Deutsch Intern
Center for Artificial Intelligence and Data Science

AI Talks @JMU 2022

In the 2nd series of talks organised by CAIDAS, members and selected guests present their research with exciting talks on current activities and projects. The talks will take place every Friday in the middle of the month at 15:00 (st) starting from December 2021. The talks will be 45 minutes, followed by a 15-minute discussion.

Due to the pandemic contact restrictions, the series of talks are held as a zoom meeting:

upcoming talks:

past talks:

See below for details.

past series:

upcoming talks

postponed - date to be announced

From Data Art to Enterprise AI

In recent years, models and methods of artificial intelligence (AI) have attracted a great deal of attention among scholars, managers and even the broader public.
The hype surrounding AI is driven not only by a vibrant research community, but also by massive investments on the part of large companies as well as the funding of numerous startups by venture capitalists.
The widespread expectation is that AI-related competences will become a prerequisite for competitiveness in several industries and a source for future innovations along the entire value chain.
However, it is increasingly apparent that many AI initiatives in practice suffer from an implementation problem, with few prototypes crossing the chasm between proofs-of-concept and productive real-world deployments.
Against this background, the talk will provide an overview of economic benefits and challenges organizations face in implementing their AI projects as well as implications for research, teaching and management.

past talks

6 May

Prof. Dr. Radu Timofte

Trends in Restoration and Manipulation of Images and Videos

Computer Vision is the Artificial Intelligence field aiming at capturing, representing and interpreting information from image and video data.
This talk will provide an overview of the recent research conducted at Augmented Perception Group, ETH Zurich and Chair of Computer Vision, JMU Wurzburg, focusing on the new trends in restoration and manipulation of images and videos, as key components in Computer Vision and AI.
In particular, we will discuss topics such as restoration, learning the space of solutions, quality enhancement, domain translation and synthesis, efficiency and mobile AI, compression, correspondences and tracking. A special attention will be given to the complexity of the solutions and their potential for real-world applications.

18 March

Towards a better understanding of cell fate and function – extracting biological information from multidimensional molecular data

All our organs and tissues are intricately composed of a multitude of different cell types with specialized functions, all emerging from a single germ cell when life begins. The differentiation of cell types during development of an organism, but also throughout adult life, is a highly orchestrated process, coordinated across multiple molecular layers of a cell. This process is controlled by cell-intrinsic molecular interactions, but also by extrinsic signals each cell receives from its neighbors, involving thousands of different genes switched on at variable levels across a multitude cell types. During the last decade, revolutionary experimental technologies have been developed to quantify gene expression and other molecular readouts in thousands of individual cells within a single experiment. Understanding cellular differentiation and cellular function, which will ultimately help us to explain development of diseases and to discover new therapies, requires the analysis of such large-scale multidimensional datasets, and novel computational approaches involving machine learning and artificial intelligence are rapidly being developed. In this lecture, I will give an overview of the major questions and the current strategies and future ideas to address these questions.

14 January

Prof. Dr. Ingo Scholtes

What makes teams successful? Insights from Repository Mining, Network Science, and Empirical Software Engineering

The convergence of social and technical systems provides us with a wealth of log data that capture the structure and dynamics of social organizations. It is tempting to utilize these data to better understand how social systems evolve, how collaboration patterns in teams are related to their "success" or "failure", and how the position of individuals in social networks affects their performance, motivation, and productivity.

Focusing on the empirical study of collaborative software projects, in this talk I will show how massive repository data from publicly available online platforms can be used to better understand human and social aspects in software development. Addressing an ongoing debate about the influence of team size on developer productivity, I specifically argue how we can use high-resolution time-stamped data on code editing to automatically construct meaningful collaboration networks. I further show how we can creatively use network models to test hypotheses at the intersection of software engineering, organizational theory, and network science, and which pitfalls await us in the analysis of massive data from online systems.

10 December

Prof. Dr. Bernhard Sendhof

(Honda Research, TU Darmstadt)

AI in Engineering Design: From Tool to Partner

The role of AI in Engineering and particularly in Engineering Design has made significant progress in the last years. In the first part of my presentation, I will outline the CAE/AI enhanced approach to engineering design from an industrial perspective. This will include examples from design and topology optimization and concludes with some of the remaining challenges like robustness and many-objective optimization.

In the second part of my presentation, I will introduce approaches to go beyond the tool-based AI in the engineering design process chain and enable the AI methods to improve their performance over time. Experience-based Computation: learning to optimise is an EU Horizon 2020 project that addresses the issue on how optimization can be improved through learning just like the engineer becomes more and more experienced over time. I will look at one approach inspired from data mining and knowledge extraction and one from transfer learning and the advantage of multi-task optimization.

AI as a cooperative partner in the engineering design process will be the subject of the last part of my presentation. I will briefly introduce the general concept of cooperative intelligence and then outline some of the challenges in understanding the engineer for optimal support. Many if not most engineering design decisions are made in a team, therefore, it is necessary to go beyond the cooperative interaction between the engineer and AI as a partner, but to also study the effect that an AI system can have on the decision dynamics in a team.

The presentation will conclude with a summary and some additional issues that have to be addressed to evolve AI from a tool to a partner in Engineering and in Engineering Design.