Posts by Collection

portfolio

publications

Success Factors in Business-Managed IT - A Case Study Analysis at a Large German Industrial Company

Published in SKILL 2018-Studierendenkonferenz Informatik, 2018

The terminology of Business-Managed IT refers to Shadow-IT systems which are operated overtly in the business units (BUs) and with the awareness from the IT department. They are a common phenomenon in corporations and usually emerge if the formal IT organization is unable to provide the BUs with solutions that meet their requirements. Because of this, Business-Managed IT is a highly significant area for an exploratory study, and both academia and practitioners can benefit from knowledge on how such systems can be successfully managed.

Recommended citation: Bachfischer, M. (2018). Success Factors in Business-Managed IT-A Case Study Analysis at a Large German Industrial Company. SKILL 2018-Studierendenkonferenz Informatik.

Offensive Comment Classification on German Language Microposts

Published in 14th Conference on Natural Language Processing KONVENS 2018, 2018

In this paper, we present two deep-learning based classifier systems for the identification of offensive comments in German Language microposts: A bidirectional LSTM model and a CNN model. Our objective is to compare the performance of these two systems with a traditional, machine-learning based SVM classifier and to evaluate our approach on Task 1 (binary classification) of the GermEval 2018 shared task.

Recommended citation: Bachfischer, M., Akujuobi, U., & Zhang, X. (2018). KAUSTmine-Offensive Comment Classification on German Language Microposts. In 14th Conference on Natural Language Processing KONVENS 2018.

M.Sc. Thesis: Adversarial Workload Matters - Executing a Large-Scale Poisoning Attack against Learned Index Structures

Published in M.Sc. Thesis - The University of Melbourne, 2021

Databases rely on indexes to quickly locate and retrieve data that is stored on disks. While traditional database indexes use tree data structures such as B+ Trees to find the position of a given query key in the index, a learned index structure considers this problem as a prediction task and uses a machine learning model to “predict” the position of the query key. This novel approach of implementing database indexes has inspired a surge of recent research aimed at studying the effectiveness of learned index structures. However, while the main advantage of learned index structures is their ability to adjust to the data via their underlying ML model, this also carries the risk of exploitation by a malicious adversary.

Recommended citation: Bachfischer, M. (2021). Adversarial Workload Matters - Executing a Large-Scale Poisoning Attack against Learned Index Structures

Testing the Robustness of Learned Index Structures

Published in https://arxiv.org/abs/2207.11575, 2022

While early empirical evidence has supported the case for learned index structures as having favourable average-case performance, little is known about their worst-case performance. By contrast, classical structures are known to achieve optimal worst-case behaviour.

Recommended citation: Bachfischer, M., Borovica-Gajic, R., & Rubinstein, B. I. P. (2022). Testing the Robustness of Learned Index Structures

Matthias Bachfischer