var sms = '<div id="sms_bulletin">  <div id="sms_header">  <div class="do_not_print" id="arrow_left"><a href="?next=-1"><img src="//www2.math.ethz.ch/smsjs/images/arrow_left.png"/></a></div><div class="do_not_print" id="arrow_right"><a href="?next=1"><img src="//www2.math.ethz.ch/smsjs/images/arrow_right.png"/></a></div>    <div class="sms_title">      <span style="float: left;">SMS</span>      <span>SWISS MATHEMATICAL SOCIETY</span>      <span style="float: right;">SMG</span>    </div>    <div class="sms_info">      <span style="float: left; width: 250px;">Bulletin no. 6</span>      <span style="float: right; text-align:right; width: 250px;">7.7 - 11.7.2025</span>      <span style="text-align: center; display: block; width: 250px; margin: auto; position: relative;">Summer Break 2025</span>    </div>  </div>  <div></div>  <div id="sms_content">    <div class="day">      <div class="day_title">        Monday, July 7      </div>      <div class="day_content">        <div class="day_content_item normal"><div class="event_location uni_la"><a class="external" href="https://www.epfl.ch">EPF Lausanne</a></div><div class="event_type"><a class="external" href="https://theory.epfl.ch/seminar/">Theory and Combinatorics Seminar</a></div><br /><div class="event_speaker"> Anay  Mehrotra</div><div class="event_title">Inherent Trade-offs in Language Generation: Hallucinations, Breadth, and Boosting <span class="abstract-button do_not_print">abstract</span></div><div class="abstract-text" style="display: none"><strong>Abstract:</strong><br />Can language models avoid hallucinations while being expressive enough to capture the full breadth of the target language (thus preventing mode collapse)? Recently, Kleinberg and Mullainathan [KM24], building on the classical works of Gold and Angluin, showed that surprisingly one can construct a model that avoids hallucinations, albeit potentially sacrificing breadth, leaving open the interesting question of whether this sacrifice in breadth is necessary.In this talk, we characterize when different notions of breadth are achievable while avoiding hallucinations. Our results answer the question raised in [KM24] by demonstrating that this tension between validity and breadth is indeed inherent in language generation. Additionally, time permitting, we will also explore some other questions, including: How many samples are needed to train a model? Given two generators, is it possible to combine them to achieve a form of boosting? Can a language model know it has learned the language?This talk is based on joint works with Steve Hanneke, Alkis Kalavasis, Amin Karbasi, and Grigoris Velegkas.</div><div class="event_time">12:15 &bull; EPF Lausanne, INJ114</div>        </div>        <div class="day_content_item normal"><div class="event_location uni_zh"><a class="external" href="https://ethz.ch">ETH Z&uuml;rich</a></div><div class="event_type"><a class="external" href="https://math.ethz.ch/s/talks-in-mathematical-physics">Talks in Mathematical Physics</a></div><br /><div class="event_speaker"> Iakovos Androulidakis (University of Athens)</div><div class="event_title">Singular polarizations <span class="abstract-button do_not_print">abstract</span></div><div class="abstract-text" style="display: none"><strong>Abstract:</strong><br />Weinstein\'s program for quantization using symplectic groupoids goes backto the 1990s. In 2012, E. Hawkins proposed a realization of this programfor geometric quantization of polarised observables in particular, usingthe deformation to normal cone. In this scheme, polarization is requiredto satisfy certain assumptions. However, Bonechi, Ciccoli et al gaveexamples of rather singular polarizations on the Podles sphere and complexprojective space. In this talk we discuss a reformulation of Hawkins\'scheme which accommodates singular polarizations. This is joint work withF. Bonechi, N. Ciccoli, V. Zenobi and P. Antonini.</div><div class="event_time">13:30 &bull; ETH Zentrum, Winterthurerstrasse 190, Z&uuml;rich, Building Y27, Room H 25</div>        </div>      </div>    </div>    <div class="day">      <div class="day_title">        Tuesday, July 8      </div>      <div class="day_content">        <div class="day_content_item no_event">        No events scheduled today!        </div>      </div>    </div>    <div class="day">      <div class="day_title">        Wednesday, July 9      </div>      <div class="day_content">        <div class="day_content_item no_event">        No events scheduled today!        </div>      </div>    </div>    <div class="day">      <div class="day_title">        Thursday, July 10      </div>      <div class="day_content">        <div class="day_content_item normal"><div class="event_location uni_la"><a class="external" href="https://www.epfl.ch">EPF Lausanne</a></div><div class="event_type"><a class="external" href="https://calendar.google.com/calendar/embed?mode=AGENDA&height=600&wkst=2&bgcolor=%23FFFFFF&src=t36it03v52cus2detrvinlgk18%40group.calendar.google.com&color=%235F6B02&ctz=Europe%2FZurich">Groups, Arithmetic and Algebraic Geometry Seminar </a></div><br /><div class="event_speaker"> Jihao Liu (Peking University)</div><div class="event_title">Abundance conjecture when &#957; &#8804; 1. <span class="abstract-button do_not_print">abstract</span></div><div class="abstract-text" style="display: none"><strong>Abstract:</strong><br /> In a recent joint work with Zheng Xu, we show that the non-vanishing conjecture implies the abundance conjecture when &#957; &#8804; 1. We also prove the abundance conjecture in dimension &#8804; 5 when &#954; &#8805; 0 and &#957; &#8804; 1 unconditionally. I will discuss this work in my talk.</div><div class="event_time">14:00 &bull; EPF Lausanne, CM 1 517</div>        </div>        <div class="day_content_item normal"><div class="event_location uni_la"><a class="external" href="https://www.epfl.ch">EPF Lausanne</a></div><div class="event_type"><a class="external" href="https://www.epfl.ch/schools/sb/research/math/statistics-seminars/">Statistics Seminar </a></div><br /><div class="event_speaker"> Riccardo  Passeggeri (Imperial College)</div><div class="event_title">Random signed measures <span class="abstract-button do_not_print">abstract</span></div><div class="abstract-text" style="display: none"><strong>Abstract:</strong><br />Point processes and, more generally, random measures are ubiquitous in statistics. However, they can only take positive values, which is a severe limitation in many situations. In this presentation, we introduce random signed measures, also known as real-valued random measures, and apply them to construct various Bayesian non-parametric models.In particular, we provide an existence result for random signed measures, allowing us to obtain a canonical definition for random signed measures and solve a long-standing open problem.Further, we provide a representation of completely random signed measures (CRSMs), which extends the celebrated Kingman\'s representation for completely random measures (CRMs) to the real-valued case. We then introduce specific classes of random signed measures, including the Skellam point process, which plays the role of the Poisson point process in the real-valued case, and the Gaussian random measure. We use the theoretical results to develop two Bayesian nonparametric models -- one for sentiment topic modeling and the other for random graphs -- and to investigate mean function estimation in Bayesian nonparametric regression.</div><div class="event_time">15:15 &bull; EPF Lausanne, CM 1 1000</div>        </div>      </div>    </div>    <div class="day">      <div class="day_title">        Friday, July 11      </div>      <div class="day_content">        <div class="day_content_item no_event">        No events scheduled today!        </div>      </div>    </div>  </div>  <div id="sms_footer">    address: Alessandra Naldi Windler, Departement Mathematik, HG G 51.4, ETH Zurich, Raemistrasse 101, CH-8092 Zurich<br />    email: <a href="mailto:bulletin@math.ch">bulletin@math.ch</a> | phone: +41 44 632 3684 | url: <a href="https://math.ch">https://math.ch</a>    <br /><br /><img src="//www2.math.ethz.ch/smsjs/sms_logo.gif" width="150" alt="SMS/SMG Logo" title="SMS/SMG Logo" />  </div></div>';