 |
The Fourth International Conference on Ambient Computing, Applications, Services and Technologies AMBIENT 2014 August 24 - 28, 2014 - Rome, Italy |
| Tutorials |
T1. Computing Techniques for Parallel and Distributed Systems with an Application to Data Compression
Prof. Dr. Sergio De Agostino, Sapienza University of Rome, Italy
T2. Applying In-Memory Technology to Genome Data Analysis
Cindy Fähnrich, Hasso Plattner Institute, Germany
T3. Semantic-Web-based Mobile Knowledge Management
Prof. Dr. Rachid Benlamri, Lakehead University, Canada
Detailed Description
T1. Computing Techniques for Parallel and Distributed Systems with an Application to Data Compression
Prof. Dr. Sergio De Agostino, Sapienza University of Rome, Italy
The computing techniques involved in the design of parallel and distributed algorithms strictly relate to the computational model on which the parallel or distributed system is based. The efficiency of a technique designed for a specific model can consistently deteriorates when applied to a different system. This is particularly evident when a technique designed for a shared memory parallel random access machine (PRAM) is implemented on a distributed system. Indeed, when the system is scaled up the communication cost is a bottleneck to linear speed-up. So, we need to limit the interprocessor communication either involving more local computation or bounding the number of global computation steps in order to obtain a practical algorithm. Local computation might cause a lack of robustness when scalability properties are required.
On the other hand, scalability and robustness are generally guaranteed if bounding the number of global computation steps is possible for a specific problem, Parallel prefix and list ranking are the two fundamental computing techniques for the design of a parallel algorithm running on a random access shared memory machine with a linear input data structure. A more general pointer jumping technique applies to non-linear structures and the so-called Euler tour technique reduces it to list ranking in order to avoid memory access conflicts. Such computing techniques can perform well on a distributed system when the number of iterations is very limited and we will describe them in the first part of the tutorial. A second part of the tutorial will concern the application of these techniques to lossless file compression.
The most popular compressors are based on Lempel-Ziv coding methods. Zip compressors employ the sliding window method, while other applications use the so-called LZW compressor. Zipping and unzipping files is parallelizable in theory by means of the above mentioned computing techniques. However, the number of global computation steps is not bounded by a constant and the local computation approach is more advantageous on a distributed system. LZW compression is less effective but faster than the zipping applications. Differently from the Zip compressors,, the LZW encoder/decoder .presents an asymmetry with respect to global parallel computation since the encoder is not parallelizable while the decoder has a very efficient parallelization.
We will conclude the tutorial showing that the number of iterations of the LZW parallel decoder is less than ten units, This makes LZW more attractive than Zip in those cases (which are the most common in practice) where compression is performed only once or very rarely while the frequent reading of raw data needs fast decompression.T2. Applying In-Memory Technology to Genome Data Analysis
Cindy Fähnrich, Hasso Plattner Institute, Germany
The objective of precision medicine is to identify the best treatment decision for a patient’s disease based on all of her/his individual specifics, e.g. genetic dispositions or family anamnesis. Especially for treatment of cancer disease, identifying individual dispositions of a patient supports clinicians in finding individualized therapies while keeping side effects at a minimum.
However, identification of genetic dispositions, the connection to a concrete disease, and the prediction of the effect of pharmaceuticals is still a time-consuming and very complex process, which requires detailed analysis of big medical data, e.g. to provide evidence-based treatment alternatives.
In this tutorial, we demonstrate how in-memory technology enables real-time analysis of big medical data to support researchers and clinicians in finding individual treatments. Based on a selected patient case, we share insights on how real-time data analyses improve clinical decisions taking. Following the steps of a clinician, we provide hands-on experience on how in-memory technology improves clinical routine.
We share details about how:
- Genome data analysis pipelines are directly integrated into our in-memory computing platform,
- In-memory database functionality enables real-time data analysis, such as patient cohort analysis and patient clustering,
- Text-mining and text analysis features support the processing of unstructured medical documents, such as doctor letters,
- Personal algorithms and software tools can be integrated in our platform to adapt it to your personal needs, and
T3. Semantic-Web-based Mobile Knowledge Management
Prof. Dr. Rachid Benlamri, Lakehead University, Canada
The ability to grasp the exact knowledge required to perform specific task, anywhere and just-in-time, is a key requirement for organizations to remain competitive in the new knowledge society. Therefore, the issues of “mobility” and “Knowledge Management” have recently received much attention among the research community. The interest in these issues is often motivated by the fact that work in many organizations is “knowledge intensive” and “mobile”. Activities involving for instance e-healthcare; mobile-learning, environmental monitoring; and remote tracking for safety and homeland security require special infrastructure for mobile knowledge management tailored to the context of their users and surrounding environment. Advances in wireless communications, semantic Web technologies, sensor networks, ubiquitous and mobile computing offer unique chances to build such an infrastructure; thus, enabling higher levels of interoperability and automation across broad range of context-aware mobile services. Such services can dynamically recognize and adapt to the context in which their users operate (e.g. location, activity, surrounding environment, social context, etc.). In this talk, we present the recent progress in this field with a focus on context awareness issues in mobile services. In particular, we show the use of ontologies to model context information in a service-oriented mobile environment. Case studies in m-learning and mobile e-healthcare will be presented.
