Search

Empirically observed time series in physics, biology, or medicine, are commonly generated by some underlying dynamical system (DS) which is the target of scientific interest. There is an increasing interest to harvest machine learning methods to reconstruct this latent DS in a data-driven, unsupervised way. In many areas of science it is common to sample time series observations from many data modalities simultaneously, e.g. electrophysiological and behavioral time series in a typical neuroscience experiment. However, current machine learning tools for reconstructing DSs usually focus on just one data modality. Here we propose a general framework for multi-modal data integration for the purpose of nonlinear DS reconstruction and the analysis of cross-modal relations. This framework is based on dynamically interpretable recurrent neural networks as general approximators of nonlinear DSs, coupled to sets of modality-specific decoder models from the class of generalized linear models. Both an expectation-maximization and a variational inference algorithm for model training are advanced and compared. We show on nonlinear DS benchmarks that our algorithms can efficiently compensate for too noisy or missing information in one data channel by exploiting other channels, and demonstrate on experimental neuroscience data how the algorithm learns to link different data domains to the underlying dynamics., Comment: 19 pages, 8 figures

In this paper, we present a model to predict the spread of the Covid-19 epidemic and apply it to the specific case of Italy. We started from a simple Susceptible, Infected, Recovered (SIR) model and we added the condition that, after a certain time, the basic reproduction number $R_0$ exponentially decays in time, as empirically suggested by world data. Using this model, we were able to reproduce the real behavior of the epidemic with an average error of 5\%. Moreover, we illustrate possible future scenarios, associated to different intervals of $R_0$. This model has been used since the beginning of March 2020, predicting the Italian peak of the epidemic in April 2020 with about 100.000 detected active cases. The real peak of the epidemic happened on the 20th of April 2020, with 108.000 active cases. This result shows that the model had predictive power for the italian case., Comment: The model presented in this paper has been adopted on Covstat.it. Errata corrige in the abstract

A series of Silicon crystal undulator samples were produced based on the approach presented in PRL 90 (2003) 034801, with the periods of undulation from 0.1 mm to 1 mm, and the number of periods on the order of 10. The samples were characterized by X-rays, revealing the sine-like shape of the crystal lattice in the bulk. Next step in the characterization has been the channeling tests done with 70 GeV protons, where good channeling properties of the undulated Silicon lattice have been observed. The photon radiation tests of crystal undulators with high energy positrons are in progress on several locations: IHEP Protvino, LNF Frascati, and CERN SPS. The progress in the experimental activities and the predictions from detailed simulations are reported., Comment: Invited talk at the NATO ARW "Advanced Photon Sources and Their Application" (Nor Hamberd, Armenia August 29 – September 02, 2004)

A crystalline undulator (CU) with periodically deformed crystallographic planes is capable of deflecting charged particles with the same strength as an equivalent magnetic field of 1000 T and could provide quite a short period L in the sub-millimeter range. We present an idea for creation of a CU and report its first realization. One face of a silicon crystal was given periodic micro-scratches (grooves), with a period of 1 mm, by means of a diamond blade. The X-ray tests of the crystal deformation have shown that a sinusoidal-like shape of crystalline planes goes through the bulk of the crystal. This opens up the possibility for experiments with high-energy particles channeled in CU, a novel compact source of radiation. The first experiment on photon emission in CU has been started at LNF with 800 MeV positrons aiming to produce 50 keV undulator photons., Comment: Presented at PAC 2003 (Portland, May 12-16)