<div dir="ltr"><div><font size="2"><span style="font-family:arial,sans-serif">Boa noite! <br></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><br></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif">Reiniciamos, 2 semanas atrás, <span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span>o cicl</span></span><span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span>o de seminários do PIPGEs, com um seminário da Profa. Daiane Zuanetti (DEs-UFSCar), cuja gravação pode ser acessada <a href="http://www.pipges.ufscar.br/seminarios/modelos-de-mistura-sua-estimacao-e-algumas-aplicacoes">nesta página</a>. <br></span></span></span></font></div><div><font size="2"><br></font></div><div><font size="2">Aproveito para divulgar que <span style="font-family:arial,sans-serif"><span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span>sexta feira que vem (dia 15 de maio), teremos um seminário da Profa. Mariana Curi (ICMC-USP), as informações estão abaixo.</span></span></span></font><font size="2"><span style="font-family:arial,sans-serif"><span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span><br></span></span></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span><br></span></span></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span>Todas nossas atividades serão apresentadas pela plataforma Google-meet. Os links são enviados na sexta feira de manhã, o acesso é livre, estão convidad@s!  <br></span></span></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span><br></span></span></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span>Abraço,</span></span></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span>Sandro<br></span></span></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span></span></b></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span><br></span></b></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span><br></span></b></span></font></div><div><font size="2"><span style="font-family:arial,sans-serif"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span>SEMINÁRIO</span> do PIPGEs (UFSCAR/USP)</b></span></font></div><font size="2"><span style="font-family:arial,sans-serif"></span></font><div class="gmail_quote"><div dir="ltr"><div style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline;color:rgb(0,0,0);text-align:left"><div id="m_5855750512810540408gmail-m_816121964062394100gmail-content" style="margin:1em 0.25em 2em;padding:0px;list-style:outside none none;vertical-align:baseline;line-height:1.5em;clear:both"><div id="m_5855750512810540408gmail-m_816121964062394100gmail-content-core" style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><div id="m_5855750512810540408gmail-m_816121964062394100gmail-parent-fieldname-text-f9470ab7c03c400ea65249ff079bb67e" style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><p style="margin:0px 0px 1em;padding:0px;list-style:outside none none;vertical-align:baseline;line-height:1.8em"><font size="2"><span style="font-family:arial,sans-serif"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline">Data e Horário:</b><br style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline">15/05/2020 às 14h<br style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline">Local:</b><br style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline">Google Meet (à distância) - link a ser disponibilizado próximo ao dia do evento<b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><br></b></span></font></p><p style="margin:0px 0px 1em;padding:0px;list-style:outside none none;vertical-align:baseline;line-height:1.8em"><font size="2"><span style="font-family:arial,sans-serif"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline">Título:</b><br style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><span>Machine learning for estimation in IRT models.</span></span></font></p><p style="margin:0px 0px 1em;padding:0px;list-style:outside none none;vertical-align:baseline;line-height:1.8em"><font size="2"><span style="font-family:arial,sans-serif"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline">Palestrante:<br style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"></b><span>Mariana Curi (ICMC - USP)</span></span></font></p><p style="margin:0px 0px 1em;padding:0px;list-style:outside none none;vertical-align:baseline;line-height:1.8em"><font size="2"><span style="font-family:arial,sans-serif"><b style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline">Resumo: <span style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"></span><br style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"></b></span></font></p><font size="2"><span style="font-family:arial,sans-serif"><span>High dimensional latent space is still a challenge for usual
 estimation methods in Item Response Theory (IRT) models, like MCMC or 
maximum likelihood. In this work, we propose a Variational Autoencoder 
(VAE) architecture, a kind of unsupervised deep neural network, for a 
multidimensional IRT model parameter estimation. Our approach allows us 
to model high latent trait dimensions, overcoming some of the 
limitations concerned to “big data” analysis. The simulation studies 
show that, given enough data, the proposed method is competitive with 
the state-of-the-art ones with respect to predictive power and is much 
faster in runtime performance. The new approach is applied to a real 
data set to illustrate the usefulness of the proposed method in the 
context of educational assessment.</span><span style="white-space:nowrap"> <br></span></span></font></div><div style="margin:0px;padding:0px;list-style:outside none none;vertical-align:baseline"><font size="2"><span style="font-family:arial,sans-serif"><span style="white-space:nowrap"><br></span></span></font></div><font size="2"><span style="font-family:arial,sans-serif"></span></font><br><font size="2"><span style="font-family:arial,sans-serif"></span></font></div></div></div></div>
</div></div>