Gabi Dreo Rodosek

Sinclair Schneider, Florian Steuber, Joao A. G. Schneider et al.

With the rise of generative pre-trained transformer models such as GPT-3, GPT-NeoX, or OPT, distinguishing human-generated texts from machine-generated ones has become important. We refined five separate language models to generate synthetic tweets, uncovering that shallow learning classification algorithms, like Naive Bayes, achieve detection accuracy between 0.6 and 0.8. Shallow learning classifiers differ from human-based detection, especially when using higher temperature values during text generation, resulting in a lower detection rate. Humans prioritize linguistic acceptability, which tends to be higher at lower temperature values. In contrast, transformer-based classifiers have an accuracy of 0.9 and above. We found that using a reinforcement learning approach to refine our generative models can successfully evade BERT-based classifiers with a detection accuracy of 0.15 or less.

Sinclair Schneider, Florian Steuber, Joao A. G. Schneider et al.

The increasing popularity of large language models has not only led to widespread use but has also brought various risks, including the potential for systematically spreading fake news. Consequently, the development of classification systems such as DetectGPT has become vital. These detectors are vulnerable to evasion techniques, as demonstrated in an experimental series: Systematic changes of the generative models' temperature proofed shallow learning-detectors to be the least reliable. Fine-tuning the generative model via reinforcement learning circumvented BERT-based-detectors. Finally, rephrasing led to a >90\% evasion of zero-shot-detectors like DetectGPT, although texts stayed highly similar to the original. A comparison with existing work highlights the better performance of the presented methods. Possible implications for society and further research are discussed.

Raphael Labaca-Castro, Luis Muñoz-González, Feargus Pendlebury et al.

Machine learning classifiers are vulnerable to adversarial examples -- input-specific perturbations that manipulate models' output. Universal Adversarial Perturbations (UAPs), which identify noisy patterns that generalize across the input space, allow the attacker to greatly scale up the generation of such examples. Although UAPs have been explored in application domains beyond computer vision, little is known about their properties and implications in the specific context of realizable attacks, such as malware, where attackers must satisfy challenging problem-space constraints. In this paper we explore the challenges and strengths of UAPs in the context of malware classification. We generate sequences of problem-space transformations that induce UAPs in the corresponding feature-space embedding and evaluate their effectiveness across different malware domains. Additionally, we propose adversarial training-based mitigations using knowledge derived from the problem-space transformations, and compare against alternative feature-space defenses. Our experiments limit the effectiveness of a white box Android evasion attack to ~20% at the cost of ~3% TPR at 1% FPR. We additionally show how our method can be adapted to more restrictive domains such as Windows malware. We observe that while adversarial training in the feature space must deal with large and often unconstrained regions, UAPs in the problem space identify specific vulnerabilities that allow us to harden a classifier more effectively, shifting the challenges and associated cost of identifying new universal adversarial transformations back to the attacker.

Peter Hillmann, Frank Tietze, Gabi Dreo Rodosek

The identification of the exact path that packets are routed on in the network is quite a challenge. This paper presents a novel, efficient traceback strategy named Tracemax in context of a defense system against distributed denial of service (DDoS) attacks. A single packet can be directly traced over many more hops than the current existing techniques allow. In combination with a defense system it differentiates between multiple connections. It aims to letting non-malicious connections pass while bad ones get thwarted. The novel concept allows detailed analyses of the traffic and the transmission path through the network. The strategy can effectively reduce the effect of common bandwidth and resource consumption attacks, foster early warning and prevention as well as higher the availability of the network services for the wanted customers.

Peter Hillmann, Tobias Uhlig, Gabi Dreo Rodosek et al.

Complex scheduling problems require a large amount computation power and innovative solution methods. The objective of this paper is the conception and implementation of a multi-agent system that is applicable in various problem domains. Independent specialized agents handle small tasks, to reach a superordinate target. Effective coordination is therefore required to achieve productive cooperation. Role models and distributed artificial intelligence are employed to tackle the resulting challenges. We simulate a NP-hard scheduling problem to demonstrate the validity of our approach. In addition to the general agent based framework we propose new simulation-based optimization heuristics to given scheduling problems. Two of the described optimization algorithms are implemented using agents. This paper highlights the advantages of the agent-based approach, like the reduction in layout complexity, improved control of complicated systems, and extendability.