Andre Kaup

Large language models (LLMs) can support predictive maintenance by reasoning over text-based expert knowledge, but their reliability depends on how such knowledge is structured and retrieved. This work presents an expert-knowledge-grounded LLM inference pipeline that combines Delphi- and FMEA-derived maintenance knowledge, retrieval-augmented prompting, and engineered telemetry summaries. In a multi-robot case study, we evaluate basic knowledge queries, complex diagnostic questions, and telemetry description tasks. Retrieved expert knowledge improves LLM-based diagnostic answers, while a deterministic knowledge-graph reasoner performs best on threshold-driven telemetry questions. The results indicate a complementary design in which graph inference provides traceable rule execution and expert-knowledge-grounded LLMs synthesize diagnostic explanations from selected evidence.

The demand for high-quality, immersive video experiences has necessitated adopting higher frame rates for more realistic scene portrayal. However, there is a growing demand for energy-saving video applications, where temporal downsampling is a crucial energy-saving factor. To this end, the main contributions of this paper are twofold: First, we present an in-depth analysis of the impact of frame rate reductions on the visual quality of the video and the encoding as well as decoding energy. Second, we propose a lightweight frame rate selection method for energy- and quality-aware encoding. Concerning the first contribution, this paper performs extensive encoding and decoding measurements, followed by an investigation of the impact of temporal downsampling on the energy demand of encoding and decoding at different frame rates. Furthermore, we determine the objective visual quality of the downsampled videos. As a result of this investigation, we identify content- and quantization-setting-dependent energy-aware frame rates, i.e., the temporal downsampling factors that lead to Pareto-optimality in terms of energy and quality. We demonstrate that significant energy savings are achieved while maintaining constant visual quality. Subsequently, a subjective experiment is conducted to verify this observation regarding perceptual quality using mean opinion scores. As the second contribution, we propose an energy-aware frame rate selection method that extracts spatio-temporal features from the video sequences. Based on these features, the proposed method employs a feature-based supervised machine learning approach to predict energy-aware frame rates for a given quantization parameter and video sequence, aiming to reduce energy consumption during encoding and decoding. The experimental results demonstrate that the proposed method offers significant energy savings, with an average of 17.46% and 17.60% of encoding and decoding energy demand reduction, respectively, alongside 3.38% average bitrate savings at a constant quality.

Local rate control is a key enabler to generalize image and video compression for dedicated challenges, such as video coding for machines. While traditional hybrid video coding can easily adapt the local rate-distortion trade-off by changing the local quantization parameter, no such approach is currently available for learning-based video compression. In this paper, we propose LRC-DHVC, a hierarchical video compression network, which allows continuous local rate control on a pixel level to vary the spatial quality distribution within individual video frames. This is achieved by concatenating a quality map to the input frame and applying a weighted MSE loss which matches the pixelwise trade-off factors in the quality map. During training, the model sees a variety of quality maps due to a constrained-random generation. Our model is the first neural video compression network, which can continuously and spatially adapt to varying quality constraints. Due to the wide quality and bit rate range, a single set of network parameters is sufficient. Compared to single rate point networks, which scale linearly with the number of rate points, the memory requirements for our network parameters remain constant. The code and model are available at link-updated-upon-acceptance.

Semantic segmentation of high-resolution remote sensing images (HRRSIs) is essential for various geospatial applications. While recent transformer-based models demonstrate promising global modeling capabilities, they are susceptible to attention noise—particularly in cluttered or noisy scenes—leading to degraded segmentation quality. This challenge is amplified in HRRSIs, where spatially heterogeneous noise and ambiguous object boundaries are prevalent. To address this problem, we propose PDDFormer, a Position-aware Differential Denoising Transformer tailored for semantic segmentation of HRRSIs. Central to our design is the Spatially Gated Differential Attention (SGDA), which adaptively modulates denoising strength across spatial positions using a learnable gate. Unlike prior differential attention models, SGDA introduces position-aware denoising that dynamically adjusts attention suppression at each token, guided by spatial semantics. This allows PDDFormer to suppress redundant background responses while preserving critical structural semantics. Extensive experiments on two public benchmarks, ISPRS Potsdam and LoveDA, demonstrate that PDDFormer consistently outperforms state-of-the-art methods in both accuracy and boundary preservation. Moreover, PDDFormer achieves favorable performance-efficiency trade-offs, making it practical for large-scale geospatial analysis. Visualizations further confirm the model’s ability to generate cleaner, more coherent segmentation maps, validating the effectiveness of our spatially adaptive denoising strategy.

Semantic segmentation of high-resolution remote sensing images remains challenging due to the degradation of high-frequency semantic cues during convolutional encoding and the lack of frequency consistency in multi-stage feature fusion. To address these issues, we propose FreDNet, a frequency-guided denoising network that explicitly enhances frequency-sensitive representations throughout the segmentation process. Specifically, we introduce the Dual-path Residual Block (DRB), which incorporates a Frequency-aware Denoising Module (FDM) and a Frequency-aware Fusion Module (FFM) to suppress frequency-domain noise while preserving edge structures. Furthermore, we design a Frequency-aware Cross-level Fusion Module (FCFM) that leverages frequency intensity response maps to adaptively fuse encoder and decoder features. These components work collaboratively to enhance the frequency robustness and spatial consistency of the segmentation predictions. Extensive experiments on three challenging benchmarks, ISPRS Vaihingen, ISPRS Potsdam, and LoveDA, demonstrate that FreDNet achieves superior performance, surpassing the latest state-of-the-art approaches by up to 0.8% in mean IoU and 0.9% in overall accuracy, while maintaining a lightweight inference cost. In addition, ablation study confirms the contribution of each component of FreDNet.

In today's society, live video streaming and user generated content streamed from battery powered devices are ubiquitous. Live streaming requires real-time video encoding, and hardware video encoders are well suited for such an encoding task. In this paper, we introduce a high-level feature model using Gaussian process regression that can predict the encoding energy of a hardware video encoder. In an evaluation setup restricted to only P-frames and a single keyframe, the model can predict the encoding energy with a mean absolute percentage error of approximately 9%. Further, we demonstrate with an ablation study that spatial resolution is a key high-level feature for encoding energy prediction of a hardware encoder. A practical application of our model is that it can be used to perform a prior estimation of the energy required to encode a video at various spatial resolutions, with different coding standards and codec presets.