People by Discipline

Name	Discipline	Category	Phone	Office
Finkenthal Michael	Nonlinear	Emeriti		Danciger B
Rokni Mordechai	Nonlinear	Emeriti	02 6584 497	Danciger B 119
Schwob Yehuda	Nonlinear	Emeriti	02 6584 331	Kaplun 33
Sahaf Michal	Nonlinear	Graduate Students	02 6586 104	Levin 22
Vilenkin Arkady	Nonlinear	Senior Scientists	02 6584 471	Levin 16
Zhang MingMing	Nonlinear	Post-Docs	02 6586 104	Levin 22
Gat Omri	Nonlinear	Faculty	02 6586 361	Levin 17
Katzav Eytan	Nonlinear	Faculty	02 6585 681	Levin 105
Fineberg Jay	Nonlinear	Faculty	02 6585 207	Levin 107
Bayart Elsa	Nonlinear	Post-Docs	02 6586 604	Levin 101

Pages

Upcoming Events

Tomorrow

Physics Colloquium : "Criticality, universality and Grokking in Neural Networks"
- Start time: 12:00pm
- End date: Monday, December 9th
- End time: 01:30pm
- Lecturer: Yohai Bar-Sinai, Tel Aviv UniversityAbstract:The empirical success of neural models far surpasses their theoretical understanding, and explaining their inner workings is important both as a fundamental scientific question and for practical reasons. In this talk I’ll review some “big questions” in the field, and the prospects of applying physics-inspired methodologies to tackle them. We will focus on an intriguing phenomenon named Grokking, where a neural model learns to generalize long after it has completely fit the training data. This transition often occurs abruptly and has been observed across a range of synthetic and realistic scenarios. By investigating Grokking in simplified models—such as linear or near-linear systems—we leverage tools from statistical mechanics, random matrix theory, and optimization to gain insights. We claim that Grokking is a near-critical phenomenon, appearing near singularities in the "thermodynamic limit" (in some proper sense) of training dynamics. Specifically, delayed generalization stems from “critical slowing down”, a generic feature of systems near phase transitions, that induces slow dynamical time scales near almost-stable fixed points. This perspective enables us to derive concrete scaling predictions that are applicable even in more complex settings, owing to the universal nature of critical phenomena.
- View this event in Google Calendar

Tuesday, December 10th

Quantum Physics Seminar: ""
- Start time: 12:00pm
- End date: Tuesday, December 10th
- End time: 01:30pm
- Lecturer: Hagai Diamandi (Applicant) Abstract:
- View this event in Google Calendar
astrolunch: Fabio Pacucci
- Start time: 05:00pm
- End date: Tuesday, December 10th
- End time: 06:00pm
- Title: JWST's Little Red Dots: Masters of Disguise in the High-Redshift Universe This talk will explore the exciting journey to probe several populations of black holes, spanning from the distant to the local Universe. I will focus primarily on the "little red dots" (LRDs), a newly discovered and puzzling population of compact red sources at redshift z > 4, identified by JWST. These objects challenge our current astrophysical models in several ways. First, I will discuss the detection of overmassive black holes relative to the stellar mass of their host galaxies. Second, I will address the X-ray weakness problem, where these sources remain undetected in deep X-ray surveys. I will use advanced GRRMHD simulations to show how mildly super-Eddington accretion onto slowly spinning black holes can resolve this issue. Third, I will explore the extremely high stellar densities at the cores of these objects and their potential for triggering runaway stellar collisions. Throughout the talk, I will also discuss how the LRDs represent a perfect bridge to study other populations of black holes, with a focus on the role of future observatories such as HWO and AXIS. These include the search for black hole seeds at redshifts z = 20-30 and the detection of quiescent black holes in the local Universe.Bio: Fabio Pacucci is a Clay and BHI Fellow at the Center for Astrophysics. He obtained his PhD in Physics from the Scuola Normale Superiore in Pisa, Italy.He is an expert in the theory and observations of high-z black holes and is part of several large JWST programs.He is also involved in planning next-generation observatories, such as NASA's AXIS in the X-ray, which was just selected for Phase A, and HWO in the UV, optical, and infrared, which will be NASA's next flagship mission.contact: Zhaozhou
- View this event in Google Calendar

Search form

People by Discipline

Pages