Projects

Projects

The AIML group advances the scientific foundations of robust, responsible, and generative AI through fundamental and application-oriented research. Our work is implemented and validated through nationally and internationally funded projects in collaboration with academic, governmental, and industrial partners. It is supported by European and national funding agencies, including the European Union, the DFG, and the Volkswagen Foundation.

Completed Research Projects

EU projekt: MAMMOth - Multi-Attribute, Multimodal Bias Mitigation in AI Systems

Funding: EU as part of the call HORIZON-CL4-2021-HUMAN-01 (A HUMAN-CENTRED AND ETHICAL DEVELOPMENT OF DIGITAL AND INDUSTRIAL TECHNOLOGIES 2021)
Project duration: 36 months, from the end of 2022

The MAMMOth project focuses on reducing multiple discrimination in tabular, network, and multimodal data. With the help of its computer science and AI experts, MAMMOth aims to address the related scientific challenges by developing an innovative, fairness-oriented, data-driven AI foundation that provides the necessary tools and techniques for detecting and mitigating (multiple) discrimination and ensures the accountability of AI systems with respect to multiple protected attributes and for traditional tabular data as well as more complex network and visual data. The project aims to demonstrate the developed solutions through pilot projects in three relevant areas of interest: a) financial/credit applications, b) identity verification systems, and c) academic assessment.

 

More >

EU projekt: STELAR - Spatio-TEmporal Linked data tools for the AgRi-food data space

Funding: EU as part of the call for proposals HORIZON-CL4-2021-DATA-01-03 - Technologies for data management (IA)

Project duration: 36 months, from the end of 2022

STELAR will design, develop, evaluate and present an innovative Knowledge Lake Management System (KLMS) to support and facilitate a holistic approach to FAIR (Findable, Accessible, Interoperable, Reusable) and AI-enabled (high quality, reliably labeled) data. The system will be piloted in various real-world use cases in the agri-food sector, one of the nine data areas of strategic societal and economic importance identified in the European Data Strategy.

More >

DFG Projekt: Hephaestus - Machine learning methods for adaptive process planning of 5-axis milling

Funding: German Research Foundation (Deutsche Forschungsgemeinschaft - DFG)

Project duration: 15.04.2021 - 31.12.2023

The aim of the project is to research a framework for learning 5-axis compensation of form errors in milling processes. This is based on a process-parallel material removal simulation and sophisticated machine learning strategies. Furthermore, we want to investigate the possibility of knowledge transfer between different workpiece geometries, milling tools, and machine tools for improved process planning. To this end, we will establish a framework that supports the necessary functions for flexible and real-time filtering, fusion, and storage of data streams with different characteristics.

Subsequently, basic knowledge about the performance of different machine learning algorithms for building process knowledge is provided and suitable supervised learning methods are designed. Based on this knowledge, a method will be explored that automatically identifies novel process situations and decides whether a new model domain is required or whether existing knowledge can be transferred.

Finally, we plan to develop a compensation strategy for molding errors that combines an adjustment of the tool path using 5 axes of the machine with a local adjustment of the feed rate. Since production data is only available to the scientific community to a very limited extent, the experimental data sets and labeling will be made available online to the scientific community. This allows other research groups to reproduce our results and evaluate their own methods.

More >

DFG Projekt: SFB 1463 - Integrated Design and Operation Methodology for Offshore Megastructures

Funding: German Research Foundation (Deutsche Forschungsgemeinschaft - DFG)

Project duration: 23.08.2021 - 31.12.2024

With the help of the real-time capable simulation model of a digital twin, the CRC researches the design and operating conditions of offshore megastructures, whereby all life phases of a structure can be mapped. This enables reactive and automated adaptation of the structure and its operation to current and future conditions.

We participate in the project B01 "Integrated design process for offshore structures" with the aim of developing a prediction model for predicting the design quality over the entire life cycle of offshore structures, combining the experience and intuition of the engineer and the expert through appropriate ML methods.

More >

EU projekt: ITN NoBIAS - Artificial Intelligence without Bias

Funding: EU
Project duration: 01.01.2020 - 31.12.2024

The main purpose of NoBIAS is to research and develop novel methods for AI-based decision-making without bias. NoBIAS will deliver a group of 15 researchers trained to identify biased and discriminatory AI decisions and provide solutions that fully utilize AI while ensuring compliance with legal and social norms.

The interdisciplinary NoBIAS consortium consists of eight organizations, including one non-academic one, spread across five European states, with leading expertise in artificial intelligence, law, and sociology. The network is complemented by ten associated non-academic partners from different application areas, including banking and healthcare.

More >

Volkswagen Foundation project: BIAS - Bias and Discrimination in Big Data and Algorithmic Processing

Funding: Volkswagen Foundation
Project duration: 01.12.2018 - 30.11.2022

BIAS is an interdisciplinary research initiative that brings together experts from philosophy, law and computer science and combines epistemological, ethical, legal and technical perspectives.

Our joint research question is: How can we ensure that big data analysis and algorithm-based decision-making are free from bias and non-discriminatory? To this end, we offer philosophical analyses of relevant concepts and principles, examine their application in relevant legal frameworks and develop technical solutions such as de-biasing strategies and discrimination detection methods.

Our shared conviction is that it is not just about the algorithms, but rather about the entire system of computerized predictions and human decisions that must be qualified as free of bias and non-discriminatory. Our solutions will therefore address the entire decision-making process rather than focusing on isolated components.

More >

LernMINT - Dealing with bias and discrimination in learning analytics models

Funding: Lower Saxony Ministry of Science and Culture
Project duration: 01.08.2020 - 31.07.2023

The aim of the LernMINT doctoral network is the interdisciplinary training of doctoral students to become experts in a sub-area of digital education, data-supported learning and teaching in STEM subjects. LernMINT has five thematic areas: (1) Learning analytics methods for STEM teaching, (2) Data-based support of learning potential, (3) Informal learning in schools and universities, (4) Transition from school to university, (5) Data protection and fairness of learning analytics methods.

More >

KISWIND

Funding: Bundesministerium für Wirtschaft und Energie (BMWi)
Project duration: 01.01.2022 - 31.12.2024

The overall objective of the research project is the development of methods for efficient and automated continuous monitoring of the support structure of wind turbines to detect damage events and anomalies using acoustic emission analysis and innovative evaluation methods from the field of artificial intelligence (AI).


More >

OSCAR - Opinion Stream Classification with Ensembles and Active leaRners

Funding: Deutsche Forschungsgemeinschaft
Project duration: 2017 – 2019

Many data accumulating in the Web reflect opinions on diverse subjects - products, institutions, events (e.g., elections) or topics (e.g., earth warming). Opinionated documents constitute a continuous stream; polarity learning on them delivers insights on the attitude of people towards each subject. Polarity learning algorithms must cope with classic Big Data characteristics: high volume and velocity of the arriving data, and volatility of the learned concepts, since subjects and attitudes of people toward certain subjects change over time. In OSCAR, we will develop classifiers that operate on an evolving feature space, adapt to changes in both vocabulary and data, and operate with limited class labels.

More >

Transalpine mobility and knowledge transfer

Funding: DFG FOR 1670
Role: Postdoctoral researcher (Eirini Ntoutsi)

The project aims at the establishment of an isotopic fingerprint for bioarchaeological finds, especially cremations, and its application to archaeological and cultural-historical problems of the Late Bronze Age until Roman Times. From a computer science persective, our focus is on the development of innovative methods that allow complete scientific analysis of project related data despite their complexity. We focus on data management and automated data analysis (similarity search, cluster analysis, outlier recognition) for the establishment of small-scaled isotopic fingerprints.

More >

GeoPKDD - Geographic Privacy-aware Knowledge Discovery and Delivery

FP6/IST project
Project duration: 2005 – 2009

GeoPKDD aims at developing theory, techniques and systems for knowledge discovery and delivery, based on new automated privacy-preserving methods for extracting user-consumable forms of knowledge from large amounts of raw data referenced in both space and time dimensions.

More >

Knowledge Discovery and Pattern Management - the PBMS approach

Funding: EPEAEK II / Heracletos Programme
Project duration: 2003 – 2005

The goal of this project is the efficient management of data mining patterns extracted from large databases, with emphasis on the pattern similarity assesment problem.

More >

PANDA - Patterns for Next Generation Database Systems

Funding: IST project
Project duration: 2001 – 2004

PANDA working group studies current state-of-the-art in pattern management and explores novel theoretical and practical aspects of a Pattern Base Management System (so-called, PBMS). PANDA's goal is the efficient and effective management of patterns; just as raw data are managed by traditional DBMS.