IST Austria: PubRep (Institute of Science and Technology)
Not a member yet
6140 research outputs found
Sort by
Mechanisms of auxin action in plant growth and development
No full textThe phytohormone auxin is a major signal coordinating growth and development in plants. The variety of its effects arises from its ability to form local auxin maxima and gradients within tissues, generated through directional cell-to-cell transport and elaborate metabolic control. These auxin distribution patterns instruct cells in a context-dependent manner to undergo predefined developmental transitions. In this Review, we discuss advances in auxin action at the level of homeostasis and signalling. We highlight key insights into the structural basis of PIN-mediated intercellular auxin transport and explore two novel non-transcriptional auxin signalling mechanisms: one involving intracellular Ca2+ transients and another involving cell-surface auxin perception that mediates global, ultrafast phosphorylation. Furthermore, we examine emerging evidence indicating the involvement of cyclic adenosine monophosphate as a second messenger in the transcriptional auxin response. Together, these recent developments in auxin research have profoundly deepened our understanding of the complex and diverse activities of auxin in plant growth and development
Token-driven totally asymmetric simple exclusion processes
No full textWe consider a family of totally asymmetric simple exclusion processes (TASEPs), consisting of particles on a lattice that require binding by a "token" in various physical configurations to advance over the lattice. Using a combination of theory and simulations, we address the following questions: (i) How token binding kinetics affects the current-density relation on the lattice; (ii) How this current-density relation depends on the scarcity of tokens; (iii) How tokens propagate the effects of the locally-imposed disorder (such as a slow site) over the entire lattice; (iv) How a shared pool of tokens couples concurrent TASEPs running on multiple lattices; (v) How our results translate to TASEPs with open boundaries that exchange particles with the reservoir. Since real particle motion (including in biological systems that inspired the standard TASEP model, e.g., protein synthesis or movement of molecular motors) is often catalyzed, regulated, actuated, or otherwise mediated, the token-driven TASEP dynamics analyzed in this paper should allow for a better understanding of real systems and enable a closer match between TASEP theory and experimental observations
A giant disk galaxy two billion years after the Big Bang
No full textObservational studies have shown that galaxy disks were already in place in the first few billion years of the Universe. The early disks detected so far, with typical half-light radii of 3 kpc at stellar masses around 1011 M⊙ for redshift z ≈ 3, are significantly smaller than today’s disks with similar masses, which is in agreement with expectations from current galaxy models. Here we report observations of a giant disk at z = 3.25, when the Universe was only two billion years old, with a half-light radius of 9.6 kpc and stellar mass of (math formular). This galaxy is larger than any other kinematically confirmed disks at similar epochs and is surprisingly similar to today’s largest disks with regard to size and mass. James Webb Space Telescope imaging and spectroscopy reveal its spiral morphology and a rotational velocity consistent with a local Tully–Fisher relationship. Multiwavelength observations show that it lies in an exceptionally dense environment, where the galaxy number density is more than ten times higher than the cosmic average and mergers are frequent. The discovery of such a giant disk suggests the presence of favourable physical conditions for large-disk formation in dense environments in the early Universe, which may include efficient accretion of gas carrying coherent angular momentum and non-destructive mergers between exceptionally gas-rich progenitor galaxies
Spatiotemporal switches in progenitor cell fate govern upper hair follicle growth and maintenance
No full textThe epidermis provides a protective barrier against hostile environments. However, our knowledge of how this barrier forms during development and is subsequently maintained remains incomplete. The infundibulum is a cylindrical epidermal tissue compartment that serves as an outlet for hair follicles protruding from the skin and the excretion of the sebaceous glands that are essential for proper skin function. In this study, we applied quantitative fate mapping to address how infundibulum are maintained during adulthood. We demonstrate that progenitors build and maintain tissues through stochastic cell fate choices. Long-term analysis identified a preferential transient contribution from cells initially located at the bottom of the structure to the maintenance of the tissue, with bursts of local progenitor expansion associated with the phases of hair growth. Beyond providing compartment-wide insights into progenitor cell dynamics in infundibulum, these findings demonstrate how spatiotemporal regulation controls transient progenitor dominance
MARLIN: Mixed-precision auto-regressive parallel inference on Large Language Models
No full textAs inference on Large Language Models (LLMs) emerges as an important workload in machine learning applications, model weight quantization has become a standard technique for efficient GPU deployment. Quantization not only reduces model size, but has also been shown to yield substantial speedups for single-user inference, due to reduced memory movement, with low accuracy impact. Yet, it remains a key open question whether speedups are achievable also in batched settings with multiple parallel clients, which are highly relevant for practical serving. It is unclear whether GPU kernels can be designed to remain practically memory-bound, while supporting the substantially increased compute requirements of batched workloads.
In this paper, we resolve this question positively by introducing a new design for Mixed-precision Auto-Regressive LINear kernels, called MARLIN. Concretely, given a model whose weights are compressed via quantization to, e.g., 4 bits per element, MARLIN shows that batchsizes up to 16-32 can be practically supported with close to maximum (4×) quantization speedup, and larger batchsizes up to 64-128 with gradually decreasing, but still significant, acceleration. MARLIN accomplishes this via a combination of techniques, such as asynchronous memory access, complex task scheduling and pipelining, and bespoke quantization support. Our experiments show that MARLIN's near-optimal performance on individual LLM layers across different scenarios can also lead to significant end-to-end LLM inference speedups (of up to 2.8×) when integrated with the popular vLLM open-source serving engine. Finally, we show that MARLIN is extensible to further compression techniques, like NVIDIA 2:4 sparsity, leading to additional speedups
Method dependence in thermal conductivity and aerodynamic roughness length estimates on a debris‐covered glacier
No full textRock debris partially covers glaciers worldwide, with varying extents and distributions, and controls sub‐debris melt rates by modifying energy transfer from the atmosphere to the ice. Two key physical properties controlling this energy exchange are thermal conductivity (k) and aerodynamic roughness length (z0). Accurate representation of these properties in energy‐balance models is critical for understanding climate‐glacier interactions and predicting the behavior of debris‐covered glaciers. However, k and z0 have been derived at very few sites from limited local measurements, using different approaches, and most model applications rely on values reported from these few sites and studies. We derive k and z0 using established and modified approaches from data at three locations on Pirámide Glacier in the central Chilean Andes. By comparing methods and evaluating melt simulated with an energy‐balance model, we reveal substantial differences between approaches. These lead to discrepancies between ice melt from energy‐balance simulations and observed data, and highlight the impact of method choice on calculated ice melt. Optimizing k against measured melt appears a viable approach to constrain melt simulations. Determining z0 seems less critical, as it has a smaller impact on total melt. Profile aerodynamic method measurements for estimating z0, despite higher costs, are independent of ice melt calculations. The large, unexpected differences between methods indicate a substantial knowledge gap. The fact that field‐derived k and z0 fail to work well in energy‐balance models, suggests that model values represent bulk properties distinct from theoretical field measurements. Addressing this gap is essential for improving glacier melt predictions
Enhanced mass transfer in microgeometry using pulsating velocity inputs: Hydrodynamic analysis and numerical simulation
No full textThe current work focuses on the performance of hydrodynamics and mass transfer in a microchannel. A hydrodynamic model is developed for a gas–liquid (CO2–water) system and slug flow pattern. For the first time in literature, a concept of pulsating velocity input is introduced in an enhanced cross-T-junction microchannel to study the mass transfer using the physical absorption mechanism in ANSYS FLUENT R2 2024. The mass transfer model is associated with the hydrodynamic model and some user-defined functions in FLUENT. This work demonstrates that incorporating obstructions and applying trapezoidal and sinusoidal wave inputs improve the CO2 absorption rate. The obtained data are further compared with the plain T-junction microchannel in terms of mass transfer coefficient. Solubility of CO2 in three different solvents (ethyl alcohol, water, and ethylene glycol) has been revealed in an enhanced cross T-junction microchannel at two different temperatures, i.e., 298.15 and 303.15 K. The numerical simulations illustrate that an increase in temperature has an adverse effect on the mass transfer rate
Design of facilitated dissociation enables timing of cytokine signalling
No full textProtein design has focused on the design of ground states, ensuring that they are sufficiently low energy to be highly populated1. Designing the kinetics and dynamics of a system requires, in addition, the design of excited states that are traversed in transitions from one low-lying state to another2,3. This is a challenging task because such states must be sufficiently strained to be poorly populated, but not so strained that they are not populated at all, and because protein design methods have focused on generating near-ideal structures4,5,6,7. Here we describe a general approach for designing systems that use an induced-fit power stroke8 to generate a structurally frustrated9 and strained excited state, allosterically driving protein complex dissociation. X-ray crystallography, double electron–electron resonance spectroscopy and kinetic binding measurements show that incorporating excited states enables the design of effector-induced increases in dissociation rates as high as 5,700-fold. We highlight the power of this approach by designing rapid biosensors, kinetically controlled circuits and cytokine mimics that can be dissociated from their receptors within seconds, enabling dissection of the temporal dynamics of interleukin-2 signalling
On the existence of magic squares of powers
No full textFor any d 2, we prove that there exists an integer n0(d) such that there exists an n × n
magic square of dth powers for all n n0(d). In particular, we establish the existence of
an n × n magic square of squares for all n 4, which settles a conjecture of
Várilly-Alvarado. All previous approaches had been based on constructive methods and
the existence of n × n magic squares of dth powers had only been known for sparse
values of n. We prove our result by the Hardy-Littlewood circle method, which in this
setting essentially reduces the problem to finding a sufficient number of disjoint linearly
independent subsets of the columns of the coefficient matrix of the equations defining
magic squares. We prove an optimal (up to a constant) lower bound for this quantity
LIPIcs
No full textMany differentially private and classical non-private graph algorithms rely crucially on determining whether some property of each vertex meets a threshold. For example, for the k-core decomposition problem, the classic peeling algorithm iteratively removes a vertex if its induced degree falls below a threshold. The sparse vector technique (SVT) is generally used to transform non-private threshold queries into private ones with only a small additive loss in accuracy. However, a naive application of SVT in the graph setting leads to an amplification of the error by a factor of n due to composition, as SVT is applied to every vertex. In this paper, we resolve this problem by formulating a novel generalized sparse vector technique which we call the Multidimensional AboveThreshold (MAT) Mechanism which generalizes SVT (applied to vectors with one dimension) to vectors with multiple dimensions. When applied to vectors with n dimensions, we solve a number of important graph problems with better bounds than previous work.
Specifically, we apply our MAT mechanism to obtain a set of improved bounds for a variety of problems including k-core decomposition, densest subgraph, low out-degree ordering, and vertex coloring. We give a tight local edge differentially private (LEDP) algorithm for k-core decomposition that results in an approximation with O(ε^{-1} log n) additive error and no multiplicative error in O(n) rounds. We also give a new (2+η)-factor multiplicative, O(ε^{-1} log n) additive error algorithm in O(log² n) rounds for any constant η > 0. Both of these results are asymptotically tight against our new lower bound of Ω(log n) for any constant-factor approximation algorithm for k-core decomposition. Our new algorithms for k-core decomposition also directly lead to new algorithms for the related problems of densest subgraph and low out-degree ordering. Finally, we give novel LEDP differentially private defective coloring algorithms that use number of colors given in terms of the arboricity of the graph