IST Austria: PubRep (Institute of Science and Technology)
Not a member yet
6140 research outputs found
Sort by
Operator-valued twisted Araki–Woods algebras
No full textWe introduce operator-valued twisted Araki–Woods algebras. These are operator-valued versions of a class of second quantization algebras that includes q-Gaussian and q-Araki–Woods algebras and also generalize Shlyakhtenko’s von Neumann algebras generated by operator-valued semicircular variables. We develop a disintegration theory that reduces the isomorphism type of operator-valued twisted Araki–Woods algebras over type I factors to the scalar-valued case. Moreover, these algebras come with a natural weight, and we characterize its modular theory. We also give sufficient criteria that guarantee the factoriality of these algebras
Thin and ephemeral snow shapes melt and runoff dynamics in the Peruvian Andes
No full textThe snow and glaciers of the Peruvian Andes provide vital water supplies in a region facing water scarcity and substantial glacier change. However, there remains a lack of understanding of snow processes and quantification of the contribution of melt to runoff. Here we apply a distributed glacio-hydrological model over the Rio Santa basin to disentangle the role of the cryosphere in the Andean water cycle. Only at the highest elevations (>5000 m a.s.l.) is the snow cover continuous; at lower elevations, the snowpack is thin and ephemeral, with rapid cycles of snowfall and melt. Due to the large catchment area affected by ephemeral snow, its contribution to catchment inputs is substantial (23% and 38% in the wet and dry season, respectively). Ice melt is crucial in the mid-dry season (up to 44% of inputs). Our results improve estimates of water fluxes and call for further process-based modelling across the Andes
Evolution of the UV slope of galaxies at cosmic morning (z > 4): The properties of extremely blue galaxies
No full textWe present an analysis of the UV continuum slope, β, using a sample of 726 galaxies with z > 4, selected from a mixture of JWST ERS, GTO, and GO observational programs. We considered only spectroscopic data obtained with the low-resolution (R ∼ 30 − 300) PRISM/CLEAR NIRSpec configuration. Studying the correlation between β and MUV, we find an overall decreasing trend, described by β = ( − 0.055 ± 0.017)MUV + ( − 2.98 ± 0.34). This is consistent with previous studies, where brighter galaxies show redder β values. However, when analyzing the trend in separate redshift bins, we find that at high redshift the relation becomes much flatter and is consistent with a flat slope within 1σ. Furthermore, we find that β tends to decrease with redshift, following β = ( − 0.075 ± 0.010)z + ( − 1.496 ± 0.056). This is consistent with most recent results showing a steepening of the spectra at higher z. We selected a sample of galaxies with extremely blue slopes (i.e., β 10% in at least 25% of the XBGs, whereas all the red sources exhibit much lower fesc values. A reduced nebular continuum contribution – resulting from either a high escape fraction or a bursty star formation history – is likely the origin of the extremely blue slopes
Filamin A editing in myeloid cells reduces intestinal inflammation and protects from colitis
No full textPatho-mechanistic origins of ulcerative colitis are still poorly understood. The actin cross-linker filamin A (FLNA) impacts cellular responses through interaction with cytosolic proteins. Posttranscriptional A-to-I editing generates two forms of FLNA: genome-encoded FLNAQ and FLNAR. FLNA is edited in colon fibroblasts, smooth muscle cells, and endothelial cells. We found that the FLNA editing status determines colitis severity. Editing was highest in healthy colons and reduced during murine and human colitis. Mice that exclusively express FLNAR were highly resistant to DSS-induced colitis, whereas fully FLNAQ animals developed severe inflammation. While the genetic induction of FLNA editing influenced transcriptional states of structural cells and microbiome composition, we found that FLNAR exerts protection specifically via myeloid cells, which are physiologically unedited. Introducing fixed FLNAR did not hamper cell migration but reduced macrophage inflammation and rendered neutrophils less prone to NETosis. Thus, loss of FLNA editing correlates with colitis severity, and targeted editing of myeloid cells serves as a novel therapeutic approach in intestinal inflammation
Replication Data for: "Improving Prediction Accuracy in Chimeric Proteins with Windowed Multiple Sequence Alignment"
No full textOfficial implementation, windowed MSAs, and the predictions as reported in the manuscript titled "Improving Prediction Accuracy in Chimeric Proteins with Windowed Multiple Sequence Alignment". (2025-06-27
LIPIcs
No full textWe consider equilibria in multiplayer stochastic graph games with terminal-node rewards. In such games, Nash equilibria are defined assuming that each player seeks to maximise their expected payoff, ignoring their aversion or tolerance to risk. We therefore study risk-sensitive equilibria (RSEs), where the expected payoff is replaced by a risk measure. A classical risk measure in the literature is the entropic risk measure, where each player has a real valued parameter capturing their risk-averseness. We introduce the extreme risk measure, which corresponds to extreme cases of entropic risk measure, where players are either extreme optimists or extreme pessimists. Under extreme risk measure, every player is an extremist: an extreme optimist perceives their reward as the maximum payoff that can be achieved with positive probability, while an extreme pessimist expects the minimum payoff achievable with positive probability. We argue that the extreme risk measure, especially in multi-player graph based settings, is particularly relevant as they can model several real life instances such as interactions between secure systems and potential security threats, or distributed controls for safety critical systems. We prove that RSEs defined with the extreme risk measure are guaranteed to exist when all rewards are non-negative. Furthermore, we prove that the problem of deciding whether a given game contains an RSE that generates risk measures within specified intervals is decidable and NP-complete for our extreme risk measure, and even PTIME-complete when all players are extreme optimists, while that same problem is undecidable using the entropic risk measure or even the classical expected payoff. This establishes, to our knowledge, the first decidable fragment for equilibria in simple stochastic games without restrictions on strategy types or number of players
PMLR
No full textLocalSGD and SCAFFOLD are widely used methods in distributed stochastic optimization, with numerous applications in machine learning, large-scale data processing, and federated learning. However, rigorously establishing their theoretical advantages over simpler methods, such as minibatch SGD (MbSGD), has proven challenging, as existing analyses often rely on strong assumptions, unrealistic premises, or overly restrictive scenarios.
In this work, we revisit the convergence properties of LocalSGD and SCAFFOLD under a variety of existing or weaker conditions, including gradient similarity, Hessian similarity, weak convexity, and Lipschitz continuity of the Hessian. Our analysis shows that (i) LocalSGD achieves faster convergence compared to MbSGD for weakly convex functions without requiring stronger gradient similarity assumptions; (ii) LocalSGD benefits significantly from higher-order similarity and smoothness; and (iii) SCAFFOLD demonstrates faster convergence than MbSGD for a broader class of non-quadratic functions. These theoretical insights provide a clearer understanding of the conditions under which LocalSGD and SCAFFOLD outperform MbSGD
PMLR
No full textObservational genome-wide association studies are now widely used for causal inference in genetic epidemiology. To maintain privacy, such data is often only publicly available as summary statistics, and often studies for the endogenous covariates and the outcome are available separately. This has necessitated methods tailored to two-sample summary statistics. Current state-of-the-art methods modify linear instrumental variable (IV) regression---with genetic variants as instruments---to account for unmeasured confounding. However, since the endogenous covariates can be high dimensional, standard IV assumptions are generally insufficient to identify all causal effects simultaneously. We ensure identifiability by assuming the causal effects are sparse and propose a sparse causal effect two-sample IV estimator, spaceTSIV, adapting the spaceIV estimator by Pfister and Peters (2022) for two-sample summary statistics. We provide two methods, based on L0- and L1-penalization, respectively. We prove identifiability of the sparse causal effects in the two-sample setting and consistency of spaceTSIV. The performance of spaceTSIV is compared with existing two-sample IV methods in simulations. Finally, we showcase our methods using real proteomic and gene-expression data for drug-target discovery
Machine learning analysis of the factors influencing university-industry collaborations
No full textIn the dynamic arena of innovation, the relations between academia and industry are a keystone for breakthroughs and practical applications. Yet, the groundwork of these pivotal University-Industry (U-I) partnerships remains covered in complexity. This paper delves into these intricate relations, unraveling the factors that help successful collaborations. Grounded in the Resource-Based Theory, our study transcends traditional analytical boundaries, leveraging a neural network model to understand a comprehensive dataset from the UK’s Higher Education Statistics Agency, SCIMAGO Rankings, and Clarivate Publications. This novel approach helps to make clear the interplay of academic load, administrative support, scientific output, and university rank in sculpting U-I collaboration dynamics. Our findings suggest that reduced academic load and robust administrative support significantly bolster U-I collaborations. However, the influence of scientific output and university ranking is more nuanced, challenging the common belief. High scientific output, while indicative of expertise, doesn't always align with industry goals. Similarly, while higher-ranked universities could attract more collaborations, the benefits are not universal. This paper not only contributes to a deeper understanding of U-I collaborations, but also provides actionable insights for university administrators, policymakers, and industry leaders. In a world where innovation is key, understanding these collaborative dynamics is crucial for fostering partnerships that push the boundaries of research and practical application
Hardness of 4-colouring G-colourable graphs
No full textWe study the complexity of a class of promise graph homomorphism problems. For a fixed graph H, the H-colouring problem is to decide whether a given graph has a homomorphism to H. By a result of Hell and Nešetřil, this problem is NP-hard for any non-bipartite loop-less graph H. Brakensiek and Guruswami [SODA 2018] conjectured the hardness extends to promise graph homomorphism problems as follows: fix a pair of non-bipartite loop-less graphs G, H such that there is a homomorphism from G to H, it is NP-hard to distinguish between graphs that are G-colourable and those that are not H-colourable. We confirm this conjecture in the cases when both G and H are 4-colourable. This is a common generalisation of previous results of Khanna, Linial, and Safra [Comb. 20(3): 393-415 (2000)] and of Krokhin and Opršal [FOCS 2019]. The result is obtained by combining the algebraic approach to promise constraint satisfaction with methods of topological combinatorics and equivariant obstruction theory