Repository for Publications and Research Data
News from the ETH Library
Recently Added
Height-dependent thermal properties of extreme-ultraviolet campfires in the quiet Sun
Item type: Journal Article
Podladchikova O.; Warmuth A.; Harra L.; et al. (2026)
Context. Small-scale impulsive energy-release events are widely considered a key ingredient of quiet-Sun coronal heating. The discovery of extreme ultraviolet (EUV) campfires by Solar Orbiter has opened a new observational window at unprecedented spatial scales. Aims. We quantified the thermal-energy content of EUV campfires, examined their impulsive heating characteristics, and investigated how their properties vary with atmospheric height. Methods. We analysed the complete set of campfires obtained in a single observing sequence on 30 May 2020 with Solar Orbiter/EUI HRIEUV at 174 Å. Temperatures and emission measures were adopted from SDO/AIA diagnostics, and geometric uncertainties were explored using multiple volume models. Statistical properties were characterised using cumulative distribution functions and maximum-likelihood techniques, and results were compared with Solar Orbiter/STIX microflares and SDO/AIA nanoflare-like events for context. Results. Campfire durations show a clear dependence on height: short-duration events occur at all heights, whereas longer-lived brightenings are preferentially associated with chromospheric layers. Emitting volumes and thermal energies, by contrast, show no comparable height dependence. The background-subtracted thermal energies span ~1020–1024 erg, with rare events reaching ~5 × 105 erg. The statistically robust power-law regime extends over ~1022–1024 erg, while non-background-subtracted estimates are systematically higher by approximately one order of magnitude. Directly detected campfires contribute a sub-percent fraction of the canonical quiet-Sun heating requirement. Accounting approximately for the detection threshold and differential emission measure completeness could raise this contribution to a value of order a few percent, though this remains a single-epoch, order-of-magnitude estimate. Conclusions. Extreme ultraviolet campfires extend the flare energy distribution towards lower energies and provide height-resolved constraints on impulsive heating in the quiet Sun. The height dependence of event durations – driven mainly by an excess of long-lived events in the lower chromosphere – combined with broadly invariant energy and volume distributions, indicates that atmospheric stratification primarily modulates cooling and dissipation timescales rather than the characteristic energy scale of the heating process.
JWST/MIRI detection of molecular H2 winds from an edge-on class II source HV Tau C
Item type: Journal Article
Pathak V.C.; Puravankara M.; Tyagi H.; et al. (2026)
Context. The evolution of protoplanetary disks is regulated by the interplay between accretion onto the central star and mass loss through jets and winds. While atomic and ionised outflows are commonly observed, direct detections of molecular winds in evolved class II disks remain rare, limiting our understanding of their role in disk evolution and dispersal. Aims. We aim to characterise the spatial, thermal, kinematic, and dynamical properties of molecular hydrogen (H2) emission associated with the nearly edge-on class II disk HV Tau C, and to assess the dynamical impact of its molecular wind. We also simultaneously constrained the accretion properties using H I recombination lines detected in the same mid-infrared (mid-IR) spectrum, enabling us to investigate the accretion–ejection connection in this system. Methods. We analyse mid-IR JWST/MIRI-MRS data from the MINDS Cycle 1 GTO programme, extracting spatially resolved pure-rotational H2 emission. Rotational diagrams were used to determine excitation temperatures and column densities, while position–velocity diagrams provide the kinematics of extended emission. Combining excitation and kinematic information, we estimated the wind mass-loss rate, momentum rate, and mechanical luminosity. Accretion luminosities and mass-accretion rates are derived from H I lines, accounting for uncertainties in visual extinction due to the edge-on geometry. Results. We detect spatially extended pure-rotational H2 emission from HV Tau C, revealing a wide-angled, bi-conical molecular wind extending well beyond the near-IR scattered-light disk, the ALMA 887 µm dust continuum, and the compact 12CO (J=3–2) gas disk. The H2 rotational diagram requires at least two temperature components: a warm component at ∼600 K and a hot component at ∼2000 K, similar to those found in much younger protostars. The kinematics indicate outward motions of a few tens of km s−1 and dynamical timescales of tens to hundreds of years. The inferred mass-loss rate is about 10−8 M⊙ yr−1, while accretion rates lie between 10−10 and 10−8 M⊙ yr−1. The measured accretion rate may be underestimated due to geometrical effects, and the true accretion rate could be higher than that inferred from the mid-IR H I lines. Conclusions. These results suggest that wide-angled molecular H2 winds may persist into the class II phase, with inferred outflow rates that are comparable to those observed in some protostellar systems. Such outflows may play a significant role in angular momentum removal, disk evolution, and disk dispersal even in relatively evolved systems.
Agentic AI Systems in Radiology: Design, Applications, Evaluation, and Challenges
Item type: Review Article
Bluethgen C.; Van Veen D.; Truhn D.; et al. (2026)
Building agents that perceive and act on their environment with a degree of autonomy has long been a focus of artificial intelligence research. This pursuit has become far more practical with the emergence of large language models (LLMs) capable of using natural language to integrate information, follow instructions, and perform forms of reasoning and planning across a wide range of tasks. With its multimodal data streams and orchestrated workflows spanning multiple systems, radiology is uniquely suited to benefit from agents that can adapt to context and automate complex, repetitive tasks. In radiology, LLMs and their multimodal variants have already demonstrated promising performance for individual tasks such as information extraction and report summarization. However, using LLMs in isolation underutilizes their potential to support complex, multistep workflows where decisions depend on evolving context from multiple information sources. Equipping LLMs with external tools and feedback mechanisms enables them to drive systems that exhibit a spectrum of autonomy, ranging from semiautomated workflows to more adaptive agents capable of managing complex processes. This review article examines the design of LLM-driven agentic systems, explores representative radiology applications, discusses evaluation frameworks, and outlines key challenges, including error cascades, tool-use effectiveness, and integration with health information technology infrastructure.
The Cost of Secure Restaking vs. Proof-Of-Stake
Item type: Conference Paper
Mamageishvili A.; Sudakov B. (2026)
We compare the total capital efficiency of secure restaking and Proof-of-Stake (PoS) protocols. First, we consider the sufficient condition by EigenLayer project for the restaking graph to be secure, that can be checked in a polynomial time in the restaking graph size. The condition implies that it is always possible to transform such a restaking graph into separate secure PoS protocols. We derive two main results: upper and lower bounds on the required extra stakes to add to the validators of the secure restaking graph to be able to transform it into secure PoS protocols. In particular, we show that the restaking savings compared to PoS protocols can be very large and can asymptotically grow as a square root of the number of validators. We also study a complementary question of aggregating secure PoS protocols into a secure restaking graph and provide matching lower and upper bounds on the PoS savings.
Large Scale Middle Mile Network Design Through Efficient Local Search
Item type: Conference Paper
Sartori C.S.; Shen J.; Kireeva A.; et al. (2026)
E-commerce competition has intensified in recent years, further raising the delivery speed and cost efficiency standards that companies must meet. The middle mile network that connects fulfillment centers to local distribution centers plays a decisive role in achieving these standards. In this paper, we study the Middle Mile Network Design Problem, which determines optimal network investments that balance the trade-off between fastest delivery speed and minimal operational costs. The problem requires computing truck schedules, shipment routes, and parcel flows through intermediate facilities subject to complex capacity constraints. Two key challenges make this problem computationally hard to solve for standard exact and heuristic approaches: (i) the use of complex black-box evaluations to predict operational outcomes, and (ii) problem instances involving over 200,000 commodities that must be solved within hours for operational planning. We propose a metaheuristic based on variable neighborhood search with specialized local search operators alongside lazy neighborhood evaluation that enable efficient exploration of the high-dimensional solution space. Our algorithm achieves up to 7.6% solution improvement over an incumbent heuristic while reducing runtime by up to 44%.