405905 research outputs found
Sort by
Processing very long sequences in biological neuronal networks
No full textSensory perception, motor activity generation, language comprehension and production, planning, or solving mathematical equations are sequential processes. Learning, predicting, and recalling sequential data, as well as detecting anomalies in such data streams, are hence fundamental computations performed by the brain. A main challenge in processing such data is context dependency: the correct prediction or recall of an upcoming element in a sequence does not only depend on the previous element, but on the entire history. In [1], we devised an algorithm capable of performing these computations based on a recurrent network of spiking neurons with biophysically interpretable variables and parameters ("spiking Temporal Memory" [sTM]). The sTM algorithm learns complex sequences in a continual, unsupervised manner by means of local synaptic plasticity mechanism known from biology. Prediction and recall of sequence elements are represented by dendritic action potentials. The processing of sequential data in sTM networks is based on ultra-sparse spiking activity, and hence highly energy efficient. So far, the sequence processing capabilities of the sTM algorithm have been demonstrated only for small sequence sets containing few tens of elements. Here, we show that even small sTM networks representing local cortical microcircuits at the sub-millimeter scale can successfully process large sequence sets containing thousands of sequence elements. Mathematical and numerical analysis reveals that the network capacity is primarily limited by the network dynamics in response to unpredicted, ambiguous stimuli. Close to the capacity limit, sTM networks exhibit synapse densities and spiking activity characteristics reminiscent of the neocortex of awake, behaving mammals.References: [1] Bouhadjar, Y., Wouters, D. J., Diesmann, M., & Tetzlaff, T. (2022). Sequence learning, prediction, and replay in networks of spiking neurons. PLOS Computational Biology, 18(6), e1010233. doi:10.1371/journal.pcbi.101023
Transport of Biomass Burning Aerosol into the Extratropical Tropopause Region over Europe via Warm Conveyor Belt Uplift
No full textEmploying normalizing flows to examine neural manifold characteristics and curvatures
No full textDespite the vast number of active neurons, neuronal population activity supposedly lies on low-dimensional manifolds (Gallego et al., 2017). To learn the statistics of neural activity, we use Normalizing Flows (NFs) (Dinh et al., 2014). These neural networks are trained to estimate the probability distribution by learning an invertible map to a latent distribution.We adjust NF’s training objectives to distinguish between relevant and noise dimensions, by using a nested dropout procedure in the latent space (Bekasov & Murray, 2020). An approximation of the network for each mixture component as a quadratic mapping enables us to calculate the Riemannian curvature tensors of the neural manifold. We focus mainly on the directions in the tangent space, in which the sectional curvature shows local extrema.Finally, we apply the method to electrophysiological recordings of the visual cortex in macaques (Chen et al., 2022). We show that manifolds deviate significantly from being flat. Analyzing the curvature of the manifolds yields insights into the regimes where neuron groups interact in a non-linear manner
From cyanobacteria to cell organelle – Engineering and studying minimal endosymbiotic metabolism
No full text
Stirring the false vacuum via interacting quantized bubbles on a 5,564-qubit quantum annealer
No full textFalse vacuum decay—the transition from a metastable quantum state to a true vacuum state—plays an important role in quantum field theory and non-equilibrium phenomena such as phase transitions and dynamical metastability. The non-perturbative nature of false vacuum decay and the limited experimental access to this process make it challenging to study, leaving several open questions regarding how true vacuum bubbles form, move and interact. Here we observe quantized bubble formation in real time, a key feature of false vacuum decay dynamics, using a quantum annealer with 5,564 superconducting flux qubits. We develop an effective model that captures both initial bubble creation and subsequent interactions, and remains accurate under dissipation. The annealer reveals coherent scaling laws in the driven many-body dynamics for more than 1,000 intrinsic qubit time units. This work provides a method for investigating false vacuum dynamics of large quantum systems in quantum annealers
High‐Carbon Amendments Improve Post‐Harvest Nitrogen Retention in Reclaimed Soil: Results of a Laboratory Incubation Study
No full textEffect of molecular adsorption on the conductivity of selectively grown, interconnected 2D-MoS 2 atomically thin flake structures
No full textThe gas sensitivity of field-effect structures with 2D-MoS2 channels selectively grown between Mo electrodes using the Mo-CVD method was investigated by measuring the effect of molecular adsorption from air on the device source-drain current (Isd). The channels were composed of interconnected atomically thin MoS2 grains, with their density and average thickness varied by choosing two different distances (15 and 20 μm) between the Mo contacts. A high response to the tested stimuli, including molecule adsorption, illumination and gate voltage changes, was observed. A significant, persistent photoconduction was induced by positive charge accumulation on traps, most likely at grain boundaries and associated defects. Isd increased under high vacuum, both in the dark and under illumination. The relative dark current response to the transition from air to high vacuum reached up to 1000% at the turn-on voltage. When monitored during the gradual change in air pressure, Isd exhibited a non-monotonic function, sharply peaking at about 10−2 mbar, suggesting molecular adsorption on different defect sites and orientations of adsorbed H2O molecules, which were capable of inducing electron accumulation or depletion. Despite the screening of disorder by extra electrons, the #20 μm sample remained more sensitive to air molecules on its surface. The high vacuum state was also investigated by annealing devices at temperatures up to 340 K in high vacuum, followed by measurements down to 100 K. This revealed thermally stimulated currents and activation energies of trapping electronic states assigned to sulfur vacancies (230 meV) and other shallow levels (85–120 meV), possibly due to natural impurities, grain boundaries or disorder defects. The results demonstrate the high sensitivity of these devices to molecular adsorption, making the technology promising for the easy fabrication of chemical sensors
The Projected Changes in the Surface Energy Budget of the CMIP5 and EURO-CORDEX Models: Are We Heading toward Wetter Growing Seasons in Central Europe?
No full textWe analyze the surface energy budget from four climate model ensembles and its future changes in thetwenty-first century under the RCP8.5 or shared socioeconomic pathway (SSP) 5-8.5 scenario. High-resolution Europeandomain of the Coordinated Regional Climate Downscaling Experiment (EURO-CORDEX) regional climate models(RCMs) and their driving CMIP5 global climate models (CMIP5-D) are first tested in central Europe against observationaldatasets. Evaluation reveals the added value of RCMs in terms of spatial variability and smaller biases. CMIP5-D are af-fected by the positive bias of global irradiance that propagates into other radiation and heat fluxes. There are strong differ-ences in the projected surface energy budget components between RCMs and CMIP5-D. There is an increase in globalirradiance for most of the year in CMIP5-D and other GCM ensembles that is translated into a year-round enhancementof the absorbed solar energy and balanced by higher latent heat flux, except in summer, when the sensible heat flux risesstrongly. Together with strong warming and reduced precipitation in summer, this leads to warm, sunny, and dry conditionswith reduced evapotranspiration and higher drought stress for vegetation. In the RCMs, the reduction in global irradiancedominates, and it is translated into a round-year reduction in the net balance of longwave radiation and stronger latentheat flux. The first months of the growing season show weaker warming associated with higher evapotranspiration and pre-cipitation. In summer, precipitation drops and global irradiance and warming rise, but they fall behind the changes in theGCMs. Compared to GCMs, there are less visible signs of conditions leading to a reduction in evapotranspiration or ashortage of soil water in the RCMs in summer
An Efficient Parallel Implementation of a Hybrid Method for the Advection of High Schmidt Scalars in Flows
No full textWe present here a parallel implementation of a hybrid method combining a high order semi-Lagrangian method for the advection-diffusion of scalars and a finite element method for the Navier-Stokes equations. Then, we introduce an overloading strategy to enhance the computational performance of this method. Finally, we present an application test case of a high Schmidt scalar in a turbulent jet flow