|Postdoctoral Associate – Pattern Recognition
For this position, the researcher will work on developing models/algorithms for real time traffic prediction that particularly take advantage of contextual data (e.g. data obtained from web-mining, news feeds, weather information) that can be relevant in terms of mobility phenomena. A simple case is that of a special events, such as a sports game, or a music concert. While these may have significant impact to the transport system, the traditional sensors (e.g. car counters, GPS probes, etc.) are insufficient to perceive this impact ahead in time. On the other hand, plenty of these events are advertised long time in advance in the web.
This general goal involves a vast list of research challenges at the level of machine learning, information retrieval, information extraction and transport engineering. Specifically, the job scope is as follow:
- Typical data analysis process (cleaning, preparation, descriptive statistics, exploration);
- Pattern recognition modeling (advancing current state of the art);
- Performance assessment (complexity analysis, field testing);
- Advance research in any of these areas or related.
The ideal person will have strong background in pattern recognition and computer science, but with a multidisciplinary spirit, namely understanding or willing to learn about urban mobility, transportation and planning. S/he will necessarily have to face statistical and mathematical challenges that are hard and somewhat rare in computing courses, so a good background in these topics is sought.
Candidates should meet the following requirements:
- PhD in Computer Science or any other field that guarantees background for above description (e.g. physics, electrical engineering, mathematics, transportation);
- Experience in one or more of the following sub-topics: Bayesian machine learning framework, graphical models, statistical learning;
- Good communication skills.