Cognitive Computing M.Sc.
Capitalize on data – or be left behind – is the maxim that drives modern IT technologies like no other. However, data-driven technologies require broad expertise in interpreting, learning from, and interacting with data that is outside the scope of most computer science and software engineering programs.
Artificial Intelligence and Machine Learning; Deep Neural Networks and Neuromorphic Computing; Statistics and Data Science; Computational Linguistics and Computer Vision share a common methodological core and problem-solving point-of-view that is summarized under the umbrella of Cognitive Computing: Engineering algorithms that improve as the availability of data increases.
The further development of Cognitive Computing technologies will be a game changer in modern society and high-tech industries. New forms of interacting with machines and services, augmented and mixed reality in the automobile sector and production processes, adaptive artificial systems in households, wearable smart devices for monitoring and optimizing the self as well as simplifying life, automation devices in administrative processes, consulting tasks, education tasks, and governmental services are just some examples where cognitive computing technologies will influence society and industry.
The Institute of Cognitive Science at the University of Osnabrück is the place to be to get a well-rounded education in cognitive computing technologies. Founded in 2001 as the first Cognitive Science Institute with the first Bachelor’s program of its kind in Germany starting in 1998, we have acquired extensive, interdisciplinary research and teaching experience. Our alumni have spread far and wide throughout academia and industry, and we have worked hard to build a cooperative network with international and local industry partners around cognitive technologies.
In our work as an academic Institute, we also emphasize the impact and influence Cognitive Computing and Artificial Intelligence have on society by working towards transparent and understandable algorithms and by researching how artificially intelligent systems can interact with and be useful to people. This point of view is especially important to take when developing technology that business can be built on in the long term and accentuates the unique strength of the Institute as a place where students learn what it takes to build cognitive computing systems for tomorrow.
› Visit Our Website
In order to enable students to learn on the job and companies to establish the skill set they need to confront technological change in their workforce, we offer a part-time master degree that can be completed successfully in 2 years during which participants remain employed and working part-time in companies.
We have distilled a study program from our diverse and established set of courses that will help students catch up on what they need to enable industry transforming technology. Additionally, industry partners will give hands on courses to enable students to hone their skills on real-world problems.
To get a broader overview over the curriculum of a full Master, take a look at the following example schedules that reflect what students may have studied at the IKW in the past. The study project and master thesis are particularly important in the program, so short example abstracts are also included!
* Names and photos are fictional
After studying Business & Computer Science, Sharon is looking to work at the interface between development and sales. She is looking for the ability to quickly conceptualize problems in terms of AI algorithms, but also the skills to lead development efforts in that domain. Therefore, she specialized in Neuroinformatics/Robotics and Artificial Intelligence.
Energy prices are rising, as is the overall awareness for environmental sustainability. However, significant energy saving potentials in home heating are left unused by today's heating control systems. "Smart", adaptive techniques can tap into these saving potentials. Our project aims to develop an adaptive energy control software (e.g. for heating and electric devices) with an accompanying smartphone user interface, in order to exploit both these saving potentials. We first try to predict the user's thermal demands, thus reducing the need for us-er interaction. This prediction allows us to prearrange future energy consuming periods in an energy-efficient way. The con-troller is intended to be used via a smartphone user interface,designed to steer users towards energy-saving settings. Fur-ther goals of this project are the connection of the user interface and the adaptive controller to a back-end server, in order to arrive at a testable prototype.
This thesis is focused on the standardization and usability optimization of graphical user interfaces at a local company. The domain of the new graphical user interface standard is initially open: It should at least cover the software running on testing facilities ("Prüfstände"), but it may be applicable to further graphical interfaces as well and cover as many software systems as possible. Ideally, it will continue to be developed and optimized after Sharon has finished her work and within the next few years be implemented in as many graphical user interfaces as possible: The project is the starting point for better usability and unified GUIs at the company. First steps include an analysis of the current situation, the conceptualization of a standard, and planning of a first technical implementation of the standard.
* Names and photos are fictional
Steven has an undergraduate degree in neuroscience and focuses his studies on the interface between machine learning, artificial intelligence and knowledge representations. His background in studying the brain and complex systems helps him to focus on cognitive technologies, taking courses in fundamental neuroscience, neuronal learning and cognitive AI-technologies. He hopes to put his knowledge to good use by working on new concepts of mobility and transport, where he will contribute as an App developer.
In this study project, we will develop Souma, an open source, decentralised, and encrypted social network that emphasises the self-determined development of individual and group identity. Social media has the potential to connect people around shared interests, enabling them to exchange data ( e.g. Twitter, Pinterest, Flickr, Youtube), develop collective knowledge (e.g. Wikipedia, Dropbox) and organise joint action (e.g. Kickstarter). But actually helping like-minded people to find each other and to coordinate their actions to achieve common goals has not yet been at focus.
We will integrate insights from a cognitive science perspective - on self and identity, online and offline communication, self-management of groups, machine learning of a filter bubble, self-observation etc. - to develop the concept and functionalities of Souma.
This thesis develops the requirements that result from the cognitive concept of Human Situation Awareness (SA) in the context of driving. This concept is analyzed on its constituents and the different cognitive processes that influence and build SA in the human brain. The analysis of SA results in several general SA-requirements for automotive systems design. The second main part of the thesis investigates how the developed SA-requirements can be applied to system design in a structure and complete manner in order to result in a system which is aware of the current situation - System Context Awareness (CA).
* Names and photos are fictional
Dorothea studied Physics, but has discovered a fascination for natural language communications with machines. To acquire the skills required to develop such systems, she is studying Cognitive Science and specializes in Neuroinformatics/Robotics and Computerlinguistics. In her career, she will engineer solutions for Assisted Living, and build commercial applications of robots in product consulting and recommendation systems.
This study project will use conceptors to model complex communication patterns and vocalisations of birds. The data will come from the Bird Db database that contains tens of thousands of birds songs over thousands of hours. Conceptors will be trained to first identify the chirps and then a slower hierarchical model will be used to model bird grammar.
This thesis concerns the world knowledge required for computational natural language understanding. In particular, the present work focuses on the integration of different types of knowledge in one modular knowledge base that is itself used in an inference-based natural language understanding pipeline. First, an introduction to different research areas relevant for natural language understanding is given. In the main body of this work, we propose an integrative knowledge base combining lexical, semantic, ontological, and distributional knowledge in a modular way. We then design a reasoning procedure that is able to make use of the developed large scale knowledge base. Finally, an evaluation of the proposed knowledge base and the reasoning pipeline is presented. For evaluation, we use three different natural language understanding tasks: recognizing textual entailment, semantic role labeling, and interpretation of noun dependencies.
Applications for the Winter Term are processed via the University.
› Send your application