arXiv subjects

Philipp Raith

Publications and source records attributed to Philipp Raith.

13 recordsLinked to original sources

arXiv2026

Clear-Weighted Bit Allocation for Satellite Downlinks

Earth-observation satellites capture more imagery than intermittent ground contacts can transmit. Onboard systems threshold a cloud detector, discard frames or tiles, and compress the survivors with a fixed codec. On expert-labeled imagery, these rules remove more than one-fifth of clear pixels, primarily through detector false positives. We train a neural codec with a clear-probability-weighted reconstruction loss, reallocating coded bytes from clouds to clear ground without requiring or transmitting a cloud map onboard. Each capture is encoded into a resumable base layer and a dependent refinement layer, while clear content is estimated from features produced by the encoder. At each contact, we causally rank arrived layers using estimated clear content, unfinished bytes, deadline slack, and aggregate deadline pressure. The scheduler serves base and computational deadlines, bounds stored residual bytes, and resumes interrupted packets. We evaluate the onboard-to-downlink pipeline using real entropy-coded bytes, orbit-derived interruptible contact capacities, and measured service time and energy on resource-constrained embedded accelerators. Clear-weighted codecs require up to 47.8\% fewer bytes than learned-compression baselines at matched clear-region quality. The optimized encoder consumes less time and energy than one pass of the cloud detector used by the frame-discard rules. Relative to fixed two-stage service on the same streams, our scheduler more than doubles deadline-full clear-content delivery for the interrupted combined cohort, reaches 83.6\% of a certified clairvoyant upper bound, and exceeds replayed reference orders in deadline-usable delivery.

cs.NI↗

arXiv2025

Multi-Dimensional Autoscaling of Stream Processing Services on Edge Devices

Edge devices have limited resources, which inevitably leads to situations where stream processing services cannot satisfy their needs. While existing autoscaling mechanisms focus entirely on resource scaling, Edge devices require alternative ways to sustain the Service Level Objectives (SLOs) of competing services. To address these issues, we introduce a Multi-dimensional Autoscaling Platform (MUDAP) that supports fine-grained vertical scaling across both service- and resource-level dimensions. MUDAP supports service-specific scaling tailored to available parameters, e.g., scale data quality or model size for a particular service. To optimize the execution across services, we present a scaling agent based on Regression Analysis of Structural Knowledge (RASK). The RASK agent efficiently explores the solution space and learns a continuous regression model of the processing environment for inferring optimal scaling actions. We compared our approach with two autoscalers, the Kubernetes VPA and a reinforcement learning agent, for scaling up to 9 services on a single Edge device. Our results showed that RASK can infer an accurate regression model in merely 20 iterations (i.e., observe 200s of processing). By increasingly adding elasticity dimensions, RASK sustained the highest request load with 28% less SLO violations, compared to baselines.

cs.DC↗

arXiv2025

Leveraging Neural Graph Compilers in Machine Learning Research for Edge-Cloud Systems

This work presents a comprehensive evaluation of neural network graph compilers across heterogeneous hardware platforms, addressing the critical gap between theoretical optimization techniques and practical deployment scenarios. We demonstrate how vendor-specific optimizations can invalidate relative performance comparisons between architectural archetypes, with performance advantages sometimes completely reversing after compilation. Our systematic analysis reveals that graph compilers exhibit performance patterns highly dependent on both neural architecture and batch sizes. Through fine-grained block-level experimentation, we establish that vendor-specific compilers can leverage repeated patterns in simple architectures, yielding disproportionate throughput gains as model depth increases. We introduce novel metrics to quantify a compiler's ability to mitigate performance friction as batch size increases. Our methodology bridges the gap between academic research and practical deployment by incorporating compiler effects throughout the research process, providing actionable insights for practitioners navigating complex optimization landscapes across heterogeneous hardware environments.

cs.DC↗

arXiv2025

Towards Multi-dimensional Elasticity for Pervasive Stream Processing Services

This paper proposes a hierarchical solution to scale streaming services across quality and resource dimensions. Modern scenarios, like smart cities, heavily rely on the continuous processing of IoT data to provide real-time services and meet application targets (Service Level Objectives -- SLOs). While the tendency is to process data at nearby Edge devices, this creates a bottleneck because resources can only be provisioned up to a limited capacity. To improve elasticity in Edge environments, we propose to scale services in multiple dimensions -- either resources or, alternatively, the service quality. We rely on a two-layer architecture where (1) local, service-specific agents ensure SLO fulfillment through multi-dimensional elasticity strategies; if no more resources can be allocated, (2) a higher-level agent optimizes global SLO fulfillment by swapping resources. The experimental results show promising outcomes, outperforming regular vertical autoscalers, when operating under tight resource constraints.

cs.PF↗

arXiv2024

FOOL: Addressing the Downlink Bottleneck in Satellite Computing with Neural Feature Compression

Nanosatellite constellations equipped with sensors capturing large geographic regions provide unprecedented opportunities for Earth observation. As constellation sizes increase, network contention poses a downlink bottleneck. Orbital Edge Computing (OEC) leverages limited onboard compute resources to reduce transfer costs by processing the raw captures at the source. However, current solutions have limited practicability due to reliance on crude filtering methods or over-prioritizing particular downstream tasks. This work presents FOOL, an OEC-native and task-agnostic feature compression method that preserves prediction performance. FOOL partitions high-resolution satellite imagery to maximize throughput. Further, it embeds context and leverages inter-tile dependencies to lower transfer costs with negligible overhead. While FOOL is a feature compressor, it can recover images with competitive scores on quality measures at lower bitrates. We extensively evaluate transfer cost reduction by including the peculiarity of intermittently available network connections in low earth orbit. Lastly, we test the feasibility of our system for standardized nanosatellite form factors. We demonstrate that FOOL permits downlinking over 100x the data volume without relying on prior information on the downstream tasks.

cs.LG↗

arXiv2023

Architectural Vision for Quantum Computing in the Edge-Cloud Continuum

Quantum processing units (QPUs) are currently exclusively available from cloud vendors. However, with recent advancements, hosting QPUs is soon possible everywhere. Existing work has yet to draw from research in edge computing to explore systems exploiting mobile QPUs, or how hybrid applications can benefit from distributed heterogeneous resources. Hence, this work presents an architecture for Quantum Computing in the edge-cloud continuum. We discuss the necessity, challenges, and solution approaches for extending existing work on classical edge computing to integrate QPUs. We describe how warm-starting allows defining workflows that exploit the hierarchical resources spread across the continuum. Then, we introduce a distributed inference engine with hybrid classical-quantum neural networks (QNNs) to aid system designers in accommodating applications with complex requirements that incur the highest degree of heterogeneity. We propose solutions focusing on classical layer partitioning and quantum circuit cutting to demonstrate the potential of utilizing classical and quantum computation across the continuum. To evaluate the importance and feasibility of our vision, we provide a proof of concept that exemplifies how extending a classical partition method to integrate quantum circuits can improve the solution quality. Specifically, we implement a split neural network with optional hybrid QNN predictors. Our results show that extending classical methods with QNNs is viable and promising for future work.

quant-ph↗

arXiv2023

FrankenSplit: Efficient Neural Feature Compression with Shallow Variational Bottleneck Injection for Mobile Edge Computing

The rise of mobile AI accelerators allows latency-sensitive applications to execute lightweight Deep Neural Networks (DNNs) on the client side. However, critical applications require powerful models that edge devices cannot host and must therefore offload requests, where the high-dimensional data will compete for limited bandwidth. This work proposes shifting away from focusing on executing shallow layers of partitioned DNNs. Instead, it advocates concentrating the local resources on variational compression optimized for machine interpretability. We introduce a novel framework for resource-conscious compression models and extensively evaluate our method in an environment reflecting the asymmetric resource distribution between edge devices and servers. Our method achieves 60% lower bitrate than a state-of-the-art SC method without decreasing accuracy and is up to 16x faster than offloading with existing codec standards.

eess.IV↗

arXiv2013

Lorentz meets Fano spectral line shapes: A universal phase and its laser control

Symmetric Lorentzian and asymmetric Fano line shapes are fundamental spectroscopic signatures that quantify the structural and dynamical properties of nuclei, atoms, molecules, and solids. This study introduces a universal temporal-phase formalism, mapping the Fano asymmetry parameter q to a phase {\phi} of the time-dependent dipole-response function. The formalism is confirmed experimentally by laser-transforming Fano absorption lines of autoionizing helium into Lorentzian lines after attosecond-pulsed excitation. We also prove the inverse, the transformation of a naturally Lorentzian line into a Fano profile. A further application of this formalism amplifies resonantly interacting extreme-ultraviolet light by quantum-phase control. The quantum phase of excited states and its response to interactions can thus be extracted from line-shape analysis, with scientific applications in many branches of spectroscopy.

physics.atom-ph↗

arXiv2012

Quantum Interferometry and Correlated Two-Electron Wave-Packet Observation in Helium

The concerted motion of two or more bound electrons governs atomic and molecular non-equilibrium processes and chemical reactions. It is thus a long-standing scientific dream to measure the dynamics of two bound correlated electrons in the quantum regime. Quantum wave packets were previously observed for single-active electrons on their natural attosecond timescales. However, at least two active electrons and a nucleus are required to address the quantum three-body problem. This situation is realized in the helium atom, but direct time-resolved observation of two-electron wave-packet motion remained an unaccomplished challenge. Here, we measure a 1.2-femtosecond quantum beating among low-lying doubly-excited states in helium to evidence a correlated two-electron wave packet. Our experimental method combines attosecond transient-absorption spectroscopy at unprecedented high spectral resolution (20 meV near 60 eV) with an intensity-tuneable visible laser field to couple the quantum states from the perturbative to the strong-coupling regime. This multi-dimensional transient-coupling scheme reveals an inversion of the characteristic Fano line shapes for a range of doubly-excited states. Employing Fano-type autoionization as a natural quantum interferometer, a dynamical phase shift by laser coupling to the N=2 continuum is postulated and experimentally quantified. This phase maps a transition from effectively single-active-electron to two-electron dynamics as the electron-electron interaction increases in lower-lying quantum states. In the future, such experiments will provide benchmark data for testing dynamical few-body quantum theory. They will boost our understanding of chemically and biologically important metastable electronic transition states and their dynamics on attosecond time scales.

physics.atom-ph↗

arXiv2012

Noisy pulses enhance temporal resolution in pump-probe spectroscopy

Time-resolved measurements of quantum dynamics are based on the availability of controlled events (e.g. pump and probe pulses) that are shorter in duration than the typical evolution time scale of the dynamical processes to be observed. Here we introduce the concept of noise-enhanced pump-probe spectroscopy, allowing the measurement of dynamics significantly shorter than the average pulse duration by exploiting randomly varying, partially coherent light fields consisting of bunched colored noise. It is shown that statistically fluctuating fields can be superior by more than a factor of 10 to frequency-stabilized fields, with important implications for time-resolved pump-probe experiments at x-ray free-electron lasers (FELs) and, in general, for measurements at the frontiers of temporal resolution (e.g. attosecond spectroscopy). As an example application, the concept is used to explain the recent experimental observation of vibrational wave packet motion in a deuterium molecular ion on time scales shorter than the average pulse duration.

physics.optics↗

arXiv2011

Two-Dimensional Spectral Interferometry using the Carrier-Envelope Phase

Two- and multi-dimensional spectroscopy is used in physics and chemistry to obtain structural and dynamical information that would otherwise be invisible by the projection into a one-dimensional data set such as a single emission or absorption spectrum. Here, we introduce a qualitatively new two-dimensional spectroscopy method by employing the carrier-envelope phase (CEP). Instead of measuring spectral vs. spectral information, the combined application of spectral interferometry and CEP control allows the measurement of otherwise inseparable temporal events on an attosecond time scale. As a specific example, we apply this general method to the case of attosecond pulse train generation, where it allows to separate contributions of three different sub-cycle electron quantum paths within one and the same laser pulse, resulting in a better physical understanding and quantification of the transition region between cutoff and plateau harmonics. The CEP-dependent separation in time between two full-cycle spaced attosecond pulses was determined to modulate by (54 +/- 16) attoseconds.

physics.chem-ph↗

arXiv2010

High Harmonic Generation via Continuum Wave-Packet Interference

High-order harmonic generation (HHG) is investigated theoretically in the over-the-barrier ionization (OBI) regime revealing the strong signature of interference between two separately ionized and separately propagating free wave packets of a single electron. The interference leads to the emission of coherent light at a photon energy corresponding to the kinetic-energy difference of the two recolliding electron quantum paths, thus complementary to the well-known classical three-step picture of HHG. As will be shown by time-frequency analysis of the emitted radiation, the process entirely dominates the coherent HHG emission after the atomic ground state has been depleted by a strong field. Moreover, it can be isolated from the continuum--bound harmonics via phase-matching.

physics.atom-ph↗

arXiv2010

Sub-Cycle Strong-Field Interferometry

A nonlinear interferometry scheme is described theoretically to induce and resolve electron wave- function beating on time scales shorter than the optical cycle of the time-delayed pump and probe pulses. By employing two moderately intense few-cycle laser fields with a stable carrier-envelope phase, a large range of the entire electronic level structure of a quantum system can be retrieved. In contrast to single-photon excitation schemes, the retrieved electronic states include levels that are both dipole- and non-dipole-accessible from the ground electronic state. The results show that strong-field interferometry can reveal both high-resolution and broad-band spectral information at the same time with important consequences for quantum-beat spectroscopy on attosecond or even shorter time scales.

physics.chem-ph↗