Christian-Albrechts-Universität

Selected work

Representative Papers

LimitNet: Progressive, Content-Aware Image Offloading for Extremely Weak Devices & Networks

Jun 03, 2024ACM SIGMOBILE International Conference on Mobile Systems, Applications, and Services

This work addresses the challenge of real-time cloud-based image inference for ultra-resource-constrained IoT devices operating over LPWANs—characterized by ultra-low bandwidth, high packet loss rates, and extremely low duty cycles. We propose the first lightweight, content-aware progressive coding framework: a deep learning–based progressive encoder dynamically prioritizes transmission of semantically critical bits; a content-sensitive bit allocation mechanism and an ultra-low-overhead deployment strategy for Cortex-M7 microcontrollers enable cloud inference to commence as soon as partial data arrives. Evaluated on ImageNet-1000, CIFAR-100, and COCO, our method achieves average accuracy gains of 14.01%, 18.01%, and 0.1 mAP@0.5, respectively, while reducing bandwidth consumption by 61.24%, 83.68%, and 42.25%. Encoding overhead increases only 4% over JPEG—significantly overcoming the fundamental limitation of conventional non-progressive codecs, which fail to decode meaningfully under partial reception.

3 citationsRead paper

The Support of Bin Packing Is Exponential

Oct 01, 2025Embedded Systems and Applications

This paper investigates the lower bound on the *support number*—i.e., the number of distinct bin types (packing patterns)—in the Bin Packing problem, as a function of the number $d$ of distinct item sizes. It establishes, for the first time, a tight exponential lower bound of $2^{Omega(d)}$, resolving a long-standing gap between existing upper and lower bounds. To achieve this, the authors introduce a novel aggregation technique based on equality-constrained integer linear programming (ILP): it equivalently reduces a high-dimensional ILP with multiple constraints to a low-dimensional model while preserving variable upper bounds—thereby enabling both combinatorial structural analysis and computational complexity characterization. This method not only yields the tight support-number bound but also uncovers the fundamental complexity bottlenecks underlying classical heuristics such as First-Fit and Next-Fit. The framework provides a new paradigm for theoretical analysis and algorithm design for high-dimensional knapsack-type problems.

1 citationsRead paper