Physics

How did the universe begin? How did it evolve to what we see now?

There was a time when few people believed such questions could even be posed in scientific terms. Now, as increasingly precise instruments deliver their treasure trove of data, the answers may be within reach.

On Wednesday, October 25, Perimeter Director Emeritus Neil Turok will tackle this intriguing topic in a Perimeter Institute Public Lecture, “Secrets of the Universe: Hiding in Plain Sight?”

The performance of neural networks like large language models (LLMs) is governed by "scaling laws": the error of the network, averaged across the whole dataset, drops as a power law in the number of network parameters and the amount of data the network was trained on. While the mean error drops smoothly and predictably, scaled up LLMs seem to have qualitatively different (emergent) capabilities than smaller versions when one evaluates them at specific tasks. So how does scaling change what neural networks learn? We propose the "quantization model" of neural scaling, where smooth power laws in mean loss are understood as averaging over many small discrete jumps in network performance. Inspired by Max Planck's assumption in 1900 that energy is quantized, we make the assumption that the knowledge or skills that networks must learn are quantized, coming in discrete chunks which we call "quanta". In our model, neural networks can be understand as being implicitly a large number of modules, and scaling simply adds modules to the network. In this talk, I will discuss evidence for and against this hypothesis, its implications for interpretability and for further scaling, and how it fits in with a broader vision for a "science of deep learning".

---

Zoom link https://pitp.zoom.us/j/93886741739?pwd=NzJrcTBNS2xEUUhXajgyak94LzVvdz09

I argue that the answer is yes, by reviewing the history and current status of such a theory.  Since 1982, I have been developing a series of such theories, beginning in 1982 with an N --> infinity limit of 2+1 dimensional matrix model (the IAS model), through another N --> infinity, T --> 0 limit of a BFSS model.  During this time our work was complemented by Adler's Trace model and others.   

Beginning in 2012, Cortes and I developed a different approach to an relational quantum cosmology by adding intrinsic energy momentum to Sorkin et al's causal set models.  The addition of energy- momentum as intrinsic variables opens up a new mechanism for the emergence of space, and spacetime, plus interacting relativistic particles.  Note that the warm phase is purely algebraic, you need no prior existence of any space to get other dimensions to emerge.   In 2021 I discovered how to derive quantum non-relativistic many body theory from what has since been called the Causal Theory of Views.  Finally in papers since we report progress on the construction of special and general relativistic  Causal Theory of Views.

---

Zoom link: https://pitp.zoom.us/j/95891848248?pwd=TUpWK2RWbU9GTGxZS1lMeS81QWp1dz09

In this colloquium, the communicative boundary between science and society in Africa will be explored, using physics as an illustration. Different types of contact zones and scenes will be discussed, as well as the societal impact of staging and narrative formats. Based on our ongoing engagement project Making Science the Star, an overview of the relationship between sender and receiver will be presented, and recipes for tuning this interaction to unlock untapped potential in predominantly non-scientific communities will be analyzed.

DISCLAIMER: This talk contains Season1 Episode2 from the show series Science In The City. This is used with permission from Dr. Stéphane Kenmoe, Producer of the series and Presenter of this talk.

---

Bio: Stéphane Kenmoe got a Ph.D. in Physics from the Max Planck Institute for Iron Research in Germany in 2015. He is presently a habilitation candidate at the Faculty of Chemistry at the University of Duisburg-Essen, Germany. He is also a science popularizer on TV and on social media, a science writer and a novelist. In 2020, he produced the movie ‘’Science in the City’’ (Science dans la Cité) in Cameroon. He is very active in networking for the promotion of early career African scientists and for connecting science and society. He has won many awards for his engagement, among which the 2020 Diversity Prize for Academic Leadership at the University of Duisburg-Essen, the Falling Walls Award for Science Engagement in 2021 and the Award for Prototypes of Humanity of the Dubai Authority for arts and culture in 2022. Since 2022, he is the chief editor of the African Physics Newsletter, an electronic quarterly about physics in Africa published by the American Physical Society. 

---

Zoom link: https://pitp.zoom.us/j/92725978684?pwd=Yys0ci9JdE0zS054SGxyaWoxZkdUUT09

During this colloquium, I will discuss my journey in looking for astronomical information hiding in the ancestral knowledge of my community. I will show concrete examples of my findings and encourage communities to engage in similar practice to provide content that could be used to teach indigenous Astronomy in classrooms. 

---

Bio:

Laurie Rousseau-Nepton is a new faculty at the University of Toronto and the Dunlap Institute for Astronomy and Astrophysics. She comes with six years of experience working as a resident astronomer at the Canada-France-Hawaii Observatory supporting various instruments including wide-field cameras, high-resolution spectrographs, Fourier Transform Spectro-imager. She received her diploma from Université Laval by studying regions of star formation in spiral galaxies and helping with the development of two Fourier Transform Spectro-imagers, SpIOMM and SITELLE. She is now leading an international project called SIGNALS, the Star formation,Ionized Gas, and Nebular Abundances Legacy Survey, which sampled with the SITELLE instrument more than 50,000 of star-forming regions in 40 nearby galaxies to understand how the local environment affect the young star clusters characteristics.

This seminar will be divided in two segments: 1) New Instrumentation for Astronomy and 2) the SIGNAL-Survey of Star-forming regions in Nearby Galaxy.

1) Evolution of technologies and optics manufacturing technics are providing new interesting options for the design of astronomical instruments to increase precision and add new capabilities. In this presentation, I will discuss my new laboratory plan at the University of Toronto to include Micro-kinetic inductance detector arrays and meta-surface optics to a Fourier Transform Imaging spectrograph design. The goal is to reach high-spectral resolution (R:15,000 to 80,000) over a large field-of-view, while keeping high sensitivity.

2) SIGNALS stands for the Star formation, Ionized Gas, and Nebular Abundances Legacy Survey. Using a Fourier Transform Imaging Spectrograph SITELLE, at the Canada-France-Hawaii Telescope, we observed 40 nearby galaxies and covered over 50,000 star-forming regions in different environment at a spatial resolution from 0.5 to 40 pc. Covering several emission line spectral features including Halpha (at R: 5,000), the survey aims at characterizing the star-forming sites and their environments to produce the most complete and well resolved database on star formation.

---

Zoom link https://pitp.zoom.us/j/94273599584?pwd=TUY3UFpVa20wbkJUcEdoTmlYUzlQUT09

---

Bio:

Laurie Rousseau-Nepton is a new faculty at the University of Toronto and the Dunlap Institute for Astronomy and Astrophysics. She comes with six years of experience working as a resident astronomer at the Canada-France-Hawaii Observatory supporting various instruments including wide-field cameras, high-resolution spectrographs, Fourier Transform Spectro-imager. She received her diploma from Université Laval by studying regions of star formation in spiral galaxies and helping with the development of two Fourier Transform Spectro-imagers, SpIOMM and SITELLE. She is now leading an international project called SIGNALS, the Star formation, Ionized Gas, and Nebular Abundances Legacy Survey, which sampled with the SITELLE instrument more than 50,000 of star-forming regions in 40 nearby galaxies to understand how the local environment affect the young star clusters characteristics. 

With rapid progress in simulation of strongly interacting quantum Hamiltonians, the challenge in characterizing unknown phases becomes a bottleneck for scientific progress. We demonstrate that a Quantum-Classical hybrid approach (QuCl) of mining the projective snapshots with interpretable classical machine learning, can unveil new signatures of seemingly featureless quantum states. The Kitaev-Heisenberg model on a honeycomb lattice with bond-dependent frustrated interactions presents an ideal system to test QuCl. The model hosts a wealth of quantum spin liquid states: gapped and gapless Z2 spin liquids, and a chiral spin liquid (CSL) phase in a small external magnetic field. Recently, various simulations have found a new intermediate gapless phase (IGP), sandwiched between the CSL and a partially polarized phase, launching a debate over its elusive nature. We reveal signatures of phases in the model by contrasting two phases pairwise using an interpretable neural network, the correlator convolutional neural network (CCNN). We train the CCNN with a labeled collection of sampled projective measurements and reveal signatures of each phase through regularization path analysis. We show that QuCl reproduces known features of established spin liquid phases and ordered phases. Most significantly, we identify a signature motif of the field-induced IGP in the spin channel perpendicular to the field direction, which we interpret as a signature of Friedel oscillations of gapless spinons forming a Fermi surface. Our predictions can guide future experimental searches for U(1) spin liquids.

--

Zoom link: https://pitp.zoom.us/j/94233944575?pwd=OVljLzMrZzlKeUErNHZQRkEzMFRKUT09

The aim of the Hammers & Nails climate workshop was to give researchers at Perimeter a unique opportunity to interact with invited experts from diverse fields related to climate, energy and biodiversity. The first half of the workshop was dedicated to short research pitches, where both PI researchers as well as invited guests from the Math4Climate network (including e.g. Waterloo Center for Innovation and Complexity, UW's Climate Institute, Energy Institute, Water Institute) could present their hammers & nails. Nails refers to open challenges within climate, energy & biodiversity research. Hammers refers to methods used by PI researchers in their daily work (e.g. causal inference, techniques in algebraic topology, strong gravity simulations etc.). The second half of the workshop built on these pitches through informal group discussions. Conversations from the workshop may be continued in the future through informal lunches and reading groups.

The aim of the Hammers & Nails climate workshop was to give researchers at Perimeter a unique opportunity to interact with invited experts from diverse fields related to climate, energy and biodiversity. The first half of the workshop was dedicated to short research pitches, where both PI researchers as well as invited guests from the Math4Climate network (including e.g. Waterloo Center for Innovation and Complexity, UW's Climate Institute, Energy Institute, Water Institute) could present their hammers & nails. Nails refers to open challenges within climate, energy & biodiversity research. Hammers refers to methods used by PI researchers in their daily work (e.g. causal inference, techniques in algebraic topology, strong gravity simulations etc.). The second half of the workshop built on these pitches through informal group discussions. Conversations from the workshop may be continued in the future through informal lunches and reading groups.

Complementing the spectacular breakthroughs in gravitational wave and multi-messenger astronomy, advancements in our theoretical understanding and modelling of the strong gravity are essential to apprehending the nonlinear dynamics of spacetime, and to unlocking the full potential of the observations. I will illustrate the scope of new physics that might be probed by black holes, neutron stars, and other such systems, including testing general relativity and strong gravity signatures of new particles, and describe some recent developments that allow for uncovering novel phenomena and making detailed predictions in this regime.

Zoom Link: https://pitp.zoom.us/j/96180256322?pwd=LzEyd3ZBWGdZeDR6MjB1dGpLWFRpUT09

There are fascinating new insights we could achieve if we succeeded in connecting quantum gravity to particle physics in the visible and the dark universe. However, making such a connection is challenging, because the typical scale of quantum gravity is believed to be far from the typical scales in particle physics. I will exploit lever arms that could make it possible to bridge this gap in scales.  To provide concrete examples for these ideas, I will review recent results on the proton lifetime in quantum gravity, the effects of asymptotically safe quantum gravity on properties of Standard Model matter and the effects of asymptotically safe quantum gravity on simple models of dark energy and dark matter.

Zoom Link: https://pitp.zoom.us/j/96545839038?pwd=UG1IVzJPWUZGa2ovOFhzdHBhOFhOdz09

Tensor Processing Units are application specific integrated circuits (ASICs) built by Google to run large-scale machine learning (ML) workloads (e.g. AlphaFold). They excel at matrix multiplications, and hence can be repurposed for applications beyond ML. In this talk I will explain how TPUs can be leveraged to run large-scale density matrix renormalization group (DMRG) calculations at unprecedented size and accuracy. DMRG is a powerful tensor network algorithm originally applied to computing ground-states and low-lying excited states of strongly correlated, low-dimensional quantum systems. For certain systems, like one-dimensional gapped or quantum critical Hamiltonians, or small, strongly correlated molecules, it has today become the gold standard method for computing e.g. ground-state properties. Using a TPUv3-pod, we ran large-scale DMRG simulations for a system of 100 spinless fermions, and optimized matrix product state wave functions with a bond dimension of more than 65000 (a parameter space with more than 600 billion parameters). Our results clearly indicate that hardware accelerator platforms like Google's latest TPU versions or NVIDIAs DGX systems are ideally suited to scale tensor network algorithms to sizes that are beyond capabilities of traditional HPC architectures.

Zoom link:  https://pitp.zoom.us/j/99337818378?pwd=SGZvdFFValJQaDNMQ0U1YnJ6NU1FQT09