Johannes Schneider

Johannes Schneider, Michail Vlachos

Interpreting a large number of neurons in deep learning is difficult. Our proposed `CLAssifier-DECoder' architecture (ClaDec) facilitates the understanding of the output of an arbitrary layer of neurons or subsets thereof. It uses a decoder that transforms the incomprehensible representation of the given neurons to a representation that is more similar to the domain a human is familiar with. In an image recognition problem, one can recognize what information (or concepts) a layer maintains by contrasting reconstructed images of ClaDec with those of a conventional auto-encoder(AE) serving as reference. An extension of ClaDec allows trading comprehensibility and fidelity. We evaluate our approach for image classification using convolutional neural networks. We show that reconstructed visualizations using encodings from a classifier capture more relevant classification information than conventional AEs. This holds although AEs contain more information on the original input. Our user study highlights that even non-experts can identify a diverse set of concepts contained in images that are relevant (or irrelevant) for the classifier. We also compare against saliency based methods that focus on pixel relevance rather than concepts. We show that ClaDec tends to highlight more relevant input areas to classification though outcomes depend on classifier architecture. Code is at \url{https://github.com/JohnTailor/ClaDec}

Alfred Sopi, Johannes Schneider, Jan vom Brocke

Non-fungible tokens(NFTs) are on the rise. They can represent artworks exhibited for marketing purposes on webpages of companies or online stores -- analogously to physical artworks. Lending of NFTs is an attractive form of passive income for owners but comes with risks (e.g., items are not returned) and costs for escrow agents. Similarly, renters have difficulties in anticipating the impact of artworks, e.g., how spectators of NFTs perceive them. To address these challenges, we introduce an NFT rental solution based on a pay-per-like pricing model using blockchain technology, i.e., smart contracts based on the Ethereum chain. We find that blockchain solutions enjoy many advantages also reported for other applications, but interestingly, we also observe dark sides of (large) blockchain fees. Blockchain solutions appear unfair to niche artists and potentially hamper cultural diversity. Furthermore, a trust-cost tradeoff arises to handle fraud caused by manipulation from parties outside the blockchain. All code for the solution is publicly available at: https://github.com/asopi/rental-project

Luca Longo, Mario Brcic, Federico Cabitza et al.

As systems based on opaque Artificial Intelligence (AI) continue to flourish in diverse real-world applications, understanding these black box models has become paramount. In response, Explainable AI (XAI) has emerged as a field of research with practical and ethical benefits across various domains. This paper not only highlights the advancements in XAI and its application in real-world scenarios but also addresses the ongoing challenges within XAI, emphasizing the need for broader perspectives and collaborative efforts. We bring together experts from diverse fields to identify open problems, striving to synchronize research agendas and accelerate XAI in practical applications. By fostering collaborative discussion and interdisciplinary cooperation, we aim to propel XAI forward, contributing to its continued success. Our goal is to put forward a comprehensive proposal for advancing XAI. To achieve this goal, we present a manifesto of 27 open problems categorized into nine categories. These challenges encapsulate the complexities and nuances of XAI and offer a road map for future research. For each problem, we provide promising research directions in the hope of harnessing the collective intelligence of interested stakeholders.

Giovanni Apruzzese, Pavel Laskov, Johannes Schneider

Machine Learning (ML) has become a valuable asset to solve many real-world tasks. For Network Intrusion Detection (NID), however, scientific advances in ML are still seen with skepticism by practitioners. This disconnection is due to the intrinsically limited scope of research papers, many of which primarily aim to demonstrate new methods ``outperforming'' prior work -- oftentimes overlooking the practical implications for deploying the proposed solutions in real systems. Unfortunately, the value of ML for NID depends on a plethora of factors, such as hardware, that are often neglected in scientific literature. This paper aims to reduce the practitioners' skepticism towards ML for NID by "changing" the evaluation methodology adopted in research. After elucidating which "factors" influence the operational deployment of ML in NID, we propose the notion of "pragmatic assessment", which enable practitioners to gauge the real value of ML methods for NID. Then, we show that the state-of-research hardly allows one to estimate the value of ML for NID. As a constructive step forward, we carry out a pragmatic assessment. We re-assess existing ML methods for NID, focusing on the classification of malicious network traffic, and consider: hundreds of configuration settings; diverse adversarial scenarios; and four hardware platforms. Our large and reproducible evaluations enable estimating the quality of ML for NID. We also validate our claims through a user-study with security practitioners.

Johannes Schneider

Pre-trained language models have led to a new state-of-the-art in many NLP tasks. However, for topic modeling, statistical generative models such as LDA are still prevalent, which do not easily allow incorporating contextual word vectors. They might yield topics that do not align well with human judgment. In this work, we propose a novel topic modeling and inference algorithm. We suggest a bag of sentences (BoS) approach using sentences as the unit of analysis. We leverage pre-trained sentence embeddings by combining generative process models and clustering. We derive a fast inference algorithm based on expectation maximization, hard assignments, and an annealing process. The evaluation shows that our method yields state-of-the art results with relatively little computational demands. Our method is also more flexible compared to prior works leveraging word embeddings, since it provides the possibility to customize topic-document distributions using priors. Code and data is at \url{https://github.com/JohnTailor/BertSenClu}.

Johannes Schneider

Large language models (LLMs) are increasingly used for topic modeling, outperforming classical topic models such as LDA. Commonly, pre-trained LLM encoders such as BERT are used out-of-the-box despite the fact that fine-tuning is known to improve LLMs considerably. The challenge lies in obtaining a suitable labeled dataset for fine-tuning. In this paper, we build on the recent idea of using bags of sentences as the elementary unit for computing topics. Based on this idea, we derive an approach called FT-Topic to perform unsupervised fine-tuning, relying primarily on two steps for constructing a training dataset in an automatic fashion. First, a heuristic method identifies pairs of sentence groups that are assumed to belong either to the same topic or to different topics. Second, we remove sentence pairs that are likely labeled incorrectly. The resulting dataset is then used to fine-tune an encoder LLM, which can be leveraged by any topic modeling approach that uses embeddings. In this work, we demonstrate its effectiveness by deriving a novel state-of-the-art topic modeling method called SenClu. The method achieves fast inference through an expectation-maximization algorithm and hard assignments of sentence groups to a single topic, while allowing users to encode prior knowledge about the topic-document distribution. Code is available at https://github.com/JohnTailor/FT-Topic

Johannes Schneider, Steffi Haag, Leona Chandra Kruse

Large language models LLMs like ChatGPT have reached the 100 Mio user barrier in record time and might increasingly enter all areas of our life leading to a diverse set of interactions between those Artificial Intelligence models and humans. While many studies have discussed governance and regulations deductively from first-order principles, few studies provide an inductive, data-driven lens based on observing dialogues between humans and LLMs especially when it comes to non-collaborative, competitive situations that have the potential to pose a serious threat to people. In this work, we conduct a user study engaging over 40 individuals across all age groups in price negotiations with an LLM. We explore how people interact with an LLM, investigating differences in negotiation outcomes and strategies. Furthermore, we highlight shortcomings of LLMs with respect to their reasoning capabilities and, in turn, susceptiveness to prompt hacking, which intends to manipulate the LLM to make agreements that are against its instructions or beyond any rationality. We also show that the negotiated prices humans manage to achieve span a broad range, which points to a literacy gap in effectively interacting with LLMs.

Johannes Schneider, Giovanni Apruzzese

We propose to generate adversarial samples by modifying activations of upper layers encoding semantically meaningful concepts. The original sample is shifted towards a target sample, yielding an adversarial sample, by using the modified activations to reconstruct the original sample. A human might (and possibly should) notice differences between the original and the adversarial sample. Depending on the attacker-provided constraints, an adversarial sample can exhibit subtle differences or appear like a "forged" sample from another class. Our approach and goal are in stark contrast to common attacks involving perturbations of single pixels that are not recognizable by humans. Our approach is relevant in, e.g., multi-stage processing of inputs, where both humans and machines are involved in decision-making because invisible perturbations will not fool a human. Our evaluation focuses on deep neural networks. We also show the transferability of our adversarial examples among networks.

Johannes Schneider, Bernd Schenk, Christina Niklaus

Grading exams is an important, labor-intensive, subjective, repetitive, and frequently challenging task. The feasibility of autograding textual responses has greatly increased thanks to the availability of large language models (LLMs) such as ChatGPT and the substantial influx of data brought about by digitalization. However, entrusting AI models with decision-making roles raises ethical considerations, mainly stemming from potential biases and issues related to generating false information. Thus, in this manuscript, we provide an evaluation of a large language model for the purpose of autograding, while also highlighting how LLMs can support educators in validating their grading procedures. Our evaluation is targeted towards automatic short textual answers grading (ASAG), spanning various languages and examinations from two distinct courses. Our findings suggest that while "out-of-the-box" LLMs provide a valuable tool to provide a complementary perspective, their readiness for independent automated grading remains a work in progress, necessitating human oversight.

Johannes Schneider, Michalis Vlachos

Deep learning has made tremendous progress in the last decade. A key success factor is the large amount of architectures, layers, objectives, and optimization techniques. They include a myriad of variants related to attention, normalization, skip connections, transformers and self-supervised learning schemes -- to name a few. We provide a comprehensive overview of the most important, recent works in these areas to those who already have a basic understanding of deep learning. We hope that a holistic and unified treatment of influential, recent works helps researchers to form new connections between diverse areas of deep learning. We identify and discuss multiple patterns that summarize the key strategies for many of the successful innovations over the last decade as well as works that can be seen as rising stars. We also include a discussion on recent commercially built, closed-source models such as OpenAI's GPT-4 and Google's PaLM 2.

Johannes Schneider

Foundation models can be disruptive for future AI development by scaling up deep learning in terms of model size and training data's breadth and size. These models achieve state-of-the-art performance (often through further adaptation) on a variety of tasks in domains such as natural language processing and computer vision. Foundational models exhibit a novel {emergent behavior}: {In-context learning} enables users to provide a query and a few examples from which a model derives an answer without being trained on such queries. Additionally, {homogenization} of models might replace a myriad of task-specific models with fewer very large models controlled by few corporations leading to a shift in power and control over AI. This paper provides a short introduction to foundation models. It contributes by crafting a crisp distinction between foundation models and prior deep learning models, providing a history of machine learning leading to foundation models, elaborating more on socio-technical aspects, i.e., organizational issues and end-user interaction, and a discussion of future research.

Dario Fenoglio, Arianna Casanova, Francesco De Santis et al.

Concept-based models (CMs) enhance interpretability in deep learning by grounding predictions in human-understandable concepts. However, concept annotations are expensive to obtain and rarely available at scale within a single data source. Federated learning (FL) could alleviate this limitation by enabling cross-institutional training that leverages concept annotations distributed across multiple data owners. Yet, FL lacks interpretable modeling paradigms. Integrating CMs with FL is non-trivial: CMs assume a fixed concept space and a predefined model architecture, whereas real-world FL is heterogeneous and non-stationary, with institutions joining over time and bringing new supervision. In this work, we propose Federated Concept-based Models (F-CMs), a new methodology for deploying CMs in evolving FL settings. F-CMs aggregate concept-level information across institutions and efficiently adapt the model architecture in response to changes in the available concept supervision, while preserving institutional privacy. Empirically, F-CMs preserve the accuracy and intervention effectiveness of training settings with full concept supervision, while outperforming non-adaptive federated baselines. Notably, F-CMs enable interpretable inference on concepts not available to a given institution, a key novelty with respect to existing approaches.

Johannes Schneider, Arianna Casanova Flores, Anne-Catherine Kranz

This study explores real-world human interactions with large language models (LLMs) in diverse, unconstrained settings in contrast to most prior research focusing on ethically trimmed models like ChatGPT for specific tasks. We aim to understand the originator of toxicity. Our findings show that although LLMs are rightfully accused of providing toxic content, it is mostly demanded or at least provoked by humans who actively seek such content. Our manual analysis of hundreds of conversations judged as toxic by APIs commercial vendors, also raises questions with respect to current practices of what user requests are refused to answer. Furthermore, we conjecture based on multiple empirical indicators that humans exhibit a change of their mental model, switching from the mindset of interacting with a machine more towards interacting with a human.

Johannes Schneider

Transformers have become the de-facto standard model in artificial intelligence since 2017 despite numerous shortcomings ranging from energy inefficiency to hallucinations. Research has made a lot of progress in improving elements of transformers, and, more generally, deep learning manifesting in many proposals for architectures, layers, optimization objectives, and optimization techniques. For researchers it is difficult to keep track of such developments on a broader level. We provide a comprehensive overview of the many important, recent works in these areas to those who already have a basic understanding of deep learning. Our focus differs from other works, as we target specifically novel, alternative potentially disruptive approaches to transformers as well as successful ideas of recent deep learning. We hope that such a holistic and unified treatment of influential, recent works and novel ideas helps researchers to form new connections between diverse areas of deep learning. We identify and discuss multiple patterns that summarize the key strategies for successful innovations over the last decade as well as works that can be seen as rising stars. Especially, we discuss attempts on how to improve on transformers covering (partially) proven methods such as state space models but also including far-out ideas in deep learning that seem promising despite not achieving state-of-the-art results. We also cover a discussion on recent state-of-the-art models such as OpenAI's GPT series and Meta's LLama models and, Google's Gemini model family.

Haichao Wang, Alexander Okupnik, Yuxing Han et al.

Long-range human movement generation remains a central challenge in computer vision and graphics. Generating coherent transitions across semantically distinct motion domains remains largely unexplored. This capability is particularly important for applications such as dance choreography, where movements must fluidly transition across diverse stylistic and semantic motifs. We propose a simple and effective inference-time optimization framework inspired by diffusion-based stochastic optimal control. Specifically, a control-energy objective that explicitly regularizes the transition trajectories of a pretrained diffusion model. We show that optimizing this objective at inference time yields transitions with fidelity and temporal coherence. This is the first work to provide a general framework for controlled long-range human motion generation with explicit transition modeling.

Rene Pilz, Johannes Schneider

This paper explores the idea of using phonemes as a textual representation within a conventional multilingual simultaneous speech-to-speech translation pipeline, as opposed to the traditional reliance on text-based language representations. To investigate this, we trained an open-source sequence-to-sequence model on the WMT17 dataset in two formats: one using standard textual representation and the other employing phonemic representation. The performance of both approaches was assessed using the BLEU metric. Our findings shows that the phonemic approach provides comparable quality but offers several advantages, including lower resource requirements or better suitability for low-resource languages.

Johannes Schneider

The data distribution commonly evolves over time leading to problems such as concept drift that often decrease classifier performance. Current techniques are not adequate for this problem because they either require detailed knowledge of the transformation or are not suited for anticipating unseen domains but can only adapt to domains, where data samples are available. We seek to predict unseen data (and their labels) allowing us to tackle challenges s a non-constant data distribution in a proactive manner rather than detecting and reacting to already existing changes that might already have led to errors. To this end, we learn a domain transformer in an unsupervised manner that allows generating data of unseen domains. Our approach first matches independently learned latent representations of two given domains obtained from an auto-encoder using a Cycle-GAN. In turn, a transformation of the original samples can be learned that can be applied iteratively to extrapolate to unseen domains. Our evaluation of CNNs on image data confirms the usefulness of the approach. It also achieves very good results on the well-known problem of unsupervised domain adaption, where only labels but no samples have to be predicted. Code is available at https://github.com/JohnTailor/DoTra.

Johannes Schneider, Michalis Vlachos

We present a `CLAssifier-DECoder' architecture (\emph{ClaDec}) which facilitates the comprehension of the output of an arbitrary layer in a neural network (NN). It uses a decoder to transform the non-interpretable representation of the given layer to a representation that is more similar to the domain a human is familiar with. In an image recognition problem, one can recognize what information is represented by a layer by contrasting reconstructed images of \emph{ClaDec} with those of a conventional auto-encoder(AE) serving as reference. We also extend \emph{ClaDec} to allow the trade-off between human interpretability and fidelity. We evaluate our approach for image classification using Convolutional NNs. We show that reconstructed visualizations using encodings from a classifier capture more relevant information for classification than conventional AEs. Relevant code is available at \url{https://github.com/JohnTailor/ClaDec}

Johannes Schneider, Thomas Locher

Protecting source code against reverse engineering and theft is an important problem. The goal is to carry out computations using confidential algorithms on an untrusted party while ensuring confidentiality of algorithms. This problem has been addressed for Boolean circuits known as `circuit privacy'. Circuits corresponding to real-world programs are impractical. Well-known obfuscation techniques are highly practicable, but provide only limited security, e.g., no piracy protection. In this work, we modify source code yielding programs with adjustable performance and security guarantees ranging from indistinguishability obfuscators to (non-secure) ordinary obfuscation. The idea is to artificially generate `misleading' statements. Their results are combined with the outcome of a confidential statement using encrypted \emph{selector variables}. Thus, an attacker must `guess' the encrypted selector variables to disguise the confidential source code. We evaluated our method using more than ten programmers as well as pattern mining across open source code repositories to gain insights of (micro-)coding patterns that are relevant for generating misleading statements. The evaluation reveals that our approach is effective in that it successfully preserves source code confidentiality.

Johannes Schneider, Heinz Pampel

As Open Access continues to gain importance in science policy, understanding the proportion of Open Access publications relative to the total research output of research-performing organizations, individual countries, or even globally has become increasingly relevant. In response, dashboards are being developed to capture and communicate progress in this area. To provide an overview of these dashboards and their characteristics, an extensive survey was conducted, resulting in the identification of nearly 60 dashboards. To support a detailed and structured description, a dedicated metadata schema was developed, and the identified dashboards were systematically indexed accordingly. To foster community engagement and ensure ongoing development, a participatory process was launched, allowing interested stakeholders to contribute to the dataset. The dataset is particularly relevant for researchers in Library and Information Science (LIS) and Science and Technology Studies (STS), supporting both empirical analyses of Open Access and the methodological refinement of indicators and policy instruments in the context of Open Science.

Andrea Thomas Nava, Lijo Johny, Fabio Azzalini et al.

Data-Driven Product Development (DDPD) leverages data to learn the relationship between product design specifications and resulting properties. To discover improved designs, we train a neural network on past experiments and apply Projected Gradient Descent to identify optimal input features that maximize performance. Since many products require simultaneous optimization of multiple correlated properties, our framework employs joint neural networks to capture interdependencies among targets. Furthermore, we integrate uncertainty estimation via \emph{Conformalised Monte Carlo Dropout} (ConfMC), a novel method combining Nested Conformal Prediction with Monte Carlo dropout to provide model-agnostic, finite-sample coverage guarantees under data exchangeability. Extensive experiments on five real-world datasets show that our method matches state-of-the-art performance while offering adaptive, non-uniform prediction intervals and eliminating the need for retraining when adjusting coverage levels.

Johannes Schneider

Generative AI (GenAI) marked a shift from AI being able to recognize to AI being able to generate solutions for a wide variety of tasks. As the generated solutions and applications become increasingly more complex and multi-faceted, novel needs, objectives, and possibilities have emerged for explainability (XAI). In this work, we elaborate on why XAI has gained importance with the rise of GenAI and its challenges for explainability research. We also unveil novel and emerging desiderata that explanations should fulfill, covering aspects such as verifiability, interactivity, security, and cost. To this end, we focus on surveying existing works. Furthermore, we provide a taxonomy of relevant dimensions that allows us to better characterize existing XAI mechanisms and methods for GenAI. We discuss different avenues to ensure XAI, from training data to prompting. Our paper offers a short but concise technical background of GenAI for non-technical readers, focusing on text and images to better understand novel or adapted XAI techniques for GenAI. However, due to the vast array of works on GenAI, we decided to forego detailed aspects of XAI related to evaluation and usage of explanations. As such, the manuscript interests both technically oriented people and other disciplines, such as social scientists and information systems researchers. Our research roadmap provides more than ten directions for future investigation.

Johannes Schneider

Agentic Artificial Intelligence (AI) builds upon Generative AI (GenAI). It constitutes the next major step in the evolution of AI with much stronger reasoning and interaction capabilities that enable more autonomous behavior to tackle complex tasks. Since the initial release of ChatGPT (3.5), Generative AI has seen widespread adoption, giving users firsthand experience. However, the distinction between Agentic AI and GenAI remains less well understood. To address this gap, our survey is structured in two parts. In the first part, we compare GenAI and Agentic AI using existing literature, discussing their key characteristics, how Agentic AI remedies limitations of GenAI, and the major steps in GenAI's evolution toward Agentic AI. This section is intended for a broad audience, including academics in both social sciences and engineering, as well as industry professionals. It provides the necessary insights to comprehend novel applications that are possible with Agentic AI but not with GenAI. In the second part, we deep dive into novel aspects of Agentic AI, including recent developments and practical concerns such as defining agents. Finally, we discuss several challenges that could serve as a future research agenda, while cautioning against risks that can emerge when exceeding human intelligence.

Johannes Schneider

Neural networks are not learning optimal decision boundaries. We show that decision boundaries are situated in areas of low training data density. They are impacted by few training samples which can easily lead to overfitting. We provide a simple algorithm performing a weighted average of the prediction of a sample and its nearest neighbors' (computed in latent space) leading to a minor favorable outcomes for a variety of important measures for neural networks. In our evaluation, we employ various self-trained and pre-trained convolutional neural networks to show that our approach improves (i) resistance to label noise, (ii) robustness against adversarial attacks, (iii) classification accuracy, and to some degree even (iv) interpretability. While improvements are not necessarily large in all four areas, our approach is conceptually simple, i.e., improvements come without any modification to network architecture, training procedure or dataset. Furthermore, they are in stark contrast to prior works that often require trade-offs among the four objectives or provide valuable, but non-actionable insights.

Johannes Schneider, Pauline Kuss, Rene Abraham et al.

Generative Artificial Intelligence (GenAI), specifically large language models(LLMs) like ChatGPT, has swiftly entered organizations without adequate governance, posing both opportunities and risks. Despite extensive debates on GenAI's transformative nature and regulatory measures, limited research addresses organizational governance, encompassing technical and business perspectives. Although numerous frameworks for governance of AI exist, it is not clear to what extent they apply to GenAI. Our review paper fills this gap by surveying recent works with the purpose of better understanding fundamental characteristics of GenAI and adjusting prior frameworks specifically towards GenAI governance within companies. To do so, it extends Nickerson's framework development processes to include prior conceptualizations. Our framework outlines the scope, objectives, and governance mechanisms tailored to harness business opportunities as well as mitigate risks associated with GenAI integration. Our research contributes a focused approach to GenAI governance, offering practical insights for companies navigating the challenges of GenAI adoption and highlighting research gaps.

Giovanni De Felice, Arianna Casanova Flores, Francesco De Santis et al.

Concept-based models are an emerging paradigm in deep learning that constrains the inference process to operate through human-interpretable variables, facilitating explainability and human interaction. However, these architectures, on par with popular opaque neural models, fail to account for the true causal mechanisms underlying the target phenomena represented in the data. This hampers their ability to support causal reasoning tasks, limits out-of-distribution generalization, and hinders the implementation of fairness constraints. To overcome these issues, we propose Causally reliable Concept Bottleneck Models (C$^2$BMs), a class of concept-based architectures that enforce reasoning through a bottleneck of concepts structured according to a model of the real-world causal mechanisms. We also introduce a pipeline to automatically learn this structure from observational data and unstructured background knowledge (e.g., scientific literature). Experimental evidence suggests that C$^2$BMs are more interpretable, causally reliable, and improve responsiveness to interventions w.r.t. standard opaque and concept-based models, while maintaining their accuracy.

Johannes Schneider, Mohit Prabhushankar

The learning dynamics of deep neural networks are not well understood. The information bottleneck (IB) theory proclaimed separate fitting and compression phases. But they have since been heavily debated. We comprehensively analyze the learning dynamics by investigating a layer's reconstruction ability of the input and prediction performance based on the evolution of parameters during training. We empirically show the existence of three phases using common datasets and architectures such as ResNet and VGG: (i) near constant reconstruction loss, (ii) decrease, and (iii) increase. We also derive an empirically grounded data model and prove the existence of phases for single-layer networks. Technically, our approach leverages classical complexity analysis. It differs from IB by relying on measuring reconstruction loss rather than information theoretic measures to relate information of intermediate layers and inputs. Our work implies a new best practice for transfer learning: We show empirically that the pre-training of a classifier should stop well before its performance is optimal.

Johannes Schneider, Kilic Sinem, Daniel Stockhammer

The domain of computational design, driven by advancements in Generative AI, is transforming creative fields. We explore the transformative effects of Generative AI on the architectural design process and discuss the role of the architect. The case of architecture is interesting as designing houses is complex, involving extensive customer interaction. We employ a within-subject experiment using a popular general-purpose text-to-image tool for generating designs and providing feedback on existing designs, followed by expert interviews. The study reveals that AI can disrupt the ideation phase by enabling clients to engage in the design process through rapid visualization of their own ideas. In turn, the architect's role shifts more towards assessing the feasibility of designs generated conjointly by clients and AI. Our study also shows that while AI can provide valuable feedback on designs, it might fail to generate such designs, allowing for interesting connections to foundations in computer science, i.e., NP-completeness. AI's feedback also tends to hamper creativity and innovation by suggesting altering novel, innovative approaches toward more standardized designs. Our study also reveals that there is uncertainty among architects about the interpretative sovereignty of architecture and loss of meaning and identity when AI increasingly takes over authorship in the design process.

Johannes Schneider, Béatrice S. Hasler, Michaela Varrone et al.

We analyze anonymous interaction data of minors in class-rooms spanning several months, schools, and subjects employing a novel, simple topic modeling approach. Specifically, we categorize more than 17,000 messages generated by students, teachers, and ChatGPT in two dimensions: content (such as nature and people) and tasks (such as writing and explaining). Our hierarchical categorization done separately for each dimension includes exemplary prompts, and provides both a high-level overview as well as tangible insights. Prior works mostly lack a content or thematic categorization. While task categorizations are more prevalent in education, most have not been supported by real-world data for K-12. In turn, it is not surprising that our analysis yielded a number of novel applications. In deriving these insights, we found that many of the well-established classical and emerging computational methods, i.e., topic modeling, for analysis of large amounts of texts underperform, leading us to directly apply state-of-the-art LLMs with adequate pre-processing to achieve hierarchical topic structures with better human alignment through explicit instructions than prior approaches. Our findings support fellow researchers, teachers and students in enriching the usage of GenAI, while our discussion also highlights a number of concerns and open questions for future research.

Johannes Schneider

Autoregressive language models like GPT aim to predict next tokens, while autoencoding models such as BERT are trained on tasks such as predicting masked tokens. We train a decoder-only architecture for predicting the second to last token for a sequence of tokens. Our approach yields higher computational training efficiency than BERT-style models by employing a structured deterministic approach to masking tokens. We use our model to improve the next token predictions of a standard GPT by combining both predictions in a ``generate-then-refine'' approach. We demonstrate on different variants of GPT-2 and different datasets that (not unexpectedly) second to last token predictions are much more accurate, i.e., more than 15\% higher accuracy than standard next token predictions. The ``generate-then-refine'' approach also demonstrates notable improvements in next-token predictions, yielding smaller yet consistent and significant gains.

Francesco De Santis, Gabriele Ciravegna, Giovanni De Felice et al.

Concept Bottleneck Models (CBMs) promote interpretability by grounding predictions in human-understandable concepts. However, existing CBMs typically fix their task predictor to a single linear or Boolean expression, limiting both predictive accuracy and adaptability to diverse user needs. We propose Mixture of Concept Bottleneck Experts (M-CBEs), a framework that generalizes existing CBMs along two dimensions: the number of experts and the functional form of each expert, exposing an underexplored region of the design space. We investigate this region by instantiating two novel models: Linear M-CBE, which learns a finite set of linear expressions, and Symbolic M-CBE, which leverages symbolic regression to discover expert functions from data under user-specified operator vocabularies. Empirical evaluation demonstrates that varying the mixture size and functional form provides a robust framework for navigating the accuracy-interpretability trade-off, adapting to different user and task needs.

Alexander Okupnik, Johannes Schneider, Kyriakos Flouris

Recent success with large language models has sparked a new wave of verbal human-AI interaction. While such models support users in a variety of creative tasks, they lack the embodied nature of human interaction. Dance, as a primal form of human expression, is predestined to complement this experience. To explore creative human-AI interaction exemplified by dance, we build an interactive model based on motion capture (MoCap) data. It generates an artificial other by partially mimicking and also "creatively" enhancing an incoming sequence of movement data. It is the first model, which leverages single-person motion data and high level features in order to do so and, thus, it does not rely on low level human-human interaction data. It combines ideas of two diffusion models, motion inpainting, and motion style transfer to generate movement representations that are both temporally coherent and responsive to a chosen movement reference. The success of the model is demonstrated by quantitatively assessing the convergence of the feature distribution of the generated samples and the test set which serves as simulating the human performer. We show that our generations are first steps to creative dancing with AI as they are both diverse showing various deviations from the human partner while appearing realistic.

Johannes Schneider

We ask: Does focusing on classes predicted as likely improve model predictions? We aim for an affirmative answer by proposing two novel test-time fine-tuning methods to improve uncertain model predictions. Instead of greedily selecting the most likely class, we introduce an additional step, \emph{focus on the likely classes}, to refine predictions. By applying a theoretically motivated single gradient descent step with a large learning rate, we refine predictions when an initial forward pass indicates high uncertainty. This aligns predictions more closely with the ideal of assigning zero probability to less plausible outcomes. The experimental evaluation demonstrates accuracy gains for one of our methods, which emphasizes shared features among likely classes, across diverse text and image domain models. %Our theoretical discussion provides a deeper understanding, highlighting the varying impact of shared and non-shared features among (focus) classes. %Our discussion also suggests an interesting view on standard, offline training vs. test-time training: Opposing optimization rationales regarding breadth of feature dependence are preferable during each training phase.

Jonas Bokstaller, Julia Altheimer, Julian Dormehl et al.

Across various sectors applications of eXplainableAI (XAI) gained momentum as the increasing black-boxedness of prevailing Machine Learning (ML) models became apparent. In parallel, Large Language Models (LLMs) significantly developed in their abilities to understand human language and complex patterns. By combining both, this paper presents a novel reference architecture for the interpretation of XAI through an interactive chatbot powered by a fine-tuned LLM. We instantiate the reference architecture in the context of State-of-Health (SoH) prediction for batteries and validate its design in multiple evaluation and demonstration rounds. The evaluation indicates that the implemented prototype enhances the human interpretability of ML, especially for users with less experience with XAI.

Johannes Schneider

The learning dynamics of deep neural networks are subject to controversy. Using the information bottleneck (IB) theory separate fitting and compression phases have been put forward but have since been heavily debated. We approach learning dynamics by analyzing a layer's reconstruction ability of the input and prediction performance based on the evolution of parameters during training. We show that a prototyping phase decreasing reconstruction loss initially, followed by reducing classification loss of a few samples, which increases reconstruction loss, exists under mild assumptions on the data. Aside from providing a mathematical analysis of single layer classification networks, we also assess the behavior using common datasets and architectures from computer vision such as ResNet and VGG.

Johannes Schneider, Robin Richner, Micha Riser

Autograding short textual answers has become much more feasible due to the rise of NLP and the increased availability of question-answer pairs brought about by a shift to online education. Autograding performance is still inferior to human grading. The statistical and black-box nature of state-of-the-art machine learning models makes them untrustworthy, raising ethical concerns and limiting their practical utility. Furthermore, the evaluation of autograding is typically confined to small, monolingual datasets for a specific question type. This study uses a large dataset consisting of about 10 million question-answer pairs from multiple languages covering diverse fields such as math and language, and strong variation in question and answer syntax. We demonstrate the effectiveness of fine-tuning transformer models for autograding for such complex datasets. Our best hyperparameter-tuned model yields an accuracy of about 86.5\%, comparable to the state-of-the-art models that are less general and more tuned to a specific type of question, subject, and language. More importantly, we address trust and ethical concerns. By involving humans in the autograding process, we show how to improve the accuracy of automatically graded answers, achieving accuracy equivalent to that of teaching assistants. We also show how teachers can effectively control the type of errors made by the system and how they can validate efficiently that the autograder's performance on individual exams is close to the expected performance.

Marcus Basalla, Johannes Schneider, Jan vom Brocke

While the potential of deep learning (DL) for automating simple tasks is already well explored, recent research has started investigating the use of deep learning for creative design, both for complete artifact creation and supporting humans in the creation process. In this paper, we use insights from computational creativity to conceptualize and assess current applications of generative deep learning in creative domains identified in a literature review. We highlight parallels between current systems and different models of human creativity as well as their shortcomings. While deep learning yields results of high value, such as high-quality images, their novelty is typically limited due to multiple reasons such as being tied to a conceptual space defined by training data. Current DL methods also do not allow for changes in the internal problem representation, and they lack the capability to identify connections across highly different domains, both of which are seen as major drivers of human creativity.

Johannes Schneider, Michalis Vlachos

We examine whether data generated by explanation techniques, which promote a process of self-reflection, can improve classifier performance. Our work is based on the idea that humans have the ability to make quick, intuitive decisions as well as to reflect on their own thinking and learn from explanations. To the best of our knowledge, this is the first time that the potential of mimicking this process by using explanations generated by explainability methods has been explored. We found that combining explanations with traditional labeled data leads to significant improvements in classification accuracy and training efficiency across multiple image classification datasets and convolutional neural network architectures. It is worth noting that during training, we not only used explanations for the correct or predicted class, but also for other classes. This serves multiple purposes, including allowing for reflection on potential outcomes and enriching the data through augmentation.

Johannes Schneider, Rene Abraham, Christian Meske et al.

Artificial Intelligence (AI) governance regulates the exercise of authority and control over the management of AI. It aims at leveraging AI through effective use of data and minimization of AI-related cost and risk. While topics such as AI governance and AI ethics are thoroughly discussed on a theoretical, philosophical, societal and regulatory level, there is limited work on AI governance targeted to companies and corporations. This work views AI products as systems, where key functionality is delivered by machine learning (ML) models leveraging (training) data. We derive a conceptual framework by synthesizing literature on AI and related fields such as ML. Our framework decomposes AI governance into governance of data, (ML) models and (AI) systems along four dimensions. It relates to existing IT and data governance frameworks and practices. It can be adopted by practitioners and academics alike. For practitioners the synthesis of mainly research papers, but also practitioner publications and publications of regulatory bodies provides a valuable starting point to implement AI governance, while for academics the paper highlights a number of areas of AI governance that deserve more attention.

Johannes Schneider

Humans rely more and more on systems with AI components. The AI community typically treats human inputs as a given and optimizes AI models only. This thinking is one-sided and it neglects the fact that humans can learn, too. In this work, human inputs are optimized for better interaction with an AI model while keeping the model fixed. The optimized inputs are accompanied by instructions on how to create them. They allow humans to save time and cut on errors, while keeping required changes to original inputs limited. We propose continuous and discrete optimization methods modifying samples in an iterative fashion. Our quantitative and qualitative evaluation including a human study on different hand-generated inputs shows that the generated proposals lead to lower error rates, require less effort to create and differ only modestly from the original samples.

Johannes Schneider, Frank Breitinger

Artificial intelligence (AI) makes decisions impacting our daily lives in an increasingly autonomous manner. Their actions might cause accidents, harm, or, more generally, violate regulations. Determining whether an AI caused a specific event and, if so, what triggered the AI's action, are key forensic questions. We provide a conceptualization of the problems and strategies for forensic investigation. We focus on AI that is potentially ``malicious by design'' and grey box analysis. Our evaluation using convolutional neural networks illustrates challenges and ideas for identifying malicious AI.

Johannes Schneider

Spatial data exhibits the property that nearby points are correlated. This also holds for learnt representations across layers, but not for commonly used weight initialization methods. Our theoretical analysis quantifies the learning behavior of weights of a single spatial filter. It is thus in contrast to a large body of work that discusses statistical properties of weights. It shows that uncorrelated initialization (i) might lead to poor convergence behavior and (ii) training of (some) parameters is likely subject to slow convergence. Empirical analysis shows that these findings for a single spatial filter extend to networks with many spatial filters. The impact of (correlated) initialization depends strongly on learning rates and l2-regularization.

Johannes Schneider, Christian Meske, Michalis Vlachos

Artificial intelligence (AI) comes with great opportunities but can also pose significant risks. Automatically generated explanations for decisions can increase transparency and foster trust, especially for systems based on automated predictions by AI models. However, given, e.g., economic incentives to create dishonest AI, to what extent can we trust explanations? To address this issue, our work investigates how AI models (i.e., deep learning, and existing instruments to increase transparency regarding AI decisions) can be used to create and detect deceptive explanations. As an empirical evaluation, we focus on text classification and alter the explanations generated by GradCAM, a well-established explanation technique in neural networks. Then, we evaluate the effect of deceptive explanations on users in an experiment with 200 participants. Our findings confirm that deceptive explanations can indeed fool humans. However, one can deploy machine learning (ML) methods to detect seemingly minor deception attempts with accuracy exceeding 80% given sufficient domain knowledge. Without domain knowledge, one can still infer inconsistencies in the explanations in an unsupervised manner, given basic knowledge of the predictive model under scrutiny.

Johannes Schneider

Humans increasingly interact with Artificial intelligence(AI) systems. AI systems are optimized for objectives such as minimum computation or minimum error rate in recognizing and interpreting inputs from humans. In contrast, inputs created by humans are often treated as a given. We investigate how inputs of humans can be altered to reduce misinterpretation by the AI system and to improve efficiency of input generation for the human while altered inputs should remain as similar as possible to the original inputs. These objectives result in trade-offs that are analyzed for a deep learning system classifying handwritten digits. To create examples that serve as demonstrations for humans to improve, we develop a model based on a conditional convolutional autoencoder (CCAE). Our quantitative and qualitative evaluation shows that in many occasions the generated proposals lead to lower error rates, require less effort to create and differ only modestly from the original samples.

Johannes Schneider, Michail Vlachos

We discuss training techniques, objectives and metrics toward personalization of deep learning models. In machine learning, personalization addresses the goal of a trained model to target a particular individual by optimizing one or more performance metrics, while conforming to certain constraints. To personalize, we investigate three methods of ``curriculum learning`` and two approaches for data grouping, i.e., augmenting the data of an individual by adding similar data identified with an auto-encoder. We show that both ``curriculuum learning'' and ``personalized'' data augmentation lead to improved performance on data of an individual. Mostly, this comes at the cost of reduced performance on a more general, broader dataset.

Johannes Schneider

This work investigates fundamental questions related to learning features in convolutional neural networks (CNN). Empirical findings across multiple architectures such as VGG, ResNet, Inception, DenseNet and MobileNet indicate that weights near the center of a filter are larger than weights on the outside. Current regularization schemes violate this principle. Thus, we introduce Locality-promoting Regularization (LOCO-Reg), which yields accuracy gains across multiple architectures and datasets. We also show theoretically that the empirical finding is a consequence of maximizing feature cohesion under the assumption of spatial locality.

Johannes Schneider

Current topic models often suffer from discovering topics not matching human intuition, unnatural switching of topics within documents and high computational demands. We address these concerns by proposing a topic model and an inference algorithm based on automatically identifying characteristic keywords for topics. Keywords influence topic-assignments of nearby words. Our algorithm learns (key)word-topic scores and it self-regulates the number of topics. Inference is simple and easily parallelizable. Qualitative analysis yields comparable results to state-of-the-art models (eg. LDA), but with different strengths and weaknesses. Quantitative analysis using 9 datasets shows gains in terms of classification accuracy, PMI score, computational performance and consistency of topic assignments within documents, while most often using less topics.

Johannes Schneider, Avi Bernstein, Jan Vom Brocke et al.

All methodologies for detecting plagiarism to date have focused on the final digital "outcome", such as a document or source code. Our novel approach takes the creation process into account using logged events collected by special software or by the macro recorders found in most office applications. We look at an author's interaction logs with the software used to create the work. Detection relies on comparing the histograms of multiple logs' command use. A work is classified as plagiarism if its log deviates too much from logs of "honestly created" works or if its log is too similar to another log. The technique supports the detection of plagiarism for digital outcomes that stem from \emph{unique} tasks, such as theses and \emph{equal} tasks such as assignments for which the same problem sets are solved by multiple students. Focusing on the latter case, we evaluate this approach using logs collected by an interactive development environment (IDE) from more than sixty students who completed three programming assignments.

Johannes Schneider, Bin Lu

We derive algorithms for efficient secure numerical and logical operations using a recently introduced scheme for secure multi-party computation~\cite{sch15} in the semi-honest model ensuring statistical or perfect security. To derive our algorithms for trigonometric functions, we use basic mathematical laws in combination with properties of the additive encryption scheme in a novel way. For division and logarithm we use a new approach to compute a Taylor series at a fixed point for all numbers. All our logical operations such as comparisons and large fan-in AND gates are perfectly secure. Our empirical evaluation yields speed-ups of more than a factor of 100 for the evaluated operations compared to the state-of-the-art.

Johannes Schneider

A client wishes to outsource computation on confidential data to a network of parties. He does not trust a single party but believes that multiple parties do not collude. To solve this problem, we use the idea of treating one of the parties as a helper. A helper assists computation only. Often using more parties ensures confidentiality despite more corrupted parties. This does not hold for adding a helper. But a helper can in some cases lower the amount of communication asymptotically to the theoretical minimum of one bit per AND gate, improving significantly on schemes without a helper. It can also allow for very efficient computations of certain functions, as we show for the exponential function with public base.