@article{21915, abstract = {Hydrological models commonly use very simple snow accumulation and melt models based on air temperature information, namely, a temperature threshold for snow accumulation as well as for snowmelt, and a melt factor. This utility emerges due to the simplicity, efficiency, and generally good performance of such models if sufficient calibration information is available. At scales beyond single gauged catchments, the estimation and evaluation of the temperature thresholds and the melt factor has been difficult due to a lack of observations on snow accumulation and melt. Using a recently published Northern Hemisphere snow water equivalent dataset (NH-SWE) and co-located climate station observations of temperature and precipitation (4736 stations across the Northern Hemisphere), this work estimates melt factors and temperature thresholds for snow modelling based on station observations and provides the first large-scale and long-term (1950–2023) evaluation of a simple temperature-index snow model and its parameters across a diverse range of snow climates. Our study reveals that the 0 °C as precipitation-phase threshold captures most snowfall days (89 %) and the 0 °C as snowmelt initiation threshold captures most snowmelt days (76 %). Adjusting large-scale uniform threshold values does not consistently improve performance across all snow accumulation and melt metrics. Estimated melt factors based on observations converge towards 3–5 mm (°C d)−1 for deeper snowpack climates (peak snow water equivalent >300 mm), but their estimation may be more challenging for colder climates with shallower snowpacks (<300 mm), conditions where the derived melt factors cover a wider range (1 to 12 mm (°C d)−1) and a much higher interannual and spatial variability. The temperature-index snow model performs consistently well, on average, across the available Northern Hemisphere data set for estimating long-term mean values of seasonal snow cover onset, snowmelt season onset, mean snow accumulation and snowmelt rates, but challenges may arise due to biases in temperature records or solid precipitation undercatch. Peak snow water equivalent is likely underestimated for deep or alpine snowpacks, while it is likely overestimated for shallow snowpacks in the coldest and continental climates. The best median performance of the temperature-index approach lies on relatively shallow snowpacks in temperate climates. This study provides valuable insights into temperature-threshold snowfall modelling and temperature-index melt modelling for applications across diverse climates and environments, and the results should help refine regional modelling approaches to enhance our understanding of snowpack responses to global warming.}, author = {Fontrodona-Bach, Adrià and Schaefli, Bettina and Woods, Ross and Larsen, Joshua R.}, issn = {1607-7938}, journal = {Hydrology and Earth System Sciences}, number = {9}, pages = {2613--2636}, publisher = {Copernicus Publications}, title = {{Estimating robust melt factors and temperature thresholds for snow modelling across the Northern Hemisphere}}, doi = {10.5194/hess-30-2613-2026}, volume = {30}, year = {2026}, } @inproceedings{21916, abstract = {Social network graphs are central to graph learning research, serving as standard benchmarks for algorithm evaluation. However, existing datasets focus mainly on mainstream social media platforms whose structures are shaped notably by algorithmic recommendations. This raises an important question: would alternative, decentralized social networks exhibit different properties? We address this by studying the Fediverse; a collection of decentralized social networks (such as Mastodon and Lemmy). These platforms differ fundamentally from for-profit social media, notably in decentralization and absence of recommendation algorithms, which may yield distinct graph structures. We introduce Fedivertex, a dataset of over 400 graphs from seven decentralized networks, collected weekly over six months. The dataset, released with a companion Python package to facilitate its use, supports research on temporal and structural aspects of decentralized social networks. In particular, we benchmark applications to decentralized machine learning and community detection.}, author = {Damie, Marc and Cyffers, Edwige Audrey Lucienne}, booktitle = {2026 Proceedings of the ACM Web Conference 2026}, isbn = {9798400723070}, location = {Dubai}, pages = {8393--8396}, publisher = {ACM}, title = {{Fedivertex: A graph dataset based on decentralized Social Media}}, doi = {10.1145/3774904.3792868}, year = {2026}, } @article{21917, abstract = {A defining feature of quantum many-body systems is the exponential scaling of the Hilbert space with the number of degrees of freedom. This exponential complexity naïvely renders a complete state characterization, for instance via the complete set of bipartite Renyi entropies for all disjoint regions, a challenging task. Recently, a compact way of storing subregions' purities by encoding them as amplitudes of a fictitious quantum wave function, known as entanglement feature, was proposed. Notably, the entanglement feature can be a simple object even for highly entangled quantum states. However the complexity and practical usage of the entanglement feature for general quantum states has not been explored. In this work, we demonstrate that the entanglement feature can be efficiently learned using only a polynomial amount of samples in the number of degrees of freedom through the so-called tensor cross interpolation (TCI) algorithm, assuming it is expressible as a finite bond dimension MPS. We benchmark this learning process on Haar and random MPS states, confirming analytic expectations. Applying the TCI algorithm to quantum eigenstates of various one dimensional quantum systems, we identify cases where eigenstates have entanglement feature learnable with TCI. We conclude with possible applications of the learned entanglement feature, such as quantifying the distance between different entanglement patterns and finding the optimal one-dimensional ordering of physical indices in a given state, highlighting the potential utility of the proposed purity interpolation method.}, author = {Kolisnyk, Dmytro and Medina Ramos, Raimel A and Vasseur, Romain and Serbyn, Maksym}, issn = {2521-327X}, journal = {Quantum}, publisher = {Verein zur Förderung des Open Access Publizierens in den Quantenwissenschaften}, title = {{Tensor cross interpolation of purities in quantum many-body systems}}, doi = {10.22331/q-2026-05-22-2114}, volume = {10}, year = {2026}, } @phdthesis{21918, author = {Khudiakova, Kseniia}, issn = {2663-337X}, pages = {89}, publisher = {Institute of Science and Technology Austria}, title = {{How epistasis and purifying selection shape genetic diversity}}, doi = {10.15479/AT-ISTA-21918}, year = {2026}, } @inproceedings{21921, abstract = {A variety of problems in geometry processing boil down to finding the most parallel field relative to a connection. Instances of this prototypical problem show up in computing direction fields and stripe patterns, quadrilateral meshing, and visualization of fluid flows. When the class of allowed fields includes those with topological defects, a relaxation is required to make the problem well-posed. We observe that these problems can be viewed as synchronization problems, which admit a natural semidefinite relaxation. We propose a unified method of solving all these problems via the efficient Burer-Monteiro factorization method. Geometrically, this amounts to lifting the field values to a higher-dimensional manifold, naturally resolving the singular nature of defects. Practically, we show that our convex relaxation method achieves better and more reliable optima than previous work employing alternative relaxations}, author = {Pacheco-Tallaj, Natalia and Couplet, Matteo and Chien, Edward and Palmer, David}, booktitle = {SIGGRAPH Conference Papers}, location = {Los Angeles, CA, United States}, publisher = {ACM}, title = {{Synchronizing fields with singularities}}, doi = {10.1145/3799902.3811225}, year = {2026}, } @article{21923, abstract = {The appearance of simulated natural phenomena heavily depends on the way surfaces are textured. However, applying texture maps to dynamic deformable surfaces presents a significant challenge, due to ever-shifting differences in length scales involved. When these surfaces move and advect the texture along with them, their final appearance degrades as deformed regions dramatically distort their texture map. Modifications to the texture directly at the pixel level in response to the deformation may introduce ghosting artifacts and look unnatural. In the real world, the appearance of surface details on a deforming material changes through the interplay of physical processes such as rupturing, exposure of internal structure, or wrinkling. Motivated by these behaviors, in this work we explore how physical principles can guide the texturing methods based on the measure of surface deformation. We present two novel wave-based procedural texturing algorithms which reproduce common physical properties like advection and self-similarity, enabling the plausible animation of deforming objects with extreme texture map distortions. Our algorithms are fully procedural, require no actual physics simulation, and store no state or history of deformation besides the input UV map, making them highly parallelizable on the GPU and efficient enough for real-time applications. We show the versatility of the method by animating physical phenomena with extreme deformations such as flowing lava, stretching putty and outpouring sludge.}, author = {Kalinov, Aleksei and Ly, Mickaël and Hafner, Christian and Wojtan, Christopher J}, issn = {0730-0301}, journal = {ACM Transactions on Graphics}, keywords = {Procedural animation}, location = {Los Angeles, CA, United States}, number = {4}, publisher = {ACM}, title = {{Physics-inspired procedural texturing of extremely deformable surfaces}}, doi = {10.1145/3811353}, volume = {45}, year = {2026}, } @article{21928, abstract = {Antibiotic combination in time and space is a key strategy to combat antimicrobial resistance. The success of such treatment designs requires their robust efficacy across treatment conditions and a pathogen’s genomic diversity. This study found that an initial treatment with a β-lactam antibiotic causes robust cellular sensitization towards an aminoglycoside antibiotic across the high-risk human pathogen Pseudomonas aeruginosa, including resistant strains. This phenomenon of cellular sensitization, termed negative hysteresis, is modulated by the Cpx envelope stress response system and linked to membrane stress during growth. The increase in efficacy is achieved through a β-lactam induced elevated cellular uptake of the subsequently administered aminoglycoside. Negative hysteresis and the Cpx system are linked in several cases to the expression of synergistic drug interactions, thus enhancing efficacy of antibiotic combinations. Overall, our study identifies the phenomenon of negative hysteresis as a robustly inducible phenotype and thus a unique focus for optimizing antimicrobial therapy.}, author = {Buchholz, Florian and Upterworth, Lina M. and Tueffers, Leif and Groth, Espen E. and Haas, Kira and Schütz, Daniel and Savietto Scholz, Abigail and Batra, Aditi and Pal, Surajit and Banerjee, Samarpita and Dubey, Badri N. and Franzenburg, Sören and Kalsdorf, Barbara and Rabe, Klaus F. and Nurjadi, Dennis and Rupp, Jan and Andersson, Dan I. and Sondermann, Holger and Bramkamp, Marc and Römhild, Roderich and Schulenburg, Hinrich}, issn = {2041-1723}, journal = {Nature Communications}, publisher = {Springer Nature}, title = {{Robust antibiotic sensitization of pathogenic Pseudomonas aeruginosa via negative hysteresis in the cell envelope}}, doi = {10.1038/s41467-026-71178-5}, volume = {17}, year = {2026}, } @article{21929, abstract = {The import of proteins into mitochondria poses fundamental mechanistic challenges: aggregation-prone precursor proteins must be maintained in aqueous compartments and threaded through narrow pores without becoming stuck or mislocalized. Recent evidence from mitochondrial protein import studies and other chaperone systems underscores the critical role of dynamics in balancing sufficiently tight binding, promiscuity, specificity, and release. Dynamic binding of client precursor proteins to import machinery components arises naturally from the avidity of their interactions. Conformational entropy enhances their stability, while the multivalent nature of these interactions ensures that client transfer to downstream insertases occurs without a substantial energy barrier. Here, we discuss this emerging paradigm of dynamic protein handling, using examples where dynamic structures have been resolved and highlight outstanding questions.}, author = {Schneider, Jakob and Guillerm, Undina and Simoes Pereira, Caroline and Schanda, Paul}, issn = {1469-896X}, journal = {Protein Science}, number = {6}, publisher = {Wiley}, title = {{Dynamic disorder is crucial for mitochondrial protein import}}, doi = {10.1002/pro.70630}, volume = {35}, year = {2026}, } @article{21930, abstract = {We present the discovery of extreme nitrogen enrichment by Wolf Rayet nitrogen (WN) stars in the metal-poor (∼10%Z⊙), lensed, compact (Reff ∼ 20 pc) galaxy RXCJ2248 at z = 6.1, revealed by unprecedentedly deep JWST/NIRSpec medium-resolution spectroscopy from the GLIMPSE-D Survey. The exquisite signal-to-noise ratio reveals multiple high-ionization nebular lines and broad Balmer and [O III] components (FWHM ∼700–3000 km s −1 ). We detect broadened He II λ1640 and λ4687 (FWHM ∼ 530 km s −1 ) and strong N III λ4642 emission consistent with a population of WN stars, making RXCJ2248 the most distant galaxy with confirmed Wolf Rayet (WR) features to date. We measure the multiphase nebular density across five ions, the direct-method metallicity ( 12 + log(O/H) = 7.753 ± 0.025 ), and a nonuniform elemental enrichment pattern of extreme N/O enhancement ( log(N/O) = 0.391 ± 0.037 from N+, N+2 , and N+3 ) but suppressed C/O relative to empirical C/N trends. We show that this abundance pattern can be explained by enrichment from a dual-burst with a low WR carbon/WN ratio, as expected at low metallicities. Crucially, these signatures can only arise during a brief, rare evolutionary window shortly after a burst (∼3–6 Myr), when WN stars dominate chemical feedback but before dilution by later yields (e.g., supernovae). The observed frequency of strong N emitters at high−z implies a ∼50 Myr burst duty cycle, suggesting that N/O outliers may represent a brief but ubiquitous phase in the evolution of highly star-forming early galaxies. The WN detection in RXCJ2248, therefore, provides the first direct evidence of WR-driven nitrogen enrichment in the first billion years of the Universe and a novel timing argument for the bursty star formation cycles that shaped galaxies at cosmic dawn.}, author = {Berg, Danielle A. and Naidu, Rohan P. and Chisholm, John and Atek, Hakim and Fujimoto, Seiji and Kokorev, Vasily and Furtak, Lukas J. and Kobayashi, Chiaki and Schaerer, Daniel and Adamo, Angela and Fei, Qinyue and Korber, Damien and Matthee, Jorryt J and Marques-Chaves, Rui and Martinez, Zorayda and Mcquinn, Kristen B.W. and Muñoz, Julian B. and Oesch, Pascal A. and Saldana-Lopez, Alberto and Stark, Daniel P. and Stephenson, Mabel G. and Hsiao, Tiger Yu Yang}, issn = {1538-4357}, journal = {The Astrophysical Journal}, number = {2}, publisher = {IOP Publishing}, title = {{A fleeting GLIMPSE of N/O enrichment at cosmic dawn: Evidence for Wolf Rayet N stars in a z = 6.1 galaxy}}, doi = {10.3847/1538-4357/ae5e4c}, volume = {1003}, year = {2026}, } @article{21931, abstract = {In 1873, James C. Maxwell conjectured that the electric field generated by n point charges in generic position has at most (n-1)^2 isolated zeroes. The first (nonoptimal) upper bound was only obtained in 2007 by Gabrielov, Novikov, and Shapiro, who also posed two additional interesting conjectures. In this article, we give the best upper bound known to date on the number of zeroes of the electric field, and construct a counterexample to Conjecture 1.8 by Gabrielov, Novikov, and Shapiro that the number of equilibria cannot exceed those of the distance function defined by the unit point charges. Finally, we note that it is quite possible that Maxwell's quadratic upper bound is not tight, so it is prudent to find lower bounds. Hence, we also explore examples and construct configurations of charges achieving the highest ratios of the number of electric field zeroes by point charges found to this day.}, author = {Edelsbrunner, Herbert and Fillmore, Christopher D and Oliveira, Goncalo}, issn = {1460-244X}, journal = {Proceedings of the London Mathematical Society}, number = {5}, publisher = {Wiley}, title = {{Counting equilibria of the electrostatic potential}}, doi = {10.1112/plms.70163}, volume = {132}, year = {2026}, } @inproceedings{21932, abstract = {We present LLMQ, an end-to-end CUDA/C++ implementation for medium-sized language-model training, e.g. 3B to 32B parameters, on affordable, commodity GPUs. These devices are characterized by low memory availability and slow communication compared to datacentre-grade GPUs. Consequently, we showcase a range of optimizations that target these bottlenecks, including activation checkpointing, offloading, and copy-engine based collectives. LLMQ is able to train or fine-tune a 7B model on a single 16GB mid-range gaming card, or a 32B model on a workstation equipped with 4 RTX 4090s. This is achieved while executing a standard 8-bit training pipeline, without additional algorithmic approximations, and maintaining FLOP utilization of around 50%. The efficiency of LLMQ rivals that of production-scale systems on much more expensive cloud-grade GPUs.}, author = {Schultheis, Erik and Alistarh, Dan-Adrian}, booktitle = {2nd Conference on Parsimony and Learning}, issn = {2640-3498}, location = {Stanford, CA, United States}, pages = {265--284}, publisher = {ML Research Press}, title = {{LLMQ: Efficient lower-precision LLM training for consumer GPUs}}, volume = {328}, year = {2026}, } @article{21933, abstract = {We consider the liquid drop model with a positive background density in the thermodynamic limit. We prove a two-term asymptotics for the ground state energy per unit volume in the dilute limit. Our proof justifies the expectation that optimal configurations consist of droplets of unit size that arrange themselves according to minimizers for the Jellium problem for point particles. In particular, we provide the first rigorous derivation of what is known as the gnocchi phase in astrophysics.}, author = {Frank, Rupert L. and Lewin, Mathieu and Seiringer, Robert}, issn = {1097-0312}, journal = {Communications on Pure and Applied Mathematics}, publisher = {Wiley}, title = {{Liquid drop model for nuclear matter in the low density limit}}, doi = {10.1002/cpa.70039}, year = {2026}, } @article{21934, abstract = {An accurate characterisation of the physical properties of galaxies at cosmic dawn is key to understanding the origin of the high abundance of UV-bright galaxies at z≳10. We exploit deep (9.1-hour exposure time) NIRSpec PRISM observations of GHZ2 to constrain the sources of ionising radiation and the properties of the interstellar medium (ISM) in this bright, compact, and highly ionising galaxy at z=12.3. We measure with high significance the prominent N IV, C IV, He II, O III, C III, O II, and Ne III emission features previously detected in shallower observations, and confirm the detection of the N III] λ1750 multiplet, yielding tight constraints on the N/O ratio, which is found to be ≃2 times the solar value. We also detect the Mg II λ2800, [Fe IV] λ2833 and Si II λ1812 doublets, the H8+HeI λλ3889 blend, and the Si IV+O IV] λλ1400 absorption complex. The O III λ3133 fluorescence line is only detected in the first observing epoch, implying variability on a rest-frame time span of 19 days, strongly suggesting the presence of an active nucleus. Combining the NIRSpec dataset with available optical and far-infrared constraints from MIRI and ALMA, we show that the emission spectrum of GHZ2 cannot be reproduced by single-density spectro-photometric models, even under extreme assumptions on the ionisation parameter and electron density. Multi-zone photoionisation modelling performed with the HOMERUN code demonstrates that star formation must be occurring in a strongly stratified ISM, where both low-/intermediate-density gas and high-density regions (log(ne/cm−3 ) ≳ 4) coexist. The GHZ2 emission landscape is consistent with either a composite star-formation plus AGN scenario, or with star formation occurring in a combination of radiation- and matter-bounded regions. Purely radiation-bounded stellar models fail to reproduce the observed He II emission, making an additional hard ionising component unavoidable.}, author = {Castellano, M. and Napolitano, L. and Moreschini, B. and Calabrò, A. and Christensen, L. and Llerena, M. and Bakx, T. J.L.C. and Belfiore, F. and Bevacqua, D. and Dickinson, M. and Fontana, A. and Gandolfi, G. and Gasparetto, T. and Marconi, A. and Mascia, Sara and Merlin, E. and Morishita, T. and Nanayakkara, T. and Paris, D. and Pentericci, L. and Pérez-Díaz, B. and Roberts-Borsani, G. and Rojas-Ruiz, S. and Santini, P. and Treu, T. and Vanzella, E. and Vulcani, B. and Wang, X. and Yoon, I. and Zavala, J.}, issn = {2565-6120}, journal = {The Open Journal of Astrophysics}, publisher = {Maynooth Academic Publishing}, title = {{Investigating ionising sources and the complex interstellar medium of GHZ2 at z=12.3}}, doi = {10.33232/001c.160281}, volume = {9}, year = {2026}, } @article{21948, abstract = {The cerebral cortex comprises diverse neuron and glial cell types generated by radial glial progenitors (RGPs) during development. Although RGPs broadly differentiate according to temporally and spatially regulated molecular logics, the lineage hierarchies linking individual progenitors to defined cell (sub)types are not well understood. Clone-resolved transcriptomics, combining molecular barcoding and single-cell RNA sequencing, allow high-resolution lineage tracing at the single-clone/cell level across different species and models. In this mini-review, we synthesize recent advances in this field, uncovering unexpected lineage relationships in the developing brain, with a particular focus on the cerebral cortex. We further highlight new insights into species-specific differences in the developmental programs generating cell-type diversity, linking changes in clonal architecture to lineage diversification during cortical evolution.}, author = {Varela Martínez, Irene and Pipicelli, Fabrizia and Hippenmeyer, Simon}, issn = {1879-0380}, journal = {Current Opinion in Genetics and Development}, publisher = {Elsevier}, title = {{Tracing cell lineages in the developing brain: Insights from mosaic analysis and clone-resolved transcriptomics}}, doi = {10.1016/j.gde.2026.102487}, volume = {99}, year = {2026}, } @inproceedings{21949, abstract = {Cardiac T1 mapping provides critical quantitative insights into myocardial tissue composition, enabling the assessment of pathologies such as fibrosis, inflammation, and edema. However, the inherently dynamic nature of the heart imposes strict limits on acquisition times, making high-resolution T1 mapping a persistent challenge. Compressed sensing (CS) approaches have reduced scan durations by undersampling k-space and reconstructing images from partial data, and recent studies show that jointly optimizing the undersampling patterns with the reconstruction network can substantially improve performance. Still, most current T1 mapping pipelines rely on static, hand-crafted masks that do not exploit the full acceleration and accuracy potential. Furthermore, most existing methods do not levarage the physical T1 decay model in optimization. In this work, we introduce T1- PILOT: an end-to-end method that explicitly incorporates the T1 signal relaxation model into the sampling–reconstruction framework to guide the learning of non-Cartesian trajectories, cross-frame alignment, and T1 decay estimation. Through extensive experiments on the CMRxRecon dataset, T1-PILOT significantly outperforms several baseline strategies (including learned single-mask and fixed radial or golden-angle sampling schemes), achieving higher T1 map fidelity at greater acceleration factors. In particular, we observe consistent gains in PSNR and VIF relative to existing methods, along with marked improvements in delineating finer myocardial structures. Our results highlight that optimizing sampling trajectories in tandem with the physical relaxation model leads to both enhanced quantitative accuracy and reduced acquisition times. Code for reproducing all experiments and results is available at https://github.com/tamirshor7/T1-PILOT}, author = {Shor, Tamir and Freiman, Moti and Baskin, Chaim and Bronstein, Alexander}, booktitle = {Medical Imaging with Deep Learning}, issn = {2640-3498}, keywords = {Cardiac T1 Mapping, Trajectory Optimization and Reconstruction, PhysicsInformed Deep-Learning}, location = {Taipei, Taiwan}, pages = {1969--1982}, publisher = {ML Research Press}, title = {{T1-PILOT: Physics-informed learned optimized trajectories for T1 mapping acceleration}}, volume = {315}, year = {2026}, } @article{21951, abstract = {The central engines of Little Red Dots (LRDs) may be “black hole stars” (BH*s), early stages of black hole growth characterized by dense gas envelopes. So far, the most direct evidence for BH*s comes from a handful of sources where the host galaxy is completely outshone as suggested by their remarkably steep Balmer breaks. Here we present a novel scheme to disentangle BH*s from their host galaxies assuming that the [O III]5008˚A line arises exclusively from the host. Using a sample of 98 LRDs (z ≈ 2 − 9) with high quality NIRSpec/PRISM spectra, we demonstrate that the hostsubtracted median stack displays a Balmer break > 2× stronger than massive quiescent galaxies, with the rest-optical continuum resembling a blackbody-like SED (Teff ≈ 4050 K, log(Lbol) ≈ 43.9 erg s−1 , Reff ≈ 1300 au). We measure a steep Balmer decrement (Hα/Hβ > 10) and numerous density-sensitive features (e.g., Fe II, He I, O I). These are hallmark signatures of dense gas envelopes, providing population-level evidence that BH*s indeed power LRDs. In the median LRD, BH*s account for ∼ 20% of the UV emission, ∼ 50% at the Balmer break, and ∼ 90% at wavelengths longer than Hα with the remainder arising from the host. BH*s preferentially reside in low-mass galaxies (M⋆ ≈ 108 M⊙) undergoing recent starbursts, as evidenced by extreme emission line EWs (e.g., [O III]5008˚A≈ 1100˚A, C III]≈ 12˚A), thereby favoring BH* origins linked to star-formation. We show V-shaped LRD selections are biased to high BH*/host fractions (≳ 60% at 5500˚A) – less dominant BH*s may be powering JWST’s blue broad-line AGN. We find BH*s are so commonplace and transient (duty cycle ∼ 1%, lifetime ∼ 10 Myrs) that every massive black hole may have once shone as a BH*. }, author = {Sun, Wendy Q. and Naidu, Rohan P. and Matthee, Jorryt J and De Graaff, Anna and Chisholm, John and Greene, Jenny E. and Oesch, Pascal A. and Torralba Torregrosa, Alberto and Hviding, Raphael E. and Brammer, Gabriel and Simcoe, Robert A. and Bose, Sownak and Bouwens, Rychard and Dayal, Pratika and Eilers, Anna Christina and Fei, Qinyue and Furtak, Lukas J. and Gottumukkala, Rashmi and Goulding, Andy and Heintz, Kasper E. and Hirschmann, Michaela and Kokorev, Vasily and Leja, Joel and Liu, Zhaoran and Natarajan, Priyamvada and Santarelli, Andrew D. and Setton, David J. and Smith, Aaron and Tacchella, Sandro and Volonteri, Marta and Walter, Fabian and Weibel, Andrea and Williams, Christina C.}, issn = {2565-6120}, journal = {The Open Journal of Astrophysics}, publisher = {Maynooth Academic Publishing}, title = {{Little Red Dot - Host Galaxy = Black Hole Star: A gas-enshrouded heart at the center of every Little Red Dot}}, doi = {10.33232/001c.162505}, volume = {9}, year = {2026}, } @article{21954, abstract = {We investigate a framework for train-free MRI segmentation based on Topological Data Analysis. The pipeline proceeds in three steps, first identifying the whole object to segment via automatic thresholding, then detecting a distinctive subset whose topology is known in advance, and finally deducing the various components of the segmentation. A key ingredient is the extraction of approximate representative cycles from persistence diagrams, which provides an interpretable link between persistent features and anatomical components. To clarify the method’s scope, we make the underlying topological and intensity assumptions explicit, quantify when they hold on real data, and analyze typical failure modes. We evaluate the approach on glioblastoma and on fetal cortical plate segmentation, with comparisons to unsupervised and deep-learning references. By operating without large annotated datasets, the method is well suited to scarce-data settings and provides an interpretable baseline and practical initialization for expert refinement or learning-based pipelines.}, author = {François, Anton and Tinarrage, Raphaël}, issn = {1573-7683}, journal = {Journal of Mathematical Imaging and Vision}, number = {3}, publisher = {Springer Nature}, title = {{Train-free segmentation in MRI with cubical persistent homology}}, doi = {10.1007/s10851-026-01300-1}, volume = {68}, year = {2026}, } @article{21955, abstract = {AgRP neurons cause hunger, the drive to seek and consume food. Their activation by fasting is key for survival and is thought to be triggered by feedback when energy stores are low. However, we know that environmental cues can also regulate AgRP neurons since cues that predict future food intake rapidly inhibit AgRP neurons, but is the converse true: can the prediction of future fasting rapidly activate AgRP neurons? Here, we show in mice that such rapid fasting activation of AgRP neurons does occur. This rapid activation is driven by excitatory input from paraventricular hypothalamic (PVH) neurons expressing Sim2, which are bidirectionally sensitive to predictions of future energy state. Thus, cognitively processed contextual information conveyed by PVHSim2 neurons strongly activates AgRP neurons. Lastly, chronic silencing of PVHSim2 neurons causes persistent hypophagia. This PVHSim2-to-AgRP-neuron circuit, by anticipating and preventing negative energy balance, provides an important new dimension of hunger regulation.}, author = {Walker, Samuel J. and Lowenstein, Elijah D. and Douglass, Amelia May Barnett and Thomas, Callum M.P. and Madara, Joseph C. and Kucukdereli, Hakan and Barbosa-Meillon, Eunice A. and Tao, Jenkang and Resch, Jon M. and Lowell, Bradford B.}, issn = { 1097-4199}, journal = {Neuron}, keywords = {hunger, hypothalamus, AGRP neurons, neuroscience, metabolism, homeostasis, feeding, food intake, energy balance, appetite}, publisher = {Elsevier}, title = {{A hypothalamic circuit for anticipating future changes in energy balance}}, doi = {10.1016/j.neuron.2026.05.010}, year = {2026}, } @phdthesis{21957, abstract = {This thesis investigates algorithmic certification and approximation methods for degenerate semidefinite programs (SDPs) and the singular roots of polynomial systems. In the first part, we present a hybrid symbolic-numeric algorithm for certifying the feasibility of weakly feasible, degenerate SDPs. By reformulating linear matrix inequalities (LMIs) into a structured polynomial system via facial reduction and incidence varieties, we guarantee the existence of an isolated exact solution. This algebraic reduction enables the certification of maximum-rank numerical approximations using methods from algebraic geometry. In the second part, we address the severe ill-conditioning and loss of quadratic convergence that plague standard path-tracking methods near isolated singular roots. To overcome this, we propose tracking algorithms that achieve superlinear convergence without the computational bloat characteristic of classical deflation techniques. By modeling the solution path as a generalized fractional Puiseux series, our approach combines an explicitly derived algebraic predictor with a localized hyperplane desingularization phase during the corrector step. Furthermore, we introduce a continuous path-limit method and an extension of the geometric sequence rule to directly extract exact fractional exponents. This bypasses traditional heuristic trial-and-error methods and explicitly accommodates sparse series expansions. Numerical experiments confirm that our method significantly reduces the cumulative number of matrix inversions while achieving high-accuracy root approximations, even for heavily degenerate systems exhibiting higher coranks.}, author = {Zapata, Jeferson}, isbn = {978-3-99078-079-4}, issn = {2663-337X}, pages = {89}, publisher = {Institute of Science and Technology Austria}, title = {{Overcoming degeneracy and singularity : Techniques for semidefinite programs and homotopy continuation endgames}}, doi = {10.15479/AT-ISTA-21957}, year = {2026}, } @misc{21960, abstract = {Solitons - localized wave packets that travel without spreading - play a central role in understanding transport and properties of nonlinear systems. In quantum many-body systems, however, such robust excitations are typically destroyed by thermalization. Here, we theoretically demonstrate the existence of solitonic excitations in high-energy states of Rydberg atom chains in the regime of strong nearest-neighbor Rydberg blockade. These localized wave packets propagate directionally atop a special class of reviving initial states related to quantum many-body scars and are capable of carrying energy. Exhibiting long coherence times, these states constitute a form of non-ergodic quantum dynamics and can be efficiently implemented on Rydberg atom simulators. In this work, in addition to a phenomenological description of solitons, we identify their counterpart in a classical nonlinear dynamical system, demonstrate their potential use in quantum information transfer, and conjecture their relevance for anomalous energy transport reported in numerical studies of Rydberg atom arrays.}, author = {Kerschbaumer, Aron}, publisher = {Institute of Science and Technology Austria}, title = {{Research Data: "Quasi-solitons in Rydberg atom chains"}}, doi = {10.15479/AT-ISTA-21960}, year = {2026}, }