BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CMSA - ECPv6.15.18//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:CMSA
X-ORIGINAL-URL:https://cmsa.fas.harvard.edu
X-WR-CALDESC:Events for CMSA
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20220313T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20221106T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20230312T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20231105T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20240310T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20241103T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231115T103000
DTEND;TZID=America/New_York:20231115T113000
DTSTAMP:20260419T110543
CREATED:20240222T105852Z
LAST-MODIFIED:20240222T105913Z
UID:10002803-1700044200-1700047800@cmsa.fas.harvard.edu
SUMMARY:A bulk gap in the presence of edge states for a truncated Haldane pseudopotential
DESCRIPTION:Topological Quantum Matter Seminar \nSpeaker: Amanda Young\, UIUC \nTitle: A bulk gap in the presence of edge states for a truncated Haldane pseudopotential \nAbstract: Haldane pseudopotentials were first introduced as Hamiltonian models for the fractional quantum Hall effect\, and it has been long expected that they should exhibit the characteristic properties of this exotic phase of matter\, including a spectral gap above the ground state energy. We will discuss recent work that verified this gap conjecture for a truncated version of the 1/3-filled Haldane pseudopotential in the cylinder geometry. Numerical evidence suggested that for open boundary conditions the gap of the truncated model closes as the cylinder radius converges to zero and that this closure is due to the presence of edge modes; in contrast\, for periodic boundary conditions\, the gap remains robustly order one in the same radius limit. The standard scheme for applying spectral gap estimating techniques to the model with periodic boundary conditions\, though\, produces a lower bound on the bulk gap that still reflects the energy of the edge modes. To obtain an estimate on the bulk gap that reflects its true behavior\, a new gap estimating strategy was developed. By customizing the spectral gap method to key invariant subspaces of the Hamiltonian\, we are able to successfully avoid the edge states and produce a more accurate lower bound on the bulk gap. In this talk\, we discuss this invariant subspace strategy for proving bulk gaps in the presence of edge states. This is based off joint work with S. Warzel. \n  \n 
URL:https://cmsa.fas.harvard.edu/event/tqms_111523/
LOCATION:Virtual
CATEGORIES:Topological Quantum Matter Seminar
ATTACH;FMTTYPE=image/png:https://cmsa.fas.harvard.edu/media/CMSA-Topological-Seminar-11.15.23.png
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231115T140000
DTEND;TZID=America/New_York:20231115T150000
DTSTAMP:20260419T110543
CREATED:20240222T094758Z
LAST-MODIFIED:20240222T095355Z
UID:10002797-1700056800-1700060400@cmsa.fas.harvard.edu
SUMMARY:On the Power of Forward pass through Transformer Architectures
DESCRIPTION:New Technologies in Mathematics Seminar \nSpeaker: Abhishek Panigrahi\, Dept. of Computer Science\, Princeton University \nTitle: On the Power of Forward pass through Transformer Architectures \nAbstract: Highly trained transformers are capable of interesting computations as they infer for an input. The exact mechanism that these models use during forward passes is an interesting area of study. This talk studies two interesting phenomena. \nIn the first half\, we explore how and why pre-trained language models\, specifically BERT of moderate sizes\, can effectively learn linguistic structures like parse trees during pre-training. Specifically\, using synthetic data through PCFGs\, we show how moderate-sized transformers can perform forward-backward parsing\, also known as the inside-outside algorithm\, during inference. We further understand the role of the pre-training loss for the model to learn to parse during pre-training. \nIn the second half\, we consider in-context learning of large language models\, where they learn to reason on the fly. An ongoing hypothesis is that transformers simulate gradient descent at inference to perform in-context learning. We propose the Transformer in Transformer (TinT) framework\, which creates explicit transformer architectures that can simulate and fine-tune a small pre-trained transformer model during inference. E.g. a 1.3B parameter TINT model can simulate and fine-tune a 125 million parameter model in a single forward pass. This framework suggests that large transformers might execute intricate sub-routines during inference\, and provides insights for enhancing their capabilities through intelligent design considerations. \n 
URL:https://cmsa.fas.harvard.edu/event/nt-111523/
LOCATION:CMSA Room G10\, CMSA\, 20 Garden Street\, Cambridge\, MA\, 02138\, United States
CATEGORIES:New Technologies in Mathematics Seminar
ATTACH;FMTTYPE=image/png:https://cmsa.fas.harvard.edu/media/NTM-11.15.2023.png
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20231115T153000
DTEND;TZID=America/New_York:20231115T163000
DTSTAMP:20260419T110543
CREATED:20240223T053940Z
LAST-MODIFIED:20240223T054457Z
UID:10002818-1700062200-1700065800@cmsa.fas.harvard.edu
SUMMARY:Thresholds
DESCRIPTION:Probability Seminar \nSpeaker: Jinyoung Park (NYU) \nTitle: Thresholds \nAbstract: For a finite set X\, a family F of subsets of X is said to be increasing if any set A that contains B in F is also in F. The p-biased product measure of F increases as p increases from 0 to 1\, and often exhibits a drastic change around a specific value\, which is called a “threshold.” Thresholds of increasing families have been of great historical interest and a central focus of the study of random discrete structures (e.g. random graphs and hypergraphs)\, with estimation of thresholds for specific properties the subject of some of the most challenging work in the area. In 2006\, Jeff Kahn and Gil Kalai conjectured that a natural (and often easy to calculate) lower bound q(F) (which we refer to as the “expectation-threshold”) for the threshold is in fact never far from its actual value. A positive answer to this conjecture enables one to narrow down the location of thresholds for any increasing properties in a tiny window. In particular\, this easily implies several previously very difficult results in probabilistic combinatorics such as thresholds for perfect hypergraph matchings (Johansson–Kahn–Vu) and bounded-degree spanning trees (Montgomery). I will present recent progress on this topic. Based on joint work with Keith Frankston\, Jeff Kahn\, Bhargav Narayanan\, and Huy Tuan Pham.
URL:https://cmsa.fas.harvard.edu/event/probability-111523/
LOCATION:CMSA Room G10\, CMSA\, 20 Garden Street\, Cambridge\, MA\, 02138\, United States
CATEGORIES:Probability Seminar
ATTACH;FMTTYPE=image/png:https://cmsa.fas.harvard.edu/media/CMSA-Probability-Seminar-11.15.23.png
END:VEVENT
END:VCALENDAR