Two postdoctoral researchers in the area of Information Technology, specialization Data Mining
Both positions are for 2 years, with application deadline on 30th of November 2016 and starting date being as soon as possible.
The first position is funded by two projects, BIDAF Big Data Analytics Framework for a Smart Society (a distributed research environment of Halmstad University, SICS Swedish ICT, Högskolan i Skövde) and ARISE (collaboration between Halmstad University and Volvo Technology).
The second position is primarily funded by KK-Foundation project SeMI (Self-Monitoring for Innovation: Meta‑framework for group-based self-monitoring), which is a “Synergy” project coordinated by Halmstad University with Alfa Laval, EasyServ, HEM, HMS, Sydpumpen and Öresundskraft as industrial partners.
CAISR is a dynamic research environment with the research focus on aware intelligent systems, i.e. systems that are human aware, situation aware and to some extent "self-aware". The subject expertise in the department is in machine learning, data mining, signal analysis and mechatronics. Mining of data streams and medical records, big data analytics, self-monitoring, deviation detection are areas of particular interest of the department. The department has a very good track record of doing research in close collaboration with the Swedish industry and public sector.
The core research question of the SeMI project is “How to construct self‑monitoring systems that use joint-human machine learning to adapt to specific domains, by taking advantage of groups of peers, and ubiquitous streams of data?” In order to answer this question we propose to develop a general meta-framework that can, based on domain-specific input, create tools for group-based self-monitoring for that particular domain. This meta-framework should be capable of learning from streams of data and detecting deviations in an unsupervised fashion, but interactively exploit available expert knowledge in a joint human-machine learning fashion. The unprecedented amount of data accessible today allows ML to focus on more descriptive and explanatory analysis. Users no longer pose well-formulated, concrete questions, but instead require the system to be capable of highlighting interesting aspects such as deviations, anomalies, relations and co-occurrences. It is almost effortless to generate data, while the cost of analysing it does not change. We will support continuous learning model, where the training and usage are not easily separated, and the system improves its performance all the time, taking advantage of new data as it arrives.
The BIDAF project aims to significantly further the research within massive data analysis, by means of statistical machine learning, in response to the increasing demand of retrieving value from data in all of society. Our research focuses on scalable algorithms that can leverage the distributed framework for efficient mining of knowledge from transient data streams. In particular, we aim to move from algorithms designed to exploit limited amounts of data for as much knowledge as possible towards algorithms designed to process large amounts of data efficiently, build models that are constrained in size, and provide end users with easy to understand and traceable results.
The ARISE project aims to develop algorithms for early detection and analysis of vehicle quality issues, integrating multiple available data sources. New telematics solutions allow monitoring trucks in operation, combining on-board data with existing in-office knowledge such as warranty claims, technical reports and expert knowledge. We will provide quality analysts with data mining and machine learning methods capable of extracting patterns and finding trends in these diverse data sources.
The main way of extracting value from data is to capture the interesting aspects of it using a suitable model. The model is then used for detecting anomalies and trends, analysing key values, or making predictions. In the big data setting, however, one can create not one, but many useful models, focusing on different aspects of the data. We will develop new algorithms for building such sets of models and for ensuring sufficient diversity among them, as well as ways to combine them in flexible ways, for example into hierarchical structures of concepts and sub-concepts, or along time axis to distinguish permanent and time-limited patterns.
The unprecedented amount of data accessible today allows ML to focus on more descriptive and explanatory analysis. Users no longer pose well-formulated, concrete questions, but instead require the system to be capable of highlighting interesting aspects such as deviations, anomalies, relations and co-occurrences. It is almost effortless to generate data, while the cost of analysing it does not change. We will support continuous learning model, where the training and usage is not easily separated, and the system improves its performance all the time, taking advantage of new data as it arrives.
An important aspect of the position is to find connections to other projects within CAISR and on identifying common problems and finding solutions applicable across multiple domains.
The selected candidate will become part of the very dynamic and international research environment at the Center for Applied Intelligent Systems Research (CAISR), a part of Halmstad Embedded and Intelligent Systems Research (EIS) at the School of Information Technology. For more information please see: http://caisr.hh.se/
The post as postdoctoral researcher is a qualifying appointment with the purpose to give the employee a possibility to develop its independence as researcher and to obtain merits that can lead to a competence for another post with higher eligibility requirements. As a postdoctoral researcher you are expected to be active in the research done within the research environments CAISR and EIS. The teaching load will be at most 20% of the time working hours. Furthermore, we expect you to take an active part in the continued development of the research environment and that you will take part in applying for research funding from various financiers, both in Sweden and abroad.
The position is intended for someone with a recent PhD degree in Information Technology, Computer Science, Computer Engineering, or closely related fields. The research track record should demonstrate excellence in research areas such as machine learning, data mining, and signal processing. Strength in computer programming and/or applied mathematics is very welcome.
Salary is to be settled by negotiation. The application should include a statement of the salary level required by the candidate.
Applications should be sent via Halmstad University's recruitment system MyNetwork. The last day to apply for the position is 2016-11-30. If you are interested in both positions, make sure you submit two applications (they can be the same).
The application package shall consist of:
- a cover letter stating the purpose of the application and a brief statement of why you believe that your goals are well-matched with the goals of this position, together with a description of future research plans
- an attested CV that includes at least
- a list of previous degrees, dates, and institution, transcripts for higher-education studies until most recent available
- a complete list of publications and a description of previous research and other work experience and links to online copies of the most important publications
- contact information for at least three references.