Oliver Müller

Marcel Meyer, Sascha Kaltenpoth, Kevin Zalipski et al.

While Time Series Foundation Models (TSFMs) offer transformative capabilities for forecasting, they simultaneously risk triggering a fundamental evaluation crisis. This crisis is driven by information leakage due to overlapping training and test sets across different models, as well as the illegitimate transfer of global patterns to test data. While the ability to learn shared temporal dynamics represents a primary strength of these models, their evaluation on historical archives often permits the exploitation of observed global shocks, which violates the independence required for valid benchmarking. We introduce TS-Arena, a platform that restores the operational integrity of forecasting by treating the genuinely unknown future as the definitive test environment. By implementing a pre-registration mechanism on live data streams, the platform ensures that evaluation targets remain physically non-existent during inference, thereby enforcing a strict global temporal split. This methodology establishes a moving temporal frontier that prevents historical contamination and provides an authentic assessment of model generalization. Initially applied within the energy sector, TS-Arena provides a sustainable infrastructure for comparing foundation models under real-world constraints. A prototype of the platform is available at https://huggingface.co/spaces/DAG-UPB/TS-Arena.

Marcel Meyer, Sascha Kaltenpoth, Kevin Zalipski et al.

Time Series Foundation Models (TSFMs) represent a new paradigm for time series forecasting, offering zero-shot forecasting capabilities without the need for domain-specific pre-training or fine-tuning. However, as with Large Language Models (LLMs), evaluating TSFMs is tricky, as with ever more extensive training sets, it becomes more and more challenging to ensure the integrity of benchmarking data. Our investigation of existing TSFM evaluation highlights multiple challenges, ranging from the representativeness of the benchmark datasets, over the lack of spatiotemporal evaluation, to risks of information leakage due to overlapping and obscure datasets, and the memorization of global patterns caused by external shocks like economic crises or pandemics. Our findings reveal widespread confusion regarding data partitions, risking inflated performance estimates and incorrect transfer of global knowledge to local time series. We argue for the development of robust evaluation methodologies to prevent pitfalls already observed in LLM and classical time series benchmarking, and call upon the research community to design new, principled approaches, such as evaluations on truly out-of-sample future data, to safeguard the integrity of TSFM assessment.

Sascha Kaltenpoth, Oliver Müller

Adopting Large language models (LLMs) in organizations potentially revolutionizes our lives and work. However, they can generate off-topic, discriminating, or harmful content. This AI alignment problem often stems from misspecifications during the LLM adoption, unnoticed by the principal due to the LLM's black-box nature. While various research disciplines investigated AI alignment, they neither address the information asymmetries between organizational adopters and black-box LLM agents nor consider organizational AI adoption processes. Therefore, we propose LLM ATLAS (LLM Agency Theory-Led Alignment Strategy) a conceptual framework grounded in agency (contract) theory, to mitigate alignment problems during organizational LLM adoption. We conduct a conceptual literature analysis using the organizational LLM adoption phases and the agency theory as concepts. Our approach results in (1) providing an extended literature analysis process specific to AI alignment methods during organizational LLM adoption and (2) providing a first LLM alignment problem-solution space.

Tobias Alt, Andrea Ibisch, Clemens Meiser et al.

Generative AI models are capable of performing a wide variety of tasks that have traditionally required creativity and human understanding. During training, they learn patterns from existing data and can subsequently generate new content such as texts, images, audio, and videos that align with these patterns. Due to their versatility and generally high-quality results, they represent, on the one hand, an opportunity for digitalisation. On the other hand, the use of generative AI models introduces novel IT security risks that must be considered as part of a comprehensive analysis of the IT security threat landscape. In response to this risk potential, companies or authorities intending to use generative AI should conduct an individual risk analysis before integrating it into their workflows. The same applies to developers and operators, as many risks associated with generative AI must be addressed during development or can only be influenced by the operating organisation. Based on this, existing security measures can be adapted, and additional measures implemented.

Jan-Peter Kucklick, Oliver Müller

In the house credit process, banks and lenders rely on a fast and accurate estimation of a real estate price to determine the maximum loan value. Real estate appraisal is often based on relational data, capturing the hard facts of the property. Yet, models benefit strongly from including image data, capturing additional soft factors. The combination of the different data types requires a multi-view learning method. Therefore, the question arises which strengths and weaknesses different multi-view learning strategies have. In our study, we test multi-kernel learning, multi-view concatenation and multi-view neural networks on real estate data and satellite images from Asheville, NC. Our results suggest that multi-view learning increases the predictive performance up to 13% in MAE. Multi-view neural networks perform best, however result in intransparent black-box models. For users seeking interpretability, hybrid multi-view neural networks or a boosting strategy are a suitable alternative.

Jan-Peter Kucklick, Oliver Müller

Buying a home is one of the most important buying decisions people have to make in their life. The latest research on real-estate appraisal focuses on incorporating image data in addition to structured data into the modeling process. This research measures the prediction performance of satellite images and structured data by using convolutional neural networks. The resulting CNN model trained performs 7% better in MAE than the advanced baseline of a neural network trained on structured data. Moreover, sliding-window heatmap provides visual interpretability of satellite images, revealing that neighborhood structures are essential in the price estimation.

Dominik Welke, Joos Behncke, Marina Hader et al.

Brain-controlled robots are a promising new type of assistive device for severely impaired persons. Little is however known about how to optimize the interaction of humans and brain-controlled robots. Information about the human's perceived correctness of robot performance might provide a useful teaching signal for adaptive control algorithms and thus help enhancing robot control. Here, we studied whether watching robots perform erroneous vs. correct action elicits differential brain responses that can be decoded from single trials of electroencephalographic (EEG) recordings, and whether brain activity during human-robot interaction is modulated by the robot's visual similarity to a human. To address these topics, we designed two experiments. In experiment I, participants watched a robot arm pour liquid into a cup. The robot performed the action either erroneously or correctly, i.e. it either spilled some liquid or not. In experiment II, participants observed two different types of robots, humanoid and non-humanoid, grabbing a ball. The robots either managed to grab the ball or not. We recorded high-resolution EEG during the observation tasks in both experiments to train a Filter Bank Common Spatial Pattern (FBCSP) pipeline on the multivariate EEG signal and decode for the correctness of the observed action, and for the type of the observed robot. Our findings show that it was possible to decode both correctness and robot type for the majority of participants significantly, although often just slightly, above chance level. Our findings suggest that non-invasive recordings of brain responses elicited when observing robots indeed contain decodable information about the correctness of the robot's action and the type of observed robot.