Informatik Masterarbeit / Bachelorarbeit / Praktikum Wir haben eine Reihe von Arbeiten im Rahmen des Studiums für sehr gute Studenten der Informatik zu vergeben. Die unten aufgeführten Arbeiten werden am Deutschen Forschungszentrum für Künstliche Intelligenz (DFKI, Arbeitsgruppe Wissensmanagement, http://www.dfki.de) in Kaiserslautern angeboten. Das DFKI ist mit der Universität Kaiserslautern affiliiert. OVERVIEW We have a number of theses to offer for excellent students in computer science, all of them are centered around the project Text 2.0. (Für eine deutsche Version der Studienangebote siehe: http://text20.net/node/11) Text 2.0 is text, which can react to reading in realtime, with the help of an eye tracker. This is a frontier field of knowledge management, which got a lot of media attention recently and it is likewise closely connected to research in user interfaces, information retrieval and various fields of artificial intelligence. Existing use cases are, so far: 'hollywood-books', automatic translations and explanations, speech interactions, and many more. Beyond that, we showed that Text 2.0 offers benefits in the areas of search refinement and information management. More detailed information can be found on the page http://text20.net it also contains several videos and images. The specific topics we have to offer are: TASKS 1) iPhone Eye Tracking (http://data.text20.net/documentation/openpositions/reading.iphone.de.pdf) Currently Text 2.0 only works on large computer screens that have an attached eye tracker. Within this task you should a) try to attach an eye tracker to an iPhone (or Android device) to see how people read on these devices and b) develop algorithms based on your findings to offer Text 2.0 related functionality without a tracking device. 2) All Things Text 2.0 (http://data.text20.net/documentation/openpositions/text20.de.pdf) Text 2.0 is a very new field, with lot of potential not yet uncovered. The main task of this work is to discover new interaction paradigms Text 2.0 has to offer and evaluate them for their usefulness. This includes, but is not limited to, finding better ways to provide additional information while reading, and real time modifications of textual information based on the current user situation. 3) WebCam Eye Tracking for Text 2.0 (http://data.text20.net/documentation/openpositions/reading.webcam.de.pdf) Even though we assume that eye tracking will become common at one day, it takes some time until this moment is reached. In the meantime we try to rely on web cameras to approximate accurate eye tracking information. The task of this project is to research, to what extend web cameras are suitable to provide a replacement to expensive eye tracking equipment and implement a prototype based on your findings. 4) Emotional Text Tagging (http://data.text20.net/documentation/openpositions/reading.emotions.de.pdf) Reading text is often emotional. What if you could ask your computer to search all the exiting, boring, interesting or funny passages you have read recently. The main tasks of this project are a) use a web camera to implement an emotion detection and b) combine your output with real time eye tracking information to tag text with emotions. 5) Attentive eCommerce (http://data.text20.net/documentation/openpositions/reading.ecommerce.de.pdf ) We assume that Text 2.0 features will become common in society. How will this change eCommerce? The tasks of this project are a) to use real time eye tracking date to place and optimize the placements advertisements and b) to apply your findings to real time optimization scenarios where no appropriate eye tracking device is available. 6) "Wildcard" As we said, Text 2.0 is new and there is much to discover. Any other topic not mentioned here is highly welcome and we would like to discuss your ideas with you. REQUIREMENTS The requirements vary according to the specific task. - Java is required for all of them. However, in case you have a strong background in another language, learning Java while doing it should be no problem for you. - software engineering skills are required, we expect clean, documented code and structured solutions - JavaScript, HTML and CSS are helpful - knowledge in machine learning, data mining or pattern recognition is helpful - knowledge of design / user interfaces can be helpful for the UI related tasks - a scientific methodology is helpful DURATION The minimum time for each task is 3 months. However, we strongly recommend a time of ~4 months for bachelor's theses and ~ 6 months for master's theses. Internship can be done as well. There are no defined start and end dates, however, the sooner the better. Space is limited. CONTACT INFO Our contact information is Ralf Biedert ralf.biedert AT dfki.de and Georg Buscher georg.buscher AT dfki.de Deutsches Forschungszentrum für Künstliche Intelligenz (DFKI GmbH) Forschungsbereich Wissensmanagement Trippstadter Str. 122 67663 Kaiserslautern