Showing 1–40 of 40 results for author: Boschini, M

Selective Attention-based Modulation for Continual Learning

Authors: Giovanni Bellitto, Federica Proietto Salanitri, Matteo Pennisi, Matteo Boschini, Angelo Porrello, Simone Calderara, Simone Palazzo, Concetto Spampinato

Abstract: We present SAM, a biologically-plausible selective attention-driven modulation approach to enhance classification models in a continual learning setting. Inspired by neurophysiological evidence that the primary visual cortex does not contribute to object manifold untangling for categorization and that primordial attention biases are still embedded in the modern brain, we propose to employ auxiliar… ▽ More We present SAM, a biologically-plausible selective attention-driven modulation approach to enhance classification models in a continual learning setting. Inspired by neurophysiological evidence that the primary visual cortex does not contribute to object manifold untangling for categorization and that primordial attention biases are still embedded in the modern brain, we propose to employ auxiliary saliency prediction features as a modulation signal to drive and stabilize the learning of a sequence of non-i.i.d. classification tasks. Experimental results confirm that SAM effectively enhances the performance (in some cases up to about twenty percent points) of state-of-the-art continual learning methods, both in class-incremental and task-incremental settings. Moreover, we show that attention-based modulation successfully encourages the learning of features that are more robust to the presence of spurious features and to adversarial attacks than baseline methods. Code is available at: https://github.com/perceivelab/SAM. △ Less

Submitted 29 March, 2024; originally announced March 2024.

Semantic Residual Prompts for Continual Learning

Authors: Martin Menabue, Emanuele Frascaroli, Matteo Boschini, Enver Sangineto, Lorenzo Bonicelli, Angelo Porrello, Simone Calderara

Abstract: Prompt-tuning methods for Continual Learning (CL) freeze a large pre-trained model and focus training on a few parameter vectors termed prompts. Most of these methods organize these vectors in a pool of key-value pairs, and use the input image as query to retrieve the prompts (values). However, as keys are learned while tasks progress, the prompting selection strategy is itself subject to catastro… ▽ More Prompt-tuning methods for Continual Learning (CL) freeze a large pre-trained model and focus training on a few parameter vectors termed prompts. Most of these methods organize these vectors in a pool of key-value pairs, and use the input image as query to retrieve the prompts (values). However, as keys are learned while tasks progress, the prompting selection strategy is itself subject to catastrophic forgetting, an issue often overlooked by existing approaches. For instance, prompts introduced to accommodate new tasks might end up interfering with previously learned prompts. To make the selection strategy more stable, we ask a foundational model (CLIP) to select our prompt within a two-level adaptation mechanism. Specifically, the first level leverages standard textual prompts for the CLIP textual encoder, leading to stable class prototypes. The second level, instead, uses these prototypes along with the query image as keys to index a second pool. The retrieved prompts serve to adapt a pre-trained ViT, granting plasticity. In doing so, we also propose a novel residual mechanism to transfer CLIP semantics to the ViT layers. Through extensive analysis on established CL benchmarks, we show that our method significantly outperforms both state-of-the-art CL approaches and the zero-shot CLIP test. Notably, our findings hold true even for datasets with a substantial domain gap w.r.t. the pre-training knowledge of the backbone model, as showcased by experiments on satellite imagery and medical datasets. △ Less

Submitted 14 March, 2024; v1 submitted 11 March, 2024; originally announced March 2024.

Comments: 23 pages, 5 figures

Astrophysical and relativistic modeling of the recoiling black-hole candidate in quasar 3C 186

Authors: Matteo Boschini, Davide Gerosa, Om Sharan Salafia, Massimo Dotti

Abstract: The compact object in quasar 3C 186 is one of the most promising recoiling black-hole candidates, exhibiting both an astrometric displacement between the quasar and the host galaxy as well as a spectroscopic shift between broad and narrow lines. 3C 186 also presents a radio jet which, when projected onto the plane of the sky, appears to be perpendicular to the quasar/galaxy displacement. Assuming… ▽ More The compact object in quasar 3C 186 is one of the most promising recoiling black-hole candidates, exhibiting both an astrometric displacement between the quasar and the host galaxy as well as a spectroscopic shift between broad and narrow lines. 3C 186 also presents a radio jet which, when projected onto the plane of the sky, appears to be perpendicular to the quasar/galaxy displacement. Assuming a gravitational-wave kick is indeed responsible for the properties of 3C 186 and using state-of-the-art relativistic modeling, we show that current observations allow for exquisite modeling of the recoiling black hole. Most notably, we find that the kick velocity, the black-hole spin, and the line of sight are almost collinear and the former appear perpendicular to each other only because of a strong projection effect. The targeted configuration requires substantial fine-tuning: while there exists a region in the black-hole binary parameter space that is compatible with 3C 186, the observed system appears to be a rare occurrence. Using archival radio observations, we explore different strategies that could potentially confirm or rule out our interpretation. In particular, we develop two observational tests that rely on the brightness ratio between the approaching and receding jet as well as the asymmetry of the jet lobes. While the available radio data provide loose constraints, deeper observations have the unique potential of unveiling the nature of 3C 186. △ Less

Submitted 21 June, 2024; v1 submitted 13 February, 2024; originally announced February 2024.

Comments: 12 pages, 8 figures

Journal ref: A&A 686, A245 (2024)

Extending black-hole remnant surrogate models to extreme mass ratios

Authors: Matteo Boschini, Davide Gerosa, Vijay Varma, Cristobal Armaza, Michael Boyle, Marceline S. Bonilla, Andrea Ceja, Yitian Chen, Nils Deppe, Matthew Giesler, Lawrence E. Kidder, Prayush Kumar, Guillermo Lara, Oliver Long, Sizheng Ma, Keefe Mitman, Peter James Nee, Harald P. Pfeiffer, Antoni Ramos-Buades, Mark A. Scheel, Nils L. Vu, Jooheon Yoo

Abstract: Numerical-relativity surrogate models for both black-hole merger waveforms and remnants have emerged as important tools in gravitational-wave astronomy. While producing very accurate predictions, their applicability is limited to the region of the parameter space where numerical-relativity simulations are available and computationally feasible. Notably, this excludes extreme mass ratios. We presen… ▽ More Numerical-relativity surrogate models for both black-hole merger waveforms and remnants have emerged as important tools in gravitational-wave astronomy. While producing very accurate predictions, their applicability is limited to the region of the parameter space where numerical-relativity simulations are available and computationally feasible. Notably, this excludes extreme mass ratios. We present a machine-learning approach to extend the validity of existing and future numerical-relativity surrogate models toward the test-particle limit, targeting in particular the mass and spin of post-merger black-hole remnants. Our model is trained on both numerical-relativity simulations at comparable masses and analytical predictions at extreme mass ratios. We extend the gaussian-process-regression model NRSur7dq4Remnant, validate its performance via cross validation, and test its accuracy against additional numerical-relativity runs. Our fit, which we dub NRSur7dq4EmriRemnant, reaches an accuracy that is comparable to or higher than that of existing remnant models while providing robust predictions for arbitrary mass ratios. △ Less

Submitted 24 October, 2023; v1 submitted 7 July, 2023; originally announced July 2023.

Comments: 10 pages, 3 figures. Published in PRD. Model publicly available at https://pypi.org/project/surfinBH

Journal ref: Phys.Rev.D 108 (2023) 8, 084015

On the Effectiveness of Equivariant Regularization for Robust Online Continual Learning

Authors: Lorenzo Bonicelli, Matteo Boschini, Emanuele Frascaroli, Angelo Porrello, Matteo Pennisi, Giovanni Bellitto, Simone Palazzo, Concetto Spampinato, Simone Calderara

Abstract: Humans can learn incrementally, whereas neural networks forget previously acquired information catastrophically. Continual Learning (CL) approaches seek to bridge this gap by facilitating the transfer of knowledge to both previous tasks (backward transfer) and future ones (forward transfer) during training. Recent research has shown that self-supervision can produce versatile models that can gen… ▽ More Humans can learn incrementally, whereas neural networks forget previously acquired information catastrophically. Continual Learning (CL) approaches seek to bridge this gap by facilitating the transfer of knowledge to both previous tasks (backward transfer) and future ones (forward transfer) during training. Recent research has shown that self-supervision can produce versatile models that can generalize well to diverse downstream tasks. However, contrastive self-supervised learning (CSSL), a popular self-supervision technique, has limited effectiveness in online CL (OCL). OCL only permits one iteration of the input dataset, and CSSL's low sample efficiency hinders its use on the input data-stream. In this work, we propose Continual Learning via Equivariant Regularization (CLER), an OCL approach that leverages equivariant tasks for self-supervision, avoiding CSSL's limitations. Our method represents the first attempt at combining equivariant knowledge with CL and can be easily integrated with existing OCL methods. Extensive ablations shed light on how equivariant pretext tasks affect the network's information flow and its impact on CL dynamics. △ Less

Submitted 5 May, 2023; originally announced May 2023.

Comments: 10 pages, 4 figures

Latent Spectral Regularization for Continual Learning

Authors: Emanuele Frascaroli, Riccardo Benaglia, Matteo Boschini, Luca Moschella, Cosimo Fiorini, Emanuele Rodolà, Simone Calderara

Abstract: While biological intelligence grows organically as new knowledge is gathered throughout life, Artificial Neural Networks forget catastrophically whenever they face a changing training data distribution. Rehearsal-based Continual Learning (CL) approaches have been established as a versatile and reliable solution to overcome this limitation; however, sudden input disruptions and memory constraints a… ▽ More While biological intelligence grows organically as new knowledge is gathered throughout life, Artificial Neural Networks forget catastrophically whenever they face a changing training data distribution. Rehearsal-based Continual Learning (CL) approaches have been established as a versatile and reliable solution to overcome this limitation; however, sudden input disruptions and memory constraints are known to alter the consistency of their predictions. We study this phenomenon by investigating the geometric characteristics of the learner's latent space and find that replayed data points of different classes increasingly mix up, interfering with classification. Hence, we propose a geometric regularizer that enforces weak requirements on the Laplacian spectrum of the latent space, promoting a partitioning behavior. We show that our proposal, called Continual Spectral Regularizer (CaSpeR), can be easily combined with any rehearsal-based CL approach and improves the performance of SOTA methods on standard benchmarks. Finally, we conduct additional analysis to provide insights into CaSpeR's effects and applicability. △ Less

Submitted 7 August, 2023; v1 submitted 9 January, 2023; originally announced January 2023.

Comments: 8 pages, 3 figures

On the Effectiveness of Lipschitz-Driven Rehearsal in Continual Learning

Authors: Lorenzo Bonicelli, Matteo Boschini, Angelo Porrello, Concetto Spampinato, Simone Calderara

Abstract: Rehearsal approaches enjoy immense popularity with Continual Learning (CL) practitioners. These methods collect samples from previously encountered data distributions in a small memory buffer; subsequently, they repeatedly optimize on the latter to prevent catastrophic forgetting. This work draws attention to a hidden pitfall of this widespread practice: repeated optimization on a small pool of da… ▽ More Rehearsal approaches enjoy immense popularity with Continual Learning (CL) practitioners. These methods collect samples from previously encountered data distributions in a small memory buffer; subsequently, they repeatedly optimize on the latter to prevent catastrophic forgetting. This work draws attention to a hidden pitfall of this widespread practice: repeated optimization on a small pool of data inevitably leads to tight and unstable decision boundaries, which are a major hindrance to generalization. To address this issue, we propose Lipschitz-DrivEn Rehearsal (LiDER), a surrogate objective that induces smoothness in the backbone network by constraining its layer-wise Lipschitz constants w.r.t. replay examples. By means of extensive experiments, we show that applying LiDER delivers a stable performance gain to several state-of-the-art rehearsal CL methods across multiple datasets, both in the presence and absence of pre-training. Through additional ablative experiments, we highlight peculiar aspects of buffer overfitting in CL and better characterize the effect produced by LiDER. Code is available at https://github.com/aimagelab/LiDER △ Less

Submitted 16 October, 2022; v1 submitted 12 October, 2022; originally announced October 2022.

Comments: 22 pages, 8 figures, typos corrected. Accepted at the thirty-sixth Conference on Neural Information Processing Systems (NeurIPS 2022), New Orleans, US

Effects of Auxiliary Knowledge on Continual Learning

Authors: Giovanni Bellitto, Matteo Pennisi, Simone Palazzo, Lorenzo Bonicelli, Matteo Boschini, Simone Calderara, Concetto Spampinato

Abstract: In Continual Learning (CL), a neural network is trained on a stream of data whose distribution changes over time. In this context, the main problem is how to learn new information without forgetting old knowledge (i.e., Catastrophic Forgetting). Most existing CL approaches focus on finding solutions to preserve acquired knowledge, so working on the past of the model. However, we argue that as the… ▽ More In Continual Learning (CL), a neural network is trained on a stream of data whose distribution changes over time. In this context, the main problem is how to learn new information without forgetting old knowledge (i.e., Catastrophic Forgetting). Most existing CL approaches focus on finding solutions to preserve acquired knowledge, so working on the past of the model. However, we argue that as the model has to continually learn new tasks, it is also important to put focus on the present knowledge that could improve following tasks learning. In this paper we propose a new, simple, CL algorithm that focuses on solving the current task in a way that might facilitate the learning of the next ones. More specifically, our approach combines the main data stream with a secondary, diverse and uncorrelated stream, from which the network can draw auxiliary knowledge. This helps the model from different perspectives, since auxiliary data may contain useful features for the current and the next tasks and incoming task classes can be mapped onto auxiliary classes. Furthermore, the addition of data to the current task is implicitly making the classifier more robust as we are forcing the extraction of more discriminative features. Our method can outperform existing state-of-the-art models on the most common CL Image Classification benchmarks. △ Less

Submitted 3 June, 2022; originally announced June 2022.

Transfer without Forgetting

Authors: Matteo Boschini, Lorenzo Bonicelli, Angelo Porrello, Giovanni Bellitto, Matteo Pennisi, Simone Palazzo, Concetto Spampinato, Simone Calderara

Abstract: This work investigates the entanglement between Continual Learning (CL) and Transfer Learning (TL). In particular, we shed light on the widespread application of network pretraining, highlighting that it is itself subject to catastrophic forgetting. Unfortunately, this issue leads to the under-exploitation of knowledge transfer during later tasks. On this ground, we propose Transfer without Forget… ▽ More This work investigates the entanglement between Continual Learning (CL) and Transfer Learning (TL). In particular, we shed light on the widespread application of network pretraining, highlighting that it is itself subject to catastrophic forgetting. Unfortunately, this issue leads to the under-exploitation of knowledge transfer during later tasks. On this ground, we propose Transfer without Forgetting (TwF), a hybrid approach building upon a fixed pretrained sibling network, which continuously propagates the knowledge inherent in the source domain through a layer-wise loss term. Our experiments indicate that TwF steadily outperforms other CL methods across a variety of settings, averaging a 4.81% gain in Class-Incremental accuracy over a variety of datasets and different buffer sizes. △ Less

Submitted 25 July, 2022; v1 submitted 1 June, 2022; originally announced June 2022.

Comments: 22 pages, 3 Figures. Accepted at 17th European Conference on Computer Vision (ECCV 2022), Tel Aviv, Israel

Spectra of Cosmic Ray Sodium and Aluminum and Unexpected Aluminum Excess

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. Johannesson, G. La Vacca, N. Masi, I. V. Moskalenko, S. Pensotti, T. A. Porter, L. Quadrani, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: Since its launch, the Alpha Magnetic Spectrometer-02 (AMS-02) has delivered outstanding quality measurements of the spectra of cosmic-ray (CR) species, $\bar{p}$, $e^{\pm}$, and nuclei (H-Si, Fe), which resulted in a number of breakthroughs. The most recent AMS-02 result is the measurement of the spectra of CR sodium and aluminum up to $\sim$2 TV. Given their low solar system abundances, a signifi… ▽ More Since its launch, the Alpha Magnetic Spectrometer-02 (AMS-02) has delivered outstanding quality measurements of the spectra of cosmic-ray (CR) species, $\bar{p}$, $e^{\pm}$, and nuclei (H-Si, Fe), which resulted in a number of breakthroughs. The most recent AMS-02 result is the measurement of the spectra of CR sodium and aluminum up to $\sim$2 TV. Given their low solar system abundances, a significant fraction of each element is produced in fragmentations of heavier species, predominantly Ne, Mg, and Si. In this paper, we use precise measurements of the sodium and aluminum spectra by AMS-02 together with ACE-CRIS and Voyager 1 data to test their origin. We show that the sodium spectrum agrees well with the predictions made with the GalProp-HelMod framework, while aluminum spectrum shows a significant excess in the rigidity range from 2-7 GV. In this context, we discuss the origin of other low-energy excesses in Li, F, and Fe found earlier. The observed excesses in Li, F, and Al appear to be consistent with the local Wolf-Rayet (WR) stars hypothesis, invoked to reproduce anomalous $^{22}$Ne/$^{20}$Ne, $^{12}$C/$^{16}$O, and $^{58}$Fe/$^{56}$Fe ratios in CRs, while excess in Fe is likely connected with a past SN activity in the solar neighborhood. We also provide updated local interstellar spectra (LIS) of sodium and aluminum in the rigidity range from few MV to $\sim$2 TV. Our calculations employ the self-consistent GalProp-HelMod framework that has proved to be a reliable tool in deriving the LIS of CR $\bar{p}$, $e^{-}$, and nuclei $Z\le28$. △ Less

Submitted 16 May, 2022; v1 submitted 20 February, 2022; originally announced February 2022.

Comments: 19 pages, 16 figures, 11 tables; ApJ in press

Journal ref: ApJ 933 (2022) 147

Class-Incremental Continual Learning into the eXtended DER-verse

Authors: Matteo Boschini, Lorenzo Bonicelli, Pietro Buzzega, Angelo Porrello, Simone Calderara

Abstract: The staple of human intelligence is the capability of acquiring knowledge in a continuous fashion. In stark contrast, Deep Networks forget catastrophically and, for this reason, the sub-field of Class-Incremental Continual Learning fosters methods that learn a sequence of tasks incrementally, blending sequentially-gained knowledge into a comprehensive prediction. This work aims at assessing and… ▽ More The staple of human intelligence is the capability of acquiring knowledge in a continuous fashion. In stark contrast, Deep Networks forget catastrophically and, for this reason, the sub-field of Class-Incremental Continual Learning fosters methods that learn a sequence of tasks incrementally, blending sequentially-gained knowledge into a comprehensive prediction. This work aims at assessing and overcoming the pitfalls of our previous proposal Dark Experience Replay (DER), a simple and effective approach that combines rehearsal and Knowledge Distillation. Inspired by the way our minds constantly rewrite past recollections and set expectations for the future, we endow our model with the abilities to i) revise its replay memory to welcome novel information regarding past data ii) pave the way for learning yet unseen classes. We show that the application of these strategies leads to remarkable improvements; indeed, the resulting method - termed eXtended-DER (X-DER) - outperforms the state of the art on both standard benchmarks (such as CIFAR-100 and miniImagenet) and a novel one here introduced. To gain a better understanding, we further provide extensive ablation studies that corroborate and extend the findings of our previous research (e.g. the value of Knowledge Distillation and flatter minima in continual learning setups). △ Less

Submitted 19 September, 2022; v1 submitted 3 January, 2022; originally announced January 2022.

Comments: 23 pages, 22 figures. To appear in IEEE TPAMI

Continual Semi-Supervised Learning through Contrastive Interpolation Consistency

Authors: Matteo Boschini, Pietro Buzzega, Lorenzo Bonicelli, Angelo Porrello, Simone Calderara

Abstract: Continual Learning (CL) investigates how to train Deep Networks on a stream of tasks without incurring forgetting. CL settings proposed in literature assume that every incoming example is paired with ground-truth annotations. However, this clashes with many real-world applications: gathering labeled data, which is in itself tedious and expensive, becomes infeasible when data flow as a stream. This… ▽ More Continual Learning (CL) investigates how to train Deep Networks on a stream of tasks without incurring forgetting. CL settings proposed in literature assume that every incoming example is paired with ground-truth annotations. However, this clashes with many real-world applications: gathering labeled data, which is in itself tedious and expensive, becomes infeasible when data flow as a stream. This work explores Continual Semi-Supervised Learning (CSSL): here, only a small fraction of labeled input examples are shown to the learner. We assess how current CL methods (e.g.: EWC, LwF, iCaRL, ER, GDumb, DER) perform in this novel and challenging scenario, where overfitting entangles forgetting. Subsequently, we design a novel CSSL method that exploits metric learning and consistency regularization to leverage unlabeled examples while learning. We show that our proposal exhibits higher resilience to diminishing supervision and, even more surprisingly, relying only on 25% supervision suffices to outperform SOTA methods trained under full supervision. △ Less

Submitted 29 August, 2022; v1 submitted 14 August, 2021; originally announced August 2021.

Comments: 7 pages, 2 figures, to appear in Pattern Recognition Letters, Volume 162, October 2022, Pages 9-14

Journal ref: Pattern Recognition Letters, Volume 162, October 2022, Pages 9-14, ISSN 0167-8655

A Hint of a Low-Energy Excess in Cosmic-Ray Fluorine

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. Johannesson, G. La Vacca, N. Masi, I. V. Moskalenko, S. Pensotti, T. A. Porter, L. Quadrani, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: Since its launch, the Alpha Magnetic Spectrometer-02 (AMS-02) has delivered outstanding quality measurements of the spectra of cosmic-ray (CR) species, $\bar{p}$, $e^{\pm}$, and nuclei (H--O, Ne, Mg, Si, Fe), which resulted in a number of breakthroughs. The most recent AMS-02 result is the measurement of the spectrum of CR fluorine up to $\sim$2 TV. Given its very low solar system abundance, fluor… ▽ More Since its launch, the Alpha Magnetic Spectrometer-02 (AMS-02) has delivered outstanding quality measurements of the spectra of cosmic-ray (CR) species, $\bar{p}$, $e^{\pm}$, and nuclei (H--O, Ne, Mg, Si, Fe), which resulted in a number of breakthroughs. The most recent AMS-02 result is the measurement of the spectrum of CR fluorine up to $\sim$2 TV. Given its very low solar system abundance, fluorine in CRs is thought to be mostly secondary, produced in fragmentations of heavier species, predominantly Ne, Mg, and Si. Similar to the best-measured secondary-to-primary boron to carbon nuclei ratio that is widely used to study the origin and propagation of CR species, the precise fluorine data would allow the origin of Si-group nuclei to be studied independently. Meanwhile, the secondary origin of CR fluorine has never been tested in a wide energy range due to the lack of accurate CR data. In this paper, we use the first ever precise measurements of the fluorine spectrum by AMS-02 together with ACE-CRIS and Voyager 1 data to actually test this paradigm. Our detailed modeling shows an excess below 10 GV in the fluorine spectrum that may hint at a primary fluorine component. We also provide an updated local interstellar spectrum (LIS) of fluorine in the rigidity range from few MV to $\sim$2 TV. Our calculations employ the self-consistent GalProp-HelMod framework that has proved to be a reliable tool in deriving the LIS of CR $\bar{p}$, $e^{-}$, and nuclei $Z\le28$. △ Less

Submitted 11 October, 2021; v1 submitted 3 June, 2021; originally announced June 2021.

Comments: 14 pages, 10 figures, 8 tables; ApJ in press

Journal ref: ApJ 925 (2022) 108

A discovery of a low-energy excess in cosmic-ray iron: an evidence of the past supernova activity in the Local Bubble

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. Johannesson, G. La Vacca, N. Masi, I. V. Moskalenko, S. Pensotti, T. A. Porter, L. Quadrani, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: Since its launch, the Alpha Magnetic Spectrometer - 02 (AMS-02) has delivered outstanding quality measurements of the spectra of cosmic-ray (CR) species, $\bar{p}$, $e^{\pm}$, and nuclei, $_1$H-$_8$O, $_{10}$Ne, $_{12}$Mg, $_{14}$Si, which resulted in a number of breakthroughs. One of the latest long awaited surprises is the spectrum of $_{26}$Fe just published by AMS-02. Because of the large frag… ▽ More Since its launch, the Alpha Magnetic Spectrometer - 02 (AMS-02) has delivered outstanding quality measurements of the spectra of cosmic-ray (CR) species, $\bar{p}$, $e^{\pm}$, and nuclei, $_1$H-$_8$O, $_{10}$Ne, $_{12}$Mg, $_{14}$Si, which resulted in a number of breakthroughs. One of the latest long awaited surprises is the spectrum of $_{26}$Fe just published by AMS-02. Because of the large fragmentation cross section and large ionization energy losses, most of CR iron at low energies is local, and may harbor some features associated with relatively recent supernova (SN) activity in the solar neighborhood. Our analysis of new iron spectrum together with Voyager 1 and ACE-CRIS data reveals an unexpected bump in the iron spectrum and in the Fe/He, Fe/O, and Fe/Si ratios at 1-2 GV, while a similar feature in the spectra of He, O, Si, and in their ratios is absent, hinting at a local source of low-energy CRs. The found excess fits well with recent discoveries of radioactive $^{60}$Fe deposits in terrestrial and lunar samples, and in CRs. We provide an updated local interstellar spectrum (LIS) of iron in the energy range from 1 MeV nucleon$^{-1}$ to $\sim$10 TeV nucleon$^{-1}$. Our calculations employ the GalProp-HelMod framework that is proved to be a reliable tool in deriving the LIS of CR $\bar{p}$, $e^{-}$, and nuclei $Z\le28$. △ Less

Submitted 22 March, 2021; v1 submitted 29 January, 2021; originally announced January 2021.

Comments: 12 pages, 9 figures, 5 tables; Accepted for publication in the Astrophysical Journal. arXiv admin note: text overlap with arXiv:2006.01337

Journal ref: The Astrophysical Journal 913 (2021) 5

Rethinking Experience Replay: a Bag of Tricks for Continual Learning

Authors: Pietro Buzzega, Matteo Boschini, Angelo Porrello, Simone Calderara

Abstract: In Continual Learning, a Neural Network is trained on a stream of data whose distribution shifts over time. Under these assumptions, it is especially challenging to improve on classes appearing later in the stream while remaining accurate on previous ones. This is due to the infamous problem of catastrophic forgetting, which causes a quick performance degradation when the classifier focuses on lea… ▽ More In Continual Learning, a Neural Network is trained on a stream of data whose distribution shifts over time. Under these assumptions, it is especially challenging to improve on classes appearing later in the stream while remaining accurate on previous ones. This is due to the infamous problem of catastrophic forgetting, which causes a quick performance degradation when the classifier focuses on learning new categories. Recent literature proposed various approaches to tackle this issue, often resorting to very sophisticated techniques. In this work, we show that naive rehearsal can be patched to achieve similar performance. We point out some shortcomings that restrain Experience Replay (ER) and propose five tricks to mitigate them. Experiments show that ER, thus enhanced, displays an accuracy gain of 51.2 and 26.9 percentage points on the CIFAR-10 and CIFAR-100 datasets respectively (memory buffer size 1000). As a result, it surpasses current state-of-the-art rehearsal-based methods. △ Less

Submitted 12 October, 2020; originally announced October 2020.

Comments: 8 pages, 6 figures. Accepted at the 25th International Conference on Pattern Recognition (ICPR 2020), Milan, Italy

Inference of the Local Interstellar Spectra of Cosmic Ray Nuclei Z<=28 with the GalProp-HelMod Framework

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. Johannesson, G. La Vacca, N. Masi, I. V. Moskalenko, S. Pensotti, T. A. Porter, L. Quadrani, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: Composition and spectra of Galactic cosmic rays (CRs) are vital for studies of high-energy processes in a variety of environments and on different scales, for interpretation of gamma-ray and microwave observations, disentangling possible signatures of new phenomena, and for understanding of our local Galactic neighborhood. Since its launch, AMS-02 has delivered outstanding quality measurements of… ▽ More Composition and spectra of Galactic cosmic rays (CRs) are vital for studies of high-energy processes in a variety of environments and on different scales, for interpretation of gamma-ray and microwave observations, disentangling possible signatures of new phenomena, and for understanding of our local Galactic neighborhood. Since its launch, AMS-02 has delivered outstanding quality measurements of the spectra of antiprotons, electrons, positrons, and nuclei: H-O, Ne, Mg, Si. These measurements resulted in a number of breakthroughs, however, spectra of heavier nuclei and especially low-abundance nuclei are not expected until later in the mission. Meanwhile, a comparison of published AMS-02 results with earlier data from HEAO-3-C2 indicate that HEAO-3-C2 data may be affected by undocumented systematic errors. Utilizing such data to compensate for the lack of AMS-02 measurements could result in significant errors. In this paper we show that a fraction of HEAO-3-C2 data match available AMS-02 measurements quite well and can be used together with Voyager 1 and ACE-CRIS data to make predictions for the local interstellar spectra (LIS) of nuclei that are not yet released by AMS-02. We are also updating our already published LIS to provide a complete set from H-Ni in the energy range from 1 MeV/nucleon to ~100-500 TeV/nucleon thus covering 8-9 orders of magnitude in energy. Our calculations employ the GalProp-HelMod framework that is proved to be a reliable tool in deriving the LIS of CR antiprotons, electrons, and nuclei H-O. △ Less

Submitted 5 September, 2021; v1 submitted 1 June, 2020; originally announced June 2020.

Comments: 90 pages, 36 figures, 60 tables, ApJS; Fixed typos in Eqs.(C2) and (C32)

Journal ref: The Astrophysical Journal Supplement Series 250 (2020) 27

Dark Experience for General Continual Learning: a Strong, Simple Baseline

Authors: Pietro Buzzega, Matteo Boschini, Angelo Porrello, Davide Abati, Simone Calderara

Abstract: Continual Learning has inspired a plethora of approaches and evaluation settings; however, the majority of them overlooks the properties of a practical scenario, where the data stream cannot be shaped as a sequence of tasks and offline training is not viable. We work towards General Continual Learning (GCL), where task boundaries blur and the domain and class distributions shift either gradually o… ▽ More Continual Learning has inspired a plethora of approaches and evaluation settings; however, the majority of them overlooks the properties of a practical scenario, where the data stream cannot be shaped as a sequence of tasks and offline training is not viable. We work towards General Continual Learning (GCL), where task boundaries blur and the domain and class distributions shift either gradually or suddenly. We address it through mixing rehearsal with knowledge distillation and regularization; our simple baseline, Dark Experience Replay, matches the network's logits sampled throughout the optimization trajectory, thus promoting consistency with its past. By conducting an extensive analysis on both standard benchmarks and a novel GCL evaluation setting (MNIST-360), we show that such a seemingly simple baseline outperforms consolidated approaches and leverages limited resources. We further explore the generalization capabilities of our objective, showing its regularization being beneficial beyond mere performance. △ Less

Submitted 22 October, 2020; v1 submitted 15 April, 2020; originally announced April 2020.

Comments: 24 pages, 4 figures. Accepted at 34th Conference on Neural Information Processing Systems (NeurIPS 2020), Vancouver, Canada

Deciphering the local Interstellar spectra of secondary nuclei with GALPROP/HelMod framework and a hint for primary lithium in cosmic rays

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. Johannesson, G. La Vacca, N. Masi, I. V. Moskalenko, S. Pensotti, T. A. Porter, L. Quadrani, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: Local interstellar spectra (LIS) of secondary cosmic ray (CR) nuclei, lithium, beryllium, boron, and partially secondary nitrogen, are derived in the rigidity range from 10 MV to ~200 TV using the most recent experimental results combined with the state-of-the-art models for CR propagation in the Galaxy and in the heliosphere. The lithium spectrum appears somewhat flatter at high energies compared… ▽ More Local interstellar spectra (LIS) of secondary cosmic ray (CR) nuclei, lithium, beryllium, boron, and partially secondary nitrogen, are derived in the rigidity range from 10 MV to ~200 TV using the most recent experimental results combined with the state-of-the-art models for CR propagation in the Galaxy and in the heliosphere. The lithium spectrum appears somewhat flatter at high energies compared to other secondary species that may imply a primary lithium component. Two propagation packages, GALPROP and HelMod, are combined to provide a single framework that is run to reproduce direct measurements of CR species at different modulation levels, and at both polarities of the solar magnetic field. An iterative maximum-likelihood method is developed that uses GALPROP-predicted LIS as input to HelMod, which provides the modulated spectra for specific time periods of the selected experiments for the model-data comparison. The proposed LIS accommodate the low-energy interstellar spectra measured by Voyager 1, HEAO-3, and ACE/CRIS as well as the high-energy observations by PAMELA, AMS-02, and earlier experiments that are made deep in the heliosphere. The interstellar and heliospheric propagation parameters derived in this study are consistent with our earlier results for propagation of CR protons, helium, carbon, oxygen, antiprotons, and electrons. △ Less

Submitted 12 December, 2019; v1 submitted 8 November, 2019; originally announced November 2019.

Comments: 17 pages, 10 figures, 6 tables, ApJ accepted

Journal ref: The Astrophysical Journal 889 (2020) 167

The HelMod Model in the Works for Inner and Outer Heliosphere: from AMS to Voyager Probes Observations

Authors: Matteo Jerome Boschini, Stefano Della Torre, Massimo Gervasi, Giuseppe La Vacca, Pier Giorgio Rancoita

Abstract: HelMod is a Monte Carlo code developed to describe the transport of Galactic Cosmic Rays (GCRs) through the heliosphere from the interstellar space to the Earth. In the current HelMod version 4 the modulation process, based on Parker's equation, is applied to the propagation of GCRs in the inner and outer heliosphere, i.e., including the heliosheath. HelMod was proved to reproduce protons, nuclei… ▽ More HelMod is a Monte Carlo code developed to describe the transport of Galactic Cosmic Rays (GCRs) through the heliosphere from the interstellar space to the Earth. In the current HelMod version 4 the modulation process, based on Parker's equation, is applied to the propagation of GCRs in the inner and outer heliosphere, i.e., including the heliosheath. HelMod was proved to reproduce protons, nuclei and electrons cosmic rays spectra observed during solar cycles 23-24 by several detectors, for instance, PAMELA, BESS and AMS-02. In particular, the unprecedented accuracy of AMS-02 observations allowed one a better tuning of the description regarding the solar modulation mechanisms implemented in HelMod. In addition, HelMod demonstrated to be capable of reproducing the uxes observed by the Voyager probes in the inner and outer regions of heliosphere up to its border. △ Less

Submitted 18 March, 2019; originally announced March 2019.

Deciphering the local Interstellar spectra of primary cosmic ray species with HelMod

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. Johannesson, G. La Vacca, N. Masi, I. V. Moskalenko, S. Pensotti, T. A. Porter, L. Quadrani, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: Local interstellar spectra (LIS) of primary cosmic ray (CR) nuclei, such as helium, oxygen, and mostly primary carbon are derived for the rigidity range from 10 MV to ~200 TV using the most recent experimental results combined with the state-of-the-art models for CR propagation in the Galaxy and in the heliosphere. Two propagation packages, GALPROP and HelMod, are combined into a single framework… ▽ More Local interstellar spectra (LIS) of primary cosmic ray (CR) nuclei, such as helium, oxygen, and mostly primary carbon are derived for the rigidity range from 10 MV to ~200 TV using the most recent experimental results combined with the state-of-the-art models for CR propagation in the Galaxy and in the heliosphere. Two propagation packages, GALPROP and HelMod, are combined into a single framework that is used to reproduce direct measurements of CR species at different modulation levels, and at both polarities of the solar magnetic field. The developed iterative maximum-likelihood method uses GALPROP-predicted LIS as input to HelMod, which provides the modulated spectra for specific time periods of the selected experiments for model-data comparison. The interstellar and heliospheric propagation parameters derived in this study are consistent with our prior analyses using the same methodology for propagation of CR protons, helium, antiprotons, and electrons. The resulting LIS accommodate a variety of measurements made in the local interstellar space (Voyager 1) and deep inside the heliosphere at low (ACE/CRIS, HEAO-3) and high energies (PAMELA, AMS-02). △ Less

Submitted 18 April, 2018; originally announced April 2018.

Comments: 13 pages, 13 figures, 6 tables, ApJ in press. arXiv admin note: text overlap with arXiv:1704.06337

Journal ref: The Astrophysical Journal 858 (2018) 61

HelMod in the works: from direct observations to the local interstellar spectrum of cosmic-ray electrons

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. Johannesson, G. La Vacca, N. Masi, I. V. Moskalenko, S. Pensotti, T. A. Porter, L. Quadrani, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: The local interstellar spectrum (LIS) of cosmic-ray (CR) electrons for the energy range 1 MeV to 1 TeV is derived using the most recent experimental results combined with the state-of-the-art models for CR propagation in the Galaxy and in the heliosphere. Two propagation packages, GALPROP and HelMod, are combined to provide a single framework that is run to reproduce direct measurements of CR spec… ▽ More The local interstellar spectrum (LIS) of cosmic-ray (CR) electrons for the energy range 1 MeV to 1 TeV is derived using the most recent experimental results combined with the state-of-the-art models for CR propagation in the Galaxy and in the heliosphere. Two propagation packages, GALPROP and HelMod, are combined to provide a single framework that is run to reproduce direct measurements of CR species at different modulation levels, and at both polarities of the solar magnetic field. An iterative maximum-likelihood method is developed that uses GALPROP-predicted LIS as input to HelMod, which provides the modulated spectra for specific time periods of the selected experiments for model-data comparison. The optimized HelMod parameters are then used to adjust GALPROP parameters to predict a refined LIS with the procedure repeated subject to a convergence criterion. The parameter optimization uses an extensive data set of proton spectra from 1997-2015. The proposed CR electron LIS accommodates both the low-energy interstellar spectra measured by Voyager 1 as well as the high-energy observations by PAMELA and AMS-02 that are made deep in the heliosphere; it also accounts for Ulysses counting rate features measured out of the ecliptic plane. The interstellar and heliospheric propagation parameters derived in this study agree well with our earlier results for CR protons, helium nuclei, and anti-protons propagation and LIS obtained in the same framework. △ Less

Submitted 12 January, 2018; originally announced January 2018.

Comments: 11 pages, 14 figures, 4 tables; ApJ, in press

Journal ref: The Astrophysical Journal 854 (2018) 94

The HelMod Monte Carlo Model for the Propagation of Cosmic Rays in Heliosphere

Authors: Matteo J. Boschini, Stefano Della Torre, Massimo Gervasi, Davide Grandi, Giuseppe La Vacca, Simonetta Pensotti, Pier Giorgio Rancoita, Davide Rozza, Mauro Tacconi

Abstract: The heliospheric modulation model HelMod solve the transport-equation for the Galactic Cosmic Rays propagation through the heliosphere down to Earth. It is based on a 2-D Monte Carlo approach, that includes a general description of the symmetric and antisymmetric parts of the diffusion tensor, thus, properly treating the particle drift effects as well as convection within the solar wind and adiaba… ▽ More The heliospheric modulation model HelMod solve the transport-equation for the Galactic Cosmic Rays propagation through the heliosphere down to Earth. It is based on a 2-D Monte Carlo approach, that includes a general description of the symmetric and antisymmetric parts of the diffusion tensor, thus, properly treating the particle drift effects as well as convection within the solar wind and adiabatic energy loss. The model was tuned in order to fit 1) the data observed outside the ecliptic plane at several distances from the Earth and 2) the spectra observed near the Earth for both, high and low solar activity periods. Great importance was given to description of polar regions of the heliosphere. We present the flux for protons, antiprotons and helium nuclei computed along the solar cycle 23-24 in comparison with experimental observations and prediction for the full solar cycle 24. △ Less

Submitted 11 August, 2017; originally announced August 2017.

Comments: 4 pages, 2 figures, proceedings of IAU Symposium 335 - Space Weather of the Heliosphere: Processes and Forecasts, July 17-21, 2017, University of Exeter, UK

Solution of heliospheric propagation: unveiling the local interstellar spectra of cosmic ray species

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. Johannesson, M. Kachelriess, G. La Vacca, N. Masi, I. V. Moskalenko, E. Orlando, S. S. Ostapchenko, S. Pensotti, T. A. Porter, L. Quadrani, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: Local interstellar spectra (LIS) for protons, helium and antiprotons are built using the most recent experimental results combined with the state-of-the-art models for propagation in the Galaxy and heliosphere. Two propagation packages, GALPROP and HelMod, are combined to provide a single framework that is run to reproduce direct measurements of cosmic ray (CR) species at different modulation leve… ▽ More Local interstellar spectra (LIS) for protons, helium and antiprotons are built using the most recent experimental results combined with the state-of-the-art models for propagation in the Galaxy and heliosphere. Two propagation packages, GALPROP and HelMod, are combined to provide a single framework that is run to reproduce direct measurements of cosmic ray (CR) species at different modulation levels and at both polarities of the solar magnetic field. To do so in a self-consistent way, an iterative procedure was developed, where the GALPROP LIS output is fed into HelMod that provides modulated spectra for specific time periods of selected experiments to compare with the data; the HelMod parameters optimization is performed at this stage and looped back to adjust the LIS using the new GALPROP run. The parameters were tuned with the maximum likelihood procedure using an extensive data set of proton spectra from 1997-2015. The proposed LIS accommodate both the low energy interstellar CR spectra measured by Voyager 1 and the high energy observations by BESS, Pamela, AMS-01, and AMS-02 made from the balloons and near-Earth payloads; it also accounts for Ulysses counting rate features measured out of the ecliptic plane. The found solution is in a good agreement with proton, helium, and antiproton data by AMS-02, BESS, and PAMELA in the whole energy range. △ Less

Submitted 20 April, 2017; originally announced April 2017.

Comments: 22 pages, 23 figures, 8 tables, and 3 machine-readable tables; ApJ, in press

Journal ref: The Astrophysical Journal, 840:115 (21pp), 2017 May 10

Propagation of Cosmic Rays in Heliosphere: the HelMod Model

Authors: M. J. Boschini, S. Della Torre, M. Gervasi, G. La Vacca, P. G. Rancoita

Abstract: The heliospheric modulation model \helmod{} is a two dimensional treatment dealing with the helio-colatitude and radial distance from Sun and is employed to solve the transport-equation for the GCR propagation through the heliosphere down to Earth. This work presents the current version 3 of the \helmod{} model and reviews how main processes involved in GCR propagation were implemented.~The treatm… ▽ More The heliospheric modulation model \helmod{} is a two dimensional treatment dealing with the helio-colatitude and radial distance from Sun and is employed to solve the transport-equation for the GCR propagation through the heliosphere down to Earth. This work presents the current version 3 of the \helmod{} model and reviews how main processes involved in GCR propagation were implemented.~The treatment includes the so-called particle drift effects --~e.g., those resulting, for instance, from the extension of the neutral current sheet inside the heliosphere and from the curvature and gradient of the IMF --, which affect the transport of particles entering the solar cavity as a function of their charge sign.~The \helmod{} model is capable to provide modulated spectra which well agree within the experimental errors with those measured by AMS-01, BESS, PAMELA and AMS-02 during the solar cycles 23 and 24.~Furthermore, the counting rate measured by Ulysses at +/- 80degree of solar latitude and 1 to 5\,AU was also found in agreement with that expected by \helmod{} code version 3. △ Less

Submitted 12 April, 2017; originally announced April 2017.

Cosmic Rays Propagation with HelMod: Difference between forward-in-time and backward-in-time approaches

Authors: S. Della Torre, P. Bobik, M. J. Boschini, M. Gervasi, D. Grandi, G. La Vacca, S. Pensotti, M. Putis, P. G. Rancoita, D. Rozza, M. Tacconi, M. Zannoni

Abstract: The cosmic rays modulation inside the heliosphere is well described by a transport equation introduced by Parker in 1965. To solve this equation several approaches were followed in the past. Recently the Monte Carlo approach becomes widely used in force of his advantages with respect to other numerical methods. In the Monte Carlo approach, the transport equation is associated to a fully equivalent… ▽ More The cosmic rays modulation inside the heliosphere is well described by a transport equation introduced by Parker in 1965. To solve this equation several approaches were followed in the past. Recently the Monte Carlo approach becomes widely used in force of his advantages with respect to other numerical methods. In the Monte Carlo approach, the transport equation is associated to a fully equivalent set of Stochastic Differential Equations. This set is used to describe the stochastic path of a quasi-particle from a source, e.g., the interstellar medium, to a specific target, e.g., a detector at Earth. In this work, we present both the Forward-in-Time and Backward-in-Time Monte Carlo solutions. We present an implementation of both algorithms in the framework of HelMod Code showing that the difference between the two approach is below 5\% that can be quoted as the systematic uncertain of the Method itself. △ Less

Submitted 13 August, 2015; originally announced November 2015.

Comments: presented at the The 34th International Cosmic Ray Conference, 30 July- 6 August, 2015, The Hague, The Netherlands

Cosmic Ray Modulation studied with HelMod Monte Carlo tool and comparison with Ulysses Fast Scan Data during consecutive Solar Minima

Authors: P. Bobik, G. Boella, M. J. Boschini, S. Della Torre, M. Gervasi, D. Grandi, G. La Vacca, K. Kudela, S. Pensotti, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: The Cosmic Rays propagation was studied in details using the HelMod-2D Monte Carlo code, that includes a general description of the diffusion tensor, and polar magnetic-field. The Numerical Approach used in this work is based on a set of Stochastic Differential Equations fully equivalent to the well know Parker Equation for the transport of Cosmic Rays. In our approach the Diffusion tensor in the… ▽ More The Cosmic Rays propagation was studied in details using the HelMod-2D Monte Carlo code, that includes a general description of the diffusion tensor, and polar magnetic-field. The Numerical Approach used in this work is based on a set of Stochastic Differential Equations fully equivalent to the well know Parker Equation for the transport of Cosmic Rays. In our approach the Diffusion tensor in the frame of the magnetic field turbolence does not depends explicitly by Solar Latitude but varies with time using a diffusion parameter obtained by Neutron Monitors. The parameters of the Model were tuned using data during the solar Cycle 23 and Ulysses latitudinal Fast Scan in 1995. The actual parametrization is able to well reproduce the observed latitudinal gradient of protons and the southward shift of the minimum of latitudinal intensity. The description of the model is also available online at website www.helmod.org. The model was then applied on Pamela/Ulysses proton intensity from 2006 up to 2009. The model during this 4-year continous period agree well with both PAMELA (at 1 AU) and Ulysses data (at various solar distance and solar latitude). The agreement improves when considering the ratio between this data. Studies done also with particles with different charge (e.g. electrons) allow us to explain the presence (or not) of protons and electrons latitudinal gradients observed by Ulysses during the Latitudinal Fast Scan in 1995 and 2007. △ Less

Submitted 19 July, 2013; originally announced July 2013.

GeoMag and HelMod webmodels version for magnetosphere and heliosphere transport of cosmic rays

Authors: P. Bobik, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, G. La Vacca, S. Pensotti, M. Putis, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: We implemented a website to deal with main effects on Cosmic Ray access to the Earth, i.e. the Solar Modulation and the Geomagnetic Field effect. In helmod.org the end user can easily access a web interface to results catalog of the HelMod Monte Carlo Code. This Model uses a Monte Carlo Approach to solves the Parker Transport Equation, obtaining a modulated proton flux for a period (monthly averag… ▽ More We implemented a website to deal with main effects on Cosmic Ray access to the Earth, i.e. the Solar Modulation and the Geomagnetic Field effect. In helmod.org the end user can easily access a web interface to results catalog of the HelMod Monte Carlo Code. This Model uses a Monte Carlo Approach to solves the Parker Transport Equation, obtaining a modulated proton flux for a period (monthly average) between January 1990 and december 2007. geomagsphere.org is instead based on GeoMag Backtracing Code, that solves the Lorentz equation with a Runge-Kutta method of 6th order, and, reversing charge sign and velocity, reconstruct particle trajectories in the Earth Magnetosphere back in time. We use last models of internal (IGRF-11) and external (Tsyganenko 1996 -T96- and 2005 -T05-) field components valid up to 2015. Particles are backtraced to the outer (magnetopause) or inner boundary to separate Primary (allowed trajectory) from Secondary (forbidden) Cosmic Rays. This code has been used both for reproducing known effects as East-West effect and rigidity cutoff calculations. In geomagsphere.org the user can choose the external field model from Tsyganenko (T96 or T05) and obtain for a fixed position and date from 1st Jan. 1968 (T96) and 1st Jan. 1995 (T05) respectively till 31$^{st}$ Dec 2012, the vertical rigidity cutoff estimation obtained with the backtracing technique with a rigidity step of 0.1 GV. For a more precise calculation (0.01 GV), requiring more CPU time, results are sent to the user by email (mail model) △ Less

Submitted 19 July, 2013; originally announced July 2013.

Suprathermal particle addition to solar wind pressure: possible influence on magnetospheric transmissivity of low energy cosmic rays?

Authors: P. Bobik, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, G. La Vacca, M. Mallamaci, S. Pensotti, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: Energetic (suprathermal) solar particles, accelerated in the interplanetary medium, contribute to the solar wind pressure, in particular during high solar activity periods. We estimated the effect of the increase of solar wind pressure due to suprathermal particles on magnetospheric transmissivity of galactic cosmic rays in the case of one recent solar event. Energetic (suprathermal) solar particles, accelerated in the interplanetary medium, contribute to the solar wind pressure, in particular during high solar activity periods. We estimated the effect of the increase of solar wind pressure due to suprathermal particles on magnetospheric transmissivity of galactic cosmic rays in the case of one recent solar event. △ Less

Submitted 19 July, 2013; originally announced July 2013.

Geomagnetic Backtracing: A comparison of Tsyganenko 1996 and 2005 External Field models with AMS-02 data

Authors: M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, S. Haino, G. La Vacca, S. Pensotti, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: We used a backtracing code to reconstruct particle trajectory inside the Earth Magnetosphere during the last solar active period (2011 and 2012) when very high Solar Wind pressure values were measured. We compared our results on AMS-02 proton and electron data with 2 different External Field models, namely Tsyganenko 1996 (T96) and 2005 (T05), both for quiet (defined as the periods when the solar… ▽ More We used a backtracing code to reconstruct particle trajectory inside the Earth Magnetosphere during the last solar active period (2011 and 2012) when very high Solar Wind pressure values were measured. We compared our results on AMS-02 proton and electron data with 2 different External Field models, namely Tsyganenko 1996 (T96) and 2005 (T05), both for quiet (defined as the periods when the solar wind pressure is below the average value, set at 2nPa) and active periods. Although T05 has been specifically designed for storm events, at high energy the particle trajectory is similar for the two models. For instance at rigidities larger than 50 GV, the RMS of angular difference between reconstructed asymptotic direction outside the Magnetosphere is of the order of the millirad, while it increases at intermediate energies. We also confirmed, as a function of the pointing direction, the well known East-West effect on the trajectory of primary particles and on the access solid angle on the AMS detector. △ Less

Submitted 19 July, 2013; originally announced July 2013.

An expression for the Mott cross section of electrons and positrons on nuclei with Z up t0 118

Authors: M J Boschini, C Consoland, M Gervasi, S Giani, D Grandi, V Ivanchenko, P Nieminem, S Pensotti, P G Rancoita, M Tacconi

Abstract: In the present work, an improved numerical solution for determining the ratio,$\mathcal{R}^{\rm Mott}$, of the unscreened Mott differential cross section (MDCS) with respect to Rutherford's formula is proposed for the scattering of electrons and positrons on nuclei with $1\leq Z \leq 118$. It accounts for incoming lepton energies between 1\,keV and 900\,MeV. For both electrons and positrons, a fit… ▽ More In the present work, an improved numerical solution for determining the ratio,$\mathcal{R}^{\rm Mott}$, of the unscreened Mott differential cross section (MDCS) with respect to Rutherford's formula is proposed for the scattering of electrons and positrons on nuclei with $1\leq Z \leq 118$. It accounts for incoming lepton energies between 1\,keV and 900\,MeV. For both electrons and positrons, a fitting formula and a set of fitting coefficients for the ratio $\mathcal{R}^{\rm Mott}$ on nuclei are also presented. The found average error of the latter practical interpolated expression is typically lower than 1% also at low energy for electrons and lower than 0.05% for positrons for all nuclei over the entire energy range. △ Less

Submitted 22 April, 2013; originally announced April 2013.

Comments: accepted for publication in Radiation Physics and Chemistry

Latitudinal Dependence of Cosmic Rays Modulation at 1 AU and Interplanetary-Magnetic-Field Polar Correction

Authors: P. Bobik, G. Boella, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, S. Pensotti, P. G. Rancoita, D. Rozza, M. Tacconi

Abstract: The cosmic rays differential intensity inside the heliosphere, for energy below 30 GeV/nuc, depends on solar activity and interplanetary magnetic field polarity. This variation, termed solar modulation, is described using a 2-D (radius and colatitude) Monte Carlo approach for solving the Parker transport equation that includes diffusion, convection, magnetic drift and adiabatic energy loss. Since… ▽ More The cosmic rays differential intensity inside the heliosphere, for energy below 30 GeV/nuc, depends on solar activity and interplanetary magnetic field polarity. This variation, termed solar modulation, is described using a 2-D (radius and colatitude) Monte Carlo approach for solving the Parker transport equation that includes diffusion, convection, magnetic drift and adiabatic energy loss. Since the whole transport is strongly related to the interplanetary magnetic field (IMF) structure, a better understanding of his description is needed in order to reproduce the cosmic rays intensity at the Earth, as well as outside the ecliptic plane. In this work an interplanetary magnetic field model including the standard description on ecliptic region and a polar correction is presented. This treatment of the IMF, implemented in the HelMod Monte Carlo code (version 2.0), was used to determine the effects on the differential intensity of Proton at 1\,AU and allowed one to investigate how latitudinal gradients of proton intensities, observed in the inner heliosphere with the Ulysses spacecraft during 1995, can be affected by the modification of the IMF in the polar regions. △ Less

Submitted 10 December, 2012; v1 submitted 7 December, 2012; originally announced December 2012.

Comments: accepted for publication inAdvances in Astronomy

Heliosphere Dimension and Cosmic Ray Modulation

Authors: P. Bobik, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, F. Noventa, S. Pensotti, P. G. Rancoita, D. Rozza

Abstract: The differential intensities of Cosmic Rays at Earth were calculated using a 2D stochastic Montecarlo diffusion code and compared with observation data. We evaluated the effect of stretched and compressed heliospheres on the Cosmic Ray intensities at the Earth. This was studied introducing a dependence of the diffusion parameter on the heliospherical size. Then, we found that the optimum value of… ▽ More The differential intensities of Cosmic Rays at Earth were calculated using a 2D stochastic Montecarlo diffusion code and compared with observation data. We evaluated the effect of stretched and compressed heliospheres on the Cosmic Ray intensities at the Earth. This was studied introducing a dependence of the diffusion parameter on the heliospherical size. Then, we found that the optimum value of the heliospherical radius better accounting for experimental data. We also found that the obtained values depends on solar activity. Our results are compatible with Voyager observations and with models of heliospherical size modulation. △ Less

Submitted 24 December, 2011; originally announced December 2011.

Comments: To appear on the Proceedings of the 13th ICATPP Conference on Astroparticle, Particle, Space Physics and Detectors for Physics Applications, Villa Olmo (Como, Italy), 3-7 October, 2011, to be published by World Scientific, Singapore (2012)

Electrical Characterization of SiPM as a Function of Test Frequency and Temperature

Authors: M. J. Boschini, C. Consolandi, P. G. Fallica, M. Gervasi, D. Grandi, M. Mazzillo, S. Pensotti, P. G. Rancoita, D. Sanfilippo, M. Tacconi, G. Valvo

Abstract: Silicon Photomultipliers (SiPM) represent a promising alternative to classical photomultipliers, for instance, for the detection of photons in high energy physics and medical physics. In the present work, electrical characterizations of test devices - manufactured by ST Microelectronics - are presented. SiPMs with an area of 3.5x3.5 micron^2 and a cell pitch of 54 micron were manufactured as array… ▽ More Silicon Photomultipliers (SiPM) represent a promising alternative to classical photomultipliers, for instance, for the detection of photons in high energy physics and medical physics. In the present work, electrical characterizations of test devices - manufactured by ST Microelectronics - are presented. SiPMs with an area of 3.5x3.5 micron^2 and a cell pitch of 54 micron were manufactured as arrays of 64x64 cells and exhibiting a fill factor of 31%. The capacitance of SiPMs was measured as a function of reverse bias voltage at frequencies ranging from from 20 Hz up to 1 MHz and temperatures from 300 K down to 85 K. While leakage currents were measured at temperatures from 400 K down to 85 K. Thus, the threshold voltage - i.e., voltage corresponding to that at which the multiplication regime for the leakage current begins - could be determined as a function of temperature. Finally, an electrical model suited to reproduce the dependence of the frequency dependence of capacitance is presented. △ Less

Submitted 10 January, 2012; v1 submitted 1 December, 2011; originally announced December 2011.

Comments: To appear on the Proceedings of the 13th ICATPP Conference on Astroparticle, Particle, Space Physics and Detectors for Physics Applications, Villa Olmo (Como, Italy), 3-7 October, 2011, to be published by World Scientific (Singapore)

Nuclear and Non-Ionizing Energy-loss of Electrons with Low and Relativistic Energies in Materials and Space Environment

Authors: M. J. Boschini, C. Consolandi, M. Gervasi, S. Giani, D. Grandi, V. Ivanchenko, P. Nieminem, S. Pensotti, P. G. Rancoita, M. Tacconi

Abstract: The treatment of the electron-nucleus interaction based on the Mott differential cross section was extended to account for effects due to screened Coulomb potentials, finite sizes and finite rest masses of nuclei for electrons above 200 keV and up to ultra high energies. This treatment allows one to determine both the total and differential cross sections, thus, subsequently to calculate the resul… ▽ More The treatment of the electron-nucleus interaction based on the Mott differential cross section was extended to account for effects due to screened Coulomb potentials, finite sizes and finite rest masses of nuclei for electrons above 200 keV and up to ultra high energies. This treatment allows one to determine both the total and differential cross sections, thus, subsequently to calculate the resulting nuclear and non-ionizing stop** powers. Above a few hundreds of MeV, neglecting the effect due to finite rest masses of recoil nuclei the stop** power and NIEL result to be largely underestimated; while, above a few tens of MeV the finite size of the nuclear target prevents a further large increase of stop** powers which approach almost constant values. △ Less

Submitted 6 December, 2011; v1 submitted 17 November, 2011; originally announced November 2011.

Comments: To appear on the Proceedings of the 13th ICATPP Conference on Astroparticle, Particle, Space Physics and Detectors for Physics Applications, Villa Olmo (Como, Italy), 3--7 October, 2011, to be published by World Scientific (Singapore)

Systematic Investigation of Solar Modulation of Galactic Protons for Solar Cycle 23 using a Monte Carlo Approach with Particle Drift Effects and Latitudinal Dependence

Authors: P. Bobik, G. Boella, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, S. Pensotti, P. G. Rancoita, M. Tacconi

Abstract: A propagation model of galactic cosmic protons through the Heliosphere was implemented using a 2-D Monte Carlo approach to determine the differential intensities of protons during the solar cycle 23. The model includes the effects due to the variation of solar activity during the propagation of cosmic rays from the boundary of the heliopause down to Earth's position. Drift effects are also account… ▽ More A propagation model of galactic cosmic protons through the Heliosphere was implemented using a 2-D Monte Carlo approach to determine the differential intensities of protons during the solar cycle 23. The model includes the effects due to the variation of solar activity during the propagation of cosmic rays from the boundary of the heliopause down to Earth's position. Drift effects are also accounted for. The simulated spectra were found in agreement with those obtained with experimental observations carried out by BESS, AMS and PAMELA collaborations. In addition, the modulated spectrum determined with the present code for the year 1995 exhibits the latitudinal gradient and equatorial southward offset minimum found by Ulysses fast scan in 1995. △ Less

Submitted 19 October, 2011; originally announced October 2011.

Comments: accepted for publication on The Astrophysical Journal

Journal ref: The Astrophysical Journal, 745:132 (21pp), 2012 February 1

Antiproton modulation in the Heliosphere and AMS-02 antiproton over proton ratio prediction

Authors: P. Bobik, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, S. Pensotti, P. G. Rancoita

Abstract: We implemented a quasi time-dependent 2D stochastic model of solar modulation describing the transport of cosmic rays (CR) in the heliosphere. Our code can modulate the Local Interstellar Spectrum (LIS) of a generic charged particle (light cosmic ions and electrons), calculating the spectrum at 1AU. Several measurements of CR antiparticles have been performed. Here we focused our attention on the… ▽ More We implemented a quasi time-dependent 2D stochastic model of solar modulation describing the transport of cosmic rays (CR) in the heliosphere. Our code can modulate the Local Interstellar Spectrum (LIS) of a generic charged particle (light cosmic ions and electrons), calculating the spectrum at 1AU. Several measurements of CR antiparticles have been performed. Here we focused our attention on the CR antiproton component and the antiproton over proton ratio. We show that our model, using the same heliospheric parameters for both particles, fit the observed anti-p/p ratio. We show a good agreement with BESS-97 and PAMELA data and make a prediction for the AMS-02 experiment. △ Less

Submitted 1 February, 2011; originally announced February 2011.

Proton and antiproton modulation in the heliosphere for different solar conditions and AMS-02 measurements prediction

Authors: P. Bobik, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, S. Pensotti, P. G. Rancoita

Abstract: Galactic Cosmic Rays (GCRs) are mainly protons confined in the galactic magnetic field to form an isotropic flux inside the galaxy. Before reaching the Earth orbit they enter the Heliosphere and undergo diffusion, convection, magnetic drift and adiabatic energy loss. The result is a reduction of particles flux at low energy (below 10 GeV), called solar modulation. We realized a quasi time-dependen… ▽ More Galactic Cosmic Rays (GCRs) are mainly protons confined in the galactic magnetic field to form an isotropic flux inside the galaxy. Before reaching the Earth orbit they enter the Heliosphere and undergo diffusion, convection, magnetic drift and adiabatic energy loss. The result is a reduction of particles flux at low energy (below 10 GeV), called solar modulation. We realized a quasi time-dependent 2D Stochastic Simulation of Solar Modulation that is able to reproduce CR spectra once known the Local Interstellar Spectrum (LIS). We were able to estimate the different behaviors associated to the polarity dependence of the Heliospheric modulation for particles as well as for antiparticles. We show a good agreement with the antiproton/proton ratio measured by AMS-01, Pamela, BESS, Heat and Caprice and we performed a prediction for the AMS-02 Experiment. △ Less

Submitted 22 December, 2010; v1 submitted 14 December, 2010; originally announced December 2010.

Proton Modulation in the Heliosphere for Different Solar Conditions and Prediction for AMS-02

Authors: P. Bobik, G. Boella, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, E. Memola, S. Pensotti, P. G. Rancoita, M. Tacconi

Abstract: Spectra of Galactic Cosmic Rays (GCRs) measured at the Earth are the combination of several processes: sources production and acceleration, propagation in the interstellar medium and propagation in the heliosphere. Inside the solar cavity the flux of GCRs is reduced due to the solar modulation, the interaction which they have with the interplanetary medium. We realized a 2D stochastic simulation o… ▽ More Spectra of Galactic Cosmic Rays (GCRs) measured at the Earth are the combination of several processes: sources production and acceleration, propagation in the interstellar medium and propagation in the heliosphere. Inside the solar cavity the flux of GCRs is reduced due to the solar modulation, the interaction which they have with the interplanetary medium. We realized a 2D stochastic simulation of solar modulation to reproduce CR spectra at the Earth, and evaluated the importance in our results of the Local Interstellar Spectrum (LIS) model and its agreement with data at high energy. We show a good agreement between our model and the data taken by AMS-01 and BESS experiments during periods with different solar activity conditions. Furthermore we made a prediction for the flux which will be measured by AMS-02 experiment. △ Less

Submitted 17 March, 2011; v1 submitted 4 December, 2010; originally announced December 2010.

Comments: Accepted for publication in the Proceedings of the ICATPP Conference on Cosmic Rays for Particle and Astroparticle Physics, Villa Olmo (Como, Italy), 7-8 October, 2010, to be published by World Scientific (Singapore)

Electron and Positron solar modulation and prediction for AMS-02

Authors: P. Bobik, M. J. Boschini, C. Consolandi, S. Della Torre, M. Gervasi, D. Grandi, K. Kudela, S. Pensotti, P. G. Rancoita

Abstract: The solar modulation, a combination of diffusion, convection, magnetic drift and energy loss inside the heliosphere is usually seen as a depletion in the Galactic cosmic ray (CR) flux at low energy (less than 10 GeV/nuc). Antiparticles such as antiprotons or positrons undergo the same processes of respective particles but with a different magnitude depending on the Solar magnetic field polarity. F… ▽ More The solar modulation, a combination of diffusion, convection, magnetic drift and energy loss inside the heliosphere is usually seen as a depletion in the Galactic cosmic ray (CR) flux at low energy (less than 10 GeV/nuc). Antiparticles such as antiprotons or positrons undergo the same processes of respective particles but with a different magnitude depending on the Solar magnetic field polarity. For electrons and positrons, due to the small mass, energy loss mechanisms as inverse compton, synchrotron, bremsstrahlung and ionization have to be taken into account, together with the typical adiabatic losses considered in the heliosphere. We developed a Monte Carlo stochastic simulation with the aim to compare the solar modulation of particles and antiparticles in the same observation period. We are able to estimate the different behaviours associated to the charge sign dependent processes of the heliospheric modulation. We compared the simulated positron fraction with measurements performed by AMS-01 and PAMELA. We also present the prediction for the AMS-02 experiment. △ Less

Submitted 19 December, 2010; v1 submitted 22 November, 2010; originally announced November 2010.

Comments: Accepted for publication in the Proceedings of the ICATPP Conference on Cosmic Rays for Particle and Astroparticle Physics, Villa Olmo (Como, Italy), 7--8 October, 2010, to be published by World Scientific

Nuclear and Non-Ionizing Energy-Loss for Coulomb Scattered Particles from Low Energy up to Relativistic Regime in Space Radiation Environment

Authors: M. J. Boschini, C. Consolandi, M. Gervasi, S. Giani, D. Grandi, V. Ivanchenko, S. Pensotti, P. G. Rancoita, M. Tacconi

Abstract: In the space environment, instruments onboard of spacecrafts can be affected by displacement damage due to radiation. The differential scattering cross section for screened nucleus--nucleus interactions - i.e., including the effects due to screened Coulomb nuclear fields -, nuclear stop** powers and non-ionization energy losses are treated from about 50 keV/nucleon up to relativistic energies. In the space environment, instruments onboard of spacecrafts can be affected by displacement damage due to radiation. The differential scattering cross section for screened nucleus--nucleus interactions - i.e., including the effects due to screened Coulomb nuclear fields -, nuclear stop** powers and non-ionization energy losses are treated from about 50 keV/nucleon up to relativistic energies. △ Less

Submitted 21 January, 2014; v1 submitted 22 November, 2010; originally announced November 2010.

Comments: Accepted for publication in the Proceedings of the ICATPP Conference on Cosmic Rays for Particle and Astroparticle Physics, Villa Olmo (Como, Italy), 7--8 October, 2010, to be published by World Scientific