First Measurement of Hard Exclusive π^{-}Δ^{++} Electroproduction Beam-Spin Asymmetries off the Proton

The polarized cross-section ratio σ_{LT^{'}}/σ_{0} from hard exclusive π^{-}Δ^{++} electroproduction off an unpolarized hydrogen target has been extracted based on beam-spin asymmetry measurements using a 10.2  GeV/10.6  GeV incident electron beam and the CLAS12 spectrometer at Jefferson Lab. The study, which provides the first observation of this channel in the deep-inelastic regime, focuses on very forward-pion kinematics in the valence regime, and photon virtualities ranging from 1.5  GeV^{2} up to 7  GeV^{2}. The reaction provides a novel access to the d-quark content of the nucleon and to p→Δ^{++} transition generalized parton distributions. A comparison to existing results for hard exclusive π^{+}n and π^{0}p electroproduction is provided, which shows a clear impact of the excitation mechanism, encoded in transition generalized parton distributions, on the asymmetry.

Observation of Correlations between Spin and Transverse Momenta in Back-to-Back Dihadron Production at CLAS12

We report the first measurements of deep inelastic scattering spin-dependent azimuthal asymmetries in back-to-back dihadron electroproduction in the deep inelastic scattering process. In this reaction, two hadrons are produced in opposite hemispheres along the z axis in the virtual photon-target nucleon center-of-mass frame, with the first hadron produced in the current-fragmentation region and the second in the target-fragmentation region. The data were taken with longitudinally polarized electron beams of 10.2 and 10.6 GeV incident on an unpolarized liquid-hydrogen target using the CLAS12 spectrometer at Jefferson Lab. Observed nonzero sinΔϕ modulations in ep→e^{'}pπ^{+}X events, where Δϕ is the difference of the azimuthal angles of the proton and pion in the virtual photon and target nucleon center-of-mass frame, indicate that correlations between the spin and transverse momenta of hadrons produced in the target- and current-fragmentation regions may be significant. The measured beam-spin asymmetries provide a first access in dihadron production to a previously unexplored leading-twist spin- and transverse-momentum-dependent fracture function. The fracture functions describe the hadronization of the target remnant after the hard scattering of a virtual photon off a quark in the target particle and provide a new avenue for studying nucleonic structure and hadronization.

Observation of Azimuth-Dependent Suppression of Hadron Pairs in Electron Scattering off Nuclei

We present the first measurement of dihadron angular correlations in electron-nucleus scattering. The data were taken with the CLAS detector and a 5.0 GeV electron beam incident on deuterium, carbon, iron, and lead targets. Relative to deuterium, the nuclear yields of charged-pion pairs show a strong suppression for azimuthally opposite pairs, no suppression for azimuthally nearby pairs, and an enhancement of pairs with large invariant mass. These effects grow with increased nuclear size. The data are qualitatively described by the gibuu model, which suggests that hadrons form near the nuclear surface and undergo multiple scattering in nuclei.These results show that angular correlation studies can open a new way to elucidate how hadrons form and interact inside nuclei.

A self-interpretable module for deep image classification on small data

Deep neural networks are the driving force of the recent explosion of machine learning applications in everyday life. However, they usually require a lot of training data to work well, and they act as black-boxes, making predictions without any explanation about them. This paper presents Memory Wrap, a module (i.e, a set of layers) that can be added to deep learning models to improve their performance and interpretability in settings where few data are available. Memory Wrap adopts a sparse content-attention mechanism between the input and some memories of past training samples. We show that adding Memory Wrap to standard deep neural networks improves their performance when they learn from a limited set of data, and allows them to reach comparable performance when they learn from the full dataset. We discuss how the analysis of its structure and content-attention weights helps to get insights about its decision process and makes their predictions more interpretable, compared to the same networks without Memory Wrap. We test our approach on image classification tasks using several networks on three different datasets, namely CIFAR10, SVHN, and CINIC10.

Outracing champion Gran Turismo drivers with deep reinforcement learning

Many potential applications of artificial intelligence involve making real-time decisions in physical systems while interacting with humans. Automobile racing represents an extreme example of these conditions; drivers must execute complex tactical manoeuvres to pass or block opponents while operating their vehicles at their traction limits¹. Racing simulations, such as the PlayStation game Gran Turismo, faithfully reproduce the non-linear control challenges of real race cars while also encapsulating the complex multi-agent interactions. Here we describe how we trained agents for Gran Turismo that can compete with the world's best e-sports drivers. We combine state-of-the-art, model-free, deep reinforcement learning algorithms with mixed-scenario training to learn an integrated control policy that combines exceptional speed with impressive tactics. In addition, we construct a reward function that enables the agent to be competitive while adhering to racing's important, but under-specified, sportsmanship rules. We demonstrate the capabilities of our agent, Gran Turismo Sophy, by winning a head-to-head competition against four of the world's best Gran Turismo drivers. By describing how we trained championship-level racers, we demonstrate the possibilities and challenges of using these techniques to control complex dynamical systems in domains where agents must respect imprecisely defined human norms.

Agent-Based Markov Modeling for Improved COVID-19 Mitigation Policies

The year 2020 saw the covid-19 virus lead to one of the worst global pandemics in history. As a result, governments around the world have been faced with the challenge of protecting public health while keeping the economy running to the greatest extent possible. Epidemiological models provide insight into the spread of these types of diseases and predict the effects of possible intervention policies. However, to date, even the most data-driven intervention policies rely on heuristics. In this paper, we study how reinforcement learning (RL) and Bayesian inference can be used to optimize mitigation policies that minimize economic impact without overwhelming hospital capacity. Our main contributions are (1) a novel agent-based pandemic simulator which, unlike traditional models, is able to model fine-grained interactions among people at specific locations in a community; (2) an RLbased methodology for optimizing fine-grained mitigation policies within this simulator; and (3) a Hidden Markov Model for predicting infected individuals based on partial observations regarding test results, presence of symptoms, and past physical contacts. This article is part of the special track on AI and COVID-19.

Tauopathy in human and experimental variant Creutzfeldt-Jakob disease

Cerebral accumulation of hyperphosphorylated tau (phospho-tau) occurs in several neurodegenerative conditions including Alzheimer disease. In prion diseases, phospho-tau deposition has been described in a rare genetic form, Gerstmann-Sträussler-Scheinker disease, but is not considered part of the neuropathological picture of Creutzfeldt-Jakob disease. Aim of this study was to investigate whether changes related to phospho-tau accumulation are present in the brain of patients with variant Creutzfeldt-Jakob disease (vCJD) that shares with Gerstmann-Sträussler-Scheinker disease abundant prion protein (PrP) deposition in amyloid form. The analysis was extended to experimental mouse models of vCJD. We detected a large number of phospho-tau-immunoreactive neuritic profiles, often clustered around PrP amyloid deposits, not only in the cerebral cortex, but also in the cerebellum of all vCJD patients examined, in the absence of Abeta. Although less constantly, phospho-tau was localized in some perikaria and dendrites. The biochemical counterpart was the presence of phospho-tau in the detergent-insoluble fraction of cerebral cortex. Phospho-tau-immunoreactive neuronal profiles were also found in association with PrP deposits in mouse models of vCJD. These findings suggest that the abnormal forms of PrP associated with vCJD trigger a tauopathy, and provide a paradigm for the early stages of tau pathology associated with cerebral amyloidoses, including Alzheimer disease.

A novel phenotype of sporadic Creutzfeldt-Jakob disease

An atypical case of sporadic Creutzfeldt-Jakob disease (CJD) is described in a 78-year-old woman homozygous for methionine at codon 129 of the prion protein (PrP) gene. The neuropathological signature was the presence of PrP immunoreactive plaque-like deposits in the cerebral cortex, striatum and thalamus. Western blot analysis showed a profile of the pathological form of PrP (PrP(Sc)) previously unrecognised in sporadic CJD, marked by the absence of diglycosylated protease resistant species. These features define a novel neuropathological and molecular CJD phenotype.

A family with Alzheimer disease and strokes associated with A713T mutation of the APP gene

Three members of an Italian family with autosomal dominant dementia and multiple strokes had the A713T mutation of the APP gene. The neuropathologic examination of the proband disclosed Alzheimer disease (AD) with severe cerebral amyloid angiopathy and multiple infarcts. This indicates that the A713T mutation of the APP gene, lying at the gamma-secretase cleavage site, can be responsible for AD with symptomatic cerebral amyloid angiopathy.

Percutaneous fine-needle biopsy of deep thoracic and abdominal masses in dogs and cats

Percutaneous fine-needle biopsy was used to investigate thoracic and abdominal masses in the dog and cat. One hundred and thirty-two cases were included in the study; 20 cases were excluded from the comparative study due to poor cellularity or blood contamination (retrieval rate 86.8 per cent). One hundred samples (56 dogs and 44 cats) were classified by cytology as neoplastic. All the cytological diagnoses of neoplasia were confirmed by histological samples obtained either by non-surgical methods, at surgery or during postmortem examination. No false positive diagnoses of neoplasia were made. Thirty-two samples were cytologically classified as 'negative for neoplasia'. Subsequent histological examination revealed 18 true negative and 14 false negative results. The procedure had an overall 89.4 per cent (118 cases out of 132) agreement between the diagnosis of inflammatory disease versus neoplasia, with a sensitivity of 87.8 per cent, a specificity of 100 per cent, a predictive value of a positive test of 100 per cent and a predictive value of a negative test of 56.3 per cent.

[Informed consent in the light of the new medical deontologic code and of the state's laws]

The Authors in the light of the legislative directives as well as of some juridical judgements and the new deontological code for medicine, especially from the articles 30 to 35, consider how information should be given to the patients. Furthermore the Authors underline how a correct information does not exclude the explicit consensus expressed by the patient before being submitted to any diagnostic and/or therapeutic treatment and how the patient's consensus should be collected. In addition the Authors briefly examine some juridical judgements that in the last years have been cause of tensions for physician and emphasize some of the fundamental rights of citizen in an up to date view of the medicine.

Analysis of the receiver response in lidar measurements

We report on the calculation of the effective telescope area in lidar applications by a ray-tracing approach. This method allows one to consider the true experimental working conditions and hence to obtain accurate values of the effective telescope area as a function of the height. This in turn allows the retrieval of the signal from the ranges where the overlap function is not constant (e.g., lower ranges), thus increasing the useful range interval. Moreover, we show that the spherical mirrors are more appropriate than the parabolic ones for most of the lidar measurements, although a particular alignment procedure, such as the one we describe, must be used.

Self-aligning lidar for the continuous monitoring of the atmosphere

In a lidar system an accurate transmitter-receiver alignment is essential for correct results. Usually this optical adjustment is time consuming and requires the intervention of highly qualified personnel. As a solution to this problem, a fast and precise automatic alignment procedure is presented, based on a simple model of the transmitter-receiver overlap. The lidar mounted at the Naples University is used to test this method. A centering precision of few microradians is obtained through dedicated software controlling a gimbal-mounted mirror. The automatic alignment procedure is then assessed. In particular, the correctness of the center and of its error is determined. Finally, the system is applied to the monitoring of tropospheric aerosols, leading to the continuous retrieval of profiles with fine spatiotemporal resolution.