for this environment is based on a hybrid push/pull rumor spreading algorithm provides fully decentralised, efficient robust communication scheme offers probabilistic guarantees rather than ensuring strict consistency. We describe generic analytical model to investigate utility our propagation from...
GPS devices allow recording the movement track of moving object they are attached to. This data typically consists a stream spatio-temporal (x,y,t) points. For application purposes is transformed into finite subsequences called trajectories. Existing knowledge extraction algorithms defined for trajectories mainly assume specific context (e.g. vehicle movements) or analyze parts trajectory stops), in association with from chosen geographic sources points-of-interest, road networks). We...
Contexts and social network information have been proven to be valuable for building accurate recommender system. However, the best of our knowledge, no existing works systematically combine diverse types such further improve recommendation quality. In this paper, we propose SoCo, a novel context-aware system incorporating elaborately processed information. We handle contextual by applying random decision trees partition original user-item-rating matrix that ratings with similar contexts are...
Demand response (DR) has been known to play an important role in the electricity sector balance supply and demand. To this end, DR baseline is a key factor successful program since it influences incentive allocation mechanism customer participation. Previous studies have investigated accuracy bias for large, industrial commercial customers. However, analysis of performance residential customers received less attention. In paper, we analyze baselines Our goes beyond by understanding impact on...
Time series forecasting for streaming data plays an important role in many real applications, ranging from IoT systems, cyber-networks, to industrial systems and healthcare. However the is often complicated with anomalies change points, which can lead learned models deviating underlying patterns of time series, especially context online learning mode. In this paper we present adaptive gradient method recurrent neural networks (RNN) forecast presence points. We explore local features...
Privacy policies are the primary channel through which companies inform users about their data collection and sharing practices. These often long difficult to comprehend. Short notices based on information extracted from privacy have been shown be useful but face a significant scalability hurdle, given number of evolution over time. Companies, users, researchers, regulators still lack usable scalable tools cope with breadth depth policies. To address these hurdles, we propose an automated...
Sensor networks are increasingly being deployed in the environment for many different purposes. The observations that they produce made available with heterogeneous schemas, vocabularies and data formats, making it difficult to share reuse this data, other purposes than those which were originally set up. authors propose an ontology-based approach providing access query capabilities streaming sources, allowing users express their needs at a conceptual level, independent of implementation...
The recent development of smart meters has allowed the analysis household electricity consumption in real time. Predicting at such very low scales should help to increase efficiency distribution networks and energy pricing. However, this is by no means a trivial task since household-level much more irregular than transmission or levels. In work, we address problem improving forecasting using statistical relations between series. This done both district (hundreds houses), various machine...
The trend of time series characterizes the intermediate upward and downward behaviour series. Learning forecasting in data play an important role many real applications, ranging from resource allocation centers, load schedule smart grid, so on. Inspired by recent successes neural networks, this paper we propose TreNet, a novel end-to-end hybrid network to learn local global contextual features for predicting TreNet leverages convolutional networks (CNNs) extract salient raw Meanwhile,...