Random September 2020

random@lists.dm.unipi.it

25 participants
30 discussions

Webinar Passeggeri @Bocconi
by Giacomo Zanella 22 Sep '20

22 Sep '20

We announce the following DEC Statistics webinar: Date: Thursday, September 24, h12:00 Speaker: Riccardo Passeggeri (Sorbonne University) Title: On a class of completely random measures (CRMs) and its role in Bayesian analysis. The seminar will be mixed online and in-presence: - anybody can attend the seminar at the following link: https://zoom.us/j/92000724531 - members of Bocconi University can attend the meeting in presence at Roentgen building, floor -1, room AS-02, subject to appropriate prevention measures such as social distancing. Abstract: Quasi-infinitely divisible (QID) distributions have been recently introduced by Lindner, Pan and Sato (Trans. Amer. Math. Soc. 370 (2018) 8483-8520). A random variable X is QID if and only if there exist two infinitely divisible (ID) random variables Y and Z s.t. X +Y = Z (in distribution) and Y is independent of X. In this talk, we present QID random measures, describe their properties and their role in Bayesian analysis. In particular, we show that QID random measures are dense in the space of all CRMs with respect to convergence in distribution. In other words, *any CRM can be approximated in distribution by a QID random measure*. Moreover, we demonstrate that there exists a one to one correspondence between the law of QID random measures and certain characteristic pairs (as it happens for the Poisson point processes). That is, *the characteristic function of a QID random measure has an explicit formulation written in terms of two (unique) deterministic objects*. Finally, we present a nonparametric Bayesian statistical framework based on QID random measures, which extends recent works of Michael Jordan, Tamara Broderick, Trevor Campbell and co-authors. Kind regards, Giacomo Zanella

1 0

Online seminar Eddie Nikolova
by Marco Scarsini 20 Sep '20

20 Sep '20

Thursday 24 September, h 17:00 *Evdokia Nikolova, *University of Texas at Austin Risk-aversion and diversity in network routing <http://mailsender.luiss.it/lists/lt.php?id=cRpRCgUDSAUBBwBOVgFdBQs> Discussants: Valerio Dose and Marc Schröder Luiss | Viale Romania, 32 | Virtual Room* 70 This is the link to access the virtual room: https://luiss.webex.com/luiss/j.php?MTID=mcf4af3f6b4da7e8a19cf1fb052a70517 Information about future seminars can be found here https://economiaefinanza.luiss.it/seminari/prossimi-seminari ******************************************************* Marco Scarsini Dipartimento di Economia e Finanza Luiss Viale Romania 32 00197 Roma, ITALY URL: http://docenti.luiss.it/scarsini/

1 0

Call for interest - Assegno di ricerca @ Polimi
by Daniele Marazzina 17 Sep '20

17 Sep '20

Con preghiera di diffusione. Grazie mille -------------------------------------------------- Il Qfinlab, laboratorio di finanza quantitativa del Dipartimento di Matematica del Politecnico di Milano, in collaborazione con il gruppo di ricerca del Professor Agasisti del Dipartimento di Ingegneria Gestionale, sta valutando la possibilità di bandire un assegno di ricerca della durata di un anno su tematiche relative all'educazione finanziaria, creazione di percorsi per le scuole e analisi di impatto. Eventuali persone interessate possono inviare il loro cv all'indirizzo qfinlab(a)polimi.it<mailto:qfinlab@polimi.it> . Per informazioni sul gruppo, si rimanda a www.qfinlab.polimi.it<http://www.qfinlab.polimi.it/> Cordialmente Daniele Marazzina Emilio Barucci

1 0

Call for interest - Assegno di ricerca @ Polimi
by Daniele Marazzina 17 Sep '20

17 Sep '20

Con preghiera di diffusione. Grazie mille -------------------------------------------------- Il Qfinlab, laboratorio di finanza quantitativa del Dipartimento di Matematica del Politecnico di Milano, in collaborazione con il gruppo di ricerca del Professor Agasisti del Dipartimento di Ingegneria Gestionale, sta valutando la possibilità di bandire un assegno di ricerca della durata di un anno su tematiche relative all'educazione finanziaria, creazione di percorsi per le scuole e analisi di impatto. Eventuali persone interessate possono inviare il loro cv all'indirizzo qfinlab(a)polimi.it . Per informazioni sul gruppo, si rimanda a www.qfinlab.polimi.it Cordialmente Daniele Marazzina Emilio Barucci

1 0

seminar De Palma
by Dario Trevisan 17 Sep '20

17 Sep '20

Dear all, sorry for another message about today's seminar by G. De Palma. I just want to mention that there was an extra bracket in the google meet link. If you want to follow online please use the following link: meet.google.com/sdy-qmac-kch Regards, Dario -- Dario Trevisan, Università degli Studi di Pisa, Dipartimento di Matematica, Largo Bruno Pontecorvo 5, 56127 - Pisa (PI) Italy telephone: (+39) 050 2213832 mobile: (+39) 331 2899761 Skype: dario-trevisan e-mail: dario.trevisan AT unipi.it webpage: http://people.dm.unipi.it/trevisan/

1 0

PhD opportunity at the University of Leipzig
by Giovanni Conforti 16 Sep '20

16 Sep '20

Dear colleagues, Please find below the announcement for two Phd positions in the group of Max von Renesse. Best Regards Giovanni Conforti At the university of Leipzig we have two vacant PhD positions in the stochastic analysis group with a duration of three years under the supervision of Max von Renesse. Possible research directions concern applications of stochastic analysis e.g. in geometry, physics or statistics. Together with the Max Planck Institute for Mathematics in the Sciences (MPI-MIS) and some world class research groups Leipzig offers a lively and stimulating environment for research in a variety of intertwining mathematical disciplines. Each position comes with a teaching load of 2 hrs per week during semesters. The positions are to be filled by October 15th preferably. Some knowledge of basic German is helpful but no must. The ideal candidate (m./f.) has an excellent recent Masters or Diploma degree in mathematics or theoretical physics with a strong background in a combination of at least two of the following fields: stochastic analysis, differential geometry, (stochastic) PDE, optimal transport theory or statistics. Applications should include a CV, transcript of university diplomas and a list of up to three potential contacts for letters of recommendation. Further inquiries and applications should be sent to Prof. Dr. Max von Renesse, renesse(a)uni-leipzig.de. Application deadline is 24 September 2020.

1 0

seminar by De Palma -- change of room
by Dario Trevisan 15 Sep '20

15 Sep '20

Dear colleagues, the previously announced seminar by G. De Palma, to be held on Thursday, September 17, in Pisa, will take place in a larger room (Aula Magna), to safely accommodate more participants (up to 25 people). If you wish to join us in person please write me an email. Regards, Dario %%% Location: Aula Magna, Dipartimento di Matematica, Pisa (due to restrictions to the number of participants, we strongly suggest to attend online) Google Meet Link: meet.google.com/sdy-qmac-kch) Time: September 17, 2020, 2 PM Speaker: Giacomo De Palma (MIT) Title: Adversarial robustness guarantees for random deep neural networks Abstract: The reliability of most deep learning algorithms is fundamentally challenged by the existence of adversarial examples, which are incorrectly classified inputs that are extremely close to a correctly classified input. We study adversarial examples for deep neural networks with random weights and biases and prove that the L^1 distance of any given input from the classification boundary scales at least as sqrt(n), where n is the dimension of the input. We also extend our proof to cover all the L^p norms. Our results constitute a fundamental advance in the study of adversarial examples, and encompass a wide variety of architectures, which include any combination of convolutional or fully connected layers with skipped connections and pooling. We validate our results with experiments on both random deep neural networks and deep neural networks trained on the MNIST and CIFAR10 datasets. Given the results of our experiments on MNIST and CIFAR10, we conjecture that the proof of our adversarial robustness guarantee can be extended to trained deep neural networks. This extension will open the way to a thorough theoretical study of neural network robustness by classifying the relation between network architecture and adversarial distance. -- Dario Trevisan, Università degli Studi di Pisa, Dipartimento di Matematica, Largo Bruno Pontecorvo 5, 56127 - Pisa (PI) Italy telephone: (+39) 050 2213832 mobile: (+39) 331 2899761 Skype: dario-trevisan e-mail: dario.trevisan AT unipi.it webpage: http://people.dm.unipi.it/trevisan/

1 0

Three Postdoc positions at WIAS-Interacting random systems group
by Luisa Andreis 14 Sep '20

14 Sep '20

Dear colleagues, at the Weierstrass Institute in Berlin we are looking to fill up to three postdoc positions. We would be grateful if you could bring this advertisement to any interested candidate. Here is a brief description of the call: "The successful candidates will analyze large interacting random systems using probabilistic methods within the newly established DFG priority program SPP 2265 "Random Geometric Systems" (https://www.wias-berlin.de/projects/SPP2265/ <https://www.wias-berlin.de/projects/SPP2265/>). The main research topics of these positions are phase transitions during coagulation and condensation. The third position is financed from other funds and offers more thematic flexibility within the research group and organizational responsibility. Please direct scientific queries to Prof. Dr. W. König (Koenig(a)wias-berlin.de <mailto:Koenig@wias-berlin.de>). Deadline for applications: 30.09.2020" For more details and to apply please visit https://www.wias-berlin.de/jobs/job.jsp?lang=1&year=20&number=15 <https://www.wias-berlin.de/jobs/job.jsp?lang=1&year=20&number=15> Best regards, Luisa Andreis Post-Doc researcher WIAS-Weierstrass Institute Mohrenstrasse 39, 10117 Berlin, Germany Personal webpage: https://sites.google.com/view/luisaandreis/home <https://sites.google.com/view/luisaandreis/home> Email: andreis(a)wias-berlin.de <mailto:andreis@wias-berlin.de>

1 0

Seminar announcement
by Dario Trevisan 10 Sep '20

10 Sep '20

Dear colleagues, I would like to announce the following seminar by G. De Palma, in Pisa (and online, of course). Regards, Dario %%% Location: Sala Seminari, Dipartimento di Matematica, Pisa (due to restrictions to the number of participants, we strongly suggest to attend online) Google Meet Link: meet.google.com/sdy-qmac-kch) Time: September 17, 2020, 2 PM Speaker: Giacomo De Palma (MIT) Title: Adversarial robustness guarantees for random deep neural networks Abstract: The reliability of most deep learning algorithms is fundamentally challenged by the existence of adversarial examples, which are incorrectly classified inputs that are extremely close to a correctly classified input. We study adversarial examples for deep neural networks with random weights and biases and prove that the L^1 distance of any given input from the classification boundary scales at least as sqrt(n), where n is the dimension of the input. We also extend our proof to cover all the L^p norms. Our results constitute a fundamental advance in the study of adversarial examples, and encompass a wide variety of architectures, which include any combination of convolutional or fully connected layers with skipped connections and pooling. We validate our results with experiments on both random deep neural networks and deep neural networks trained on the MNIST and CIFAR10 datasets. Given the results of our experiments on MNIST and CIFAR10, we conjecture that the proof of our adversarial robustness guarantee can be extended to trained deep neural networks. This extension will open the way to a thorough theoretical study of neural network robustness by classifying the relation between network architecture and adversarial distance. -- Dario Trevisan, Università degli Studi di Pisa, Dipartimento di Matematica, Largo Bruno Pontecorvo 5, 56127 - Pisa (PI) Italy telephone: (+39) 050 2213832 mobile: (+39) 331 2899761 Skype: dario-trevisan e-mail: dario.trevisan AT unipi.it webpage: http://people.dm.unipi.it/trevisan/

1 0

Fw: STAR - STochastics And Risk - Online Seminars kick-off
by michele giordano 09 Sep '20

09 Sep '20

Dear all, It is soon time to kick-off the online seminar series: STAR - STochastics And Risk - Online Seminars. This series of seminars brings together all interested people in probability, stochastic analysis, control, risk evaluation, statistics, with a view towards applications, in particular to renewable energy markets and production. This series brings together the major research themes of the projects SCROLLER<https://www.mn.uio.no/math/english/research/projects/scroller/index.html>, SPATUS<https://www.mn.uio.no/math/english/research/projects/spatus/>, and STORM<https://www.mn.uio.no/math/english/research/projects/storm/>. If you would like to take part to the seminars series please register here.<https://nettskjema.no/a/159180#/page/1> <https://nettskjema.no/a/159180#/page/1> On Friday 18. September from 11:00-12:00 the first seminar will be held, and after that it will run on a weekly basis. The lecture will last for 45 minutes + questions, and afterwards you are invited to bring a cup of coffee/tea and have a chat in our "Coffe in the Stars". The first speaker is Andreas Petersson, with the seminar: Finite element approximation of Lyapunov equations for the computation of quadratic functionals of SPDEs Abstract: We consider the computation of quadratic functionals of the solution to a linear parabolic stochastic partial differential equation (SPDE) with multiplicative Gaussian noise on a bounded domain. The functionals are allowed to be path dependent and the noise is white in time and may be white in space. An operator valued Lyapunov equation, whose solution admits a deterministic representation of the functional of the SPDE solution, is used for this purpose and error estimates are shown in suitable operator norms for a fully discrete approximation of this equation. We also use these estimates to derive weak error rates for a fully discrete approximation of the SPDE itself. In the setting of finite element approximations, a computational complexity comparison reveals that approximating the Lyapunov equation allows us to compute quadratic functionals more cheaply compared to applying Monte Carlo or covariance-based methods directly to the discretized SPDE. We illustrate the theoretical results with numerical simulations. For more details on this event, see: https://www.mn.uio.no/math/english/research/projects/storm/events/seminars/… There you will find a list of all the speakers, abstracts, and more details about the seminar series. If you register to the seminars series, you will receive a mail with the details for the zoom meeting next week. We are looking forward to see you, online! Best regards, Michele and Giulia

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

Random September 2020