<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Hi GAPS,</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
We are happy to invite you to the latest talk in our monthly Early Career Scientist seminar series! This month’s talk will take place on Monday March 31st at 1 PM in the MOB Auditorium for those of you onsite (light snacks will be provided) and on Zoom for
those that aren’t (information listed below). The seminar is structured as a 40 minute talk followed by 15 minutes for you to ask any and all questions you might have the for speaker.</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
These seminars are initially focusing on machine learning and its connections with the physics we do here at TRIUMF. This month, we are happy to be re-inviting Dr. Anindita Maiti as our speaker. Anindita Maiti is a postdoctoral Fellow at Perimeter Institute
since Sept 2023. Previously, she was a postdoctoral fellow at Harvard Applied Math (theory of Deep Learning) during May-Aug 2023. Anindita received her PhD in theoretical high energy physics at Northeastern University and the NSF IAIFI, supervised by Prof.
James Halverson, in May 2023. Her research interest lies at the intersection of quantum field theory, quantum mechanics, statistical physics, and Machine Learning (ML). On one hand, Anindita leverages concepts and frameworks from fundamental physics, including
Feynman path integral formalism, renormalization group (RG) flow, random matrix theory, and Schrodinger equation, to develop a comprehensive understanding of Neural Networks and ML algorithms. On the other end, she utilizes such studies to make ML more trustworthy,
robust, and interpretable for quantum field theory and quantum simulations.</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
More information about the talk and zoom coordinates can be found below. See you on Monday!</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Your GAPS, on behalf of the Early Career Talks organizing committee</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
_______________________________________________________________________</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Speaker: Dr. Anindita Maiti (remote)</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Title: A Wilsonian RG framework for Regression Tasks in Supervised Learning</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Abstract: The performance of machine learning (ML) models fundamentally hinges on their ability to discriminate between relevant and irrelevant features in data. We introduce a first-of-its-kind Wilsonian RG framework to analyze the predictions of overparameterized
neural networks (NN), which are models characterized by an excess of parameters relative to the complexity of the task. These networks, trained via supervised learning, are known to produce noisy outputs in regression tasks. In our formulation, irrelevant
features within the data are systematically coarse-grained through momentum shell RG, inducing an RG flow that governs the evolution of noise in the predictions. When the irrelevant features follow a Gaussian distribution, this RG flow exhibits universality
across different NN architectures. In contrast, non-Gaussian features give rise to more intricate, data-dependent RG flows. This approach reveals novel behaviors in NNs that have eluded conventional ML methods. By advancing beyond philosophical analogies between
RG and ML, our framework offers a field theory-based methodology for understanding feature learning. This talk is based on the paper https://arxiv.org/abs/2405.06008.</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Zoom: https://ubc.zoom.us/j/64575929910?pwd=QAb0SbY9Zi5iqxSB4Dfa1N9wGBXBqQ.1</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Meeting ID: 645 7592 9910</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Passcode: 892489</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div class="elementToProof" style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div id="Signature" class="elementToProof" style="color: inherit;">
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<a href="http://gaps.triumf.ca" id="OWA8327e2d6-6900-852e-c2d2-a2f209bc324b" class="OWAAutoLink" title="gaps.triumf.ca">gaps.triumf.ca</a></div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Follow us on Instagram at <a href="https://www.instagram.com/triumfgaps/" id="OWA5e8b7af1-eb79-676a-4d9a-f163fdff3468" class="OWAAutoLink" title="https://www.instagram.com/triumfgaps/">
triumfgaps</a></div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Aptos, Aptos_EmbeddedFont, Aptos_MSFontService, Calibri, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<span style="font-family: Calibri, sans-serif; font-size: 14.6667px; color: rgb(36, 36, 36); background-color: rgb(255, 255, 255);"><i>TRIUMF is located on the traditional, ancestral, and unceded territory of the xʷməθkʷəy̓əm (Musqueam) people, who for millennia
have passed on their culture, history, and traditions from one generation to the next on this site.</i></span><br>
</div>
</div>
</body>
</html>