לוח שנה משולב

The analog computation at the chemical synapse that underlies cognition depends on a highly compartmentalized, tightly regulated, and complex network of interactions between synaptic activity and hundreds of proteins and the mechanisms that regulate them. For a top-down study of how the network operates in concert, I present a modular, versatile approach for combined imaging of multiprotein composition, activation states, ion and neurotransmitter fluxes, and mRNA translation across the same individual synapses. I will show how this approach extends to other subcellular systems such as mitochondria. I will discuss the use of Bayesian network inference to extract biological insight from high-dimensional, multimodal synapse distributions. Finally, I will present applications of this approach to identify convergent molecular phenotypes across autism and schizophrenia-associated genes, and for an in-depth study of the complex synaptic response to genetic and chemical perturbations of GluN2A.

In this talk I will present a useful construction introduced by Kassabov and Pak, called diagonal products. They arise naturally in the study of the space of marked groups, that is, the normal subgroups of a free group with the Chabauty topology. It turns out that diagonal products are extremely rich, and proved to be a useful tool for providing a full spectrum of various growth functions for groups. This includes answers to questions such as "How fast do Følner sets grow in an amenable group" and "How fast do residual chains grow in residually finite groups" etc. We will elaborate on a joint work with Arie Levit and Itamar Vigdorovich, where we give a general classification result for characters (á la Thoma) of many diagonal products. As a result, we deduce that there are uncountably many Hilbert–Schmidt stable groups, which are as unstable as one wants.

A $k$-dominating set ($k$-DS) is a node subset of distance at most $k$ from any node in the graph.
The task of constructing a small (i.e., existentially optimal) $k$-DS was suggested by Kutten and Peleg (Journal of Algorithms, 1998) as a useful primitive for distributed graph algorithms and received considerable attention ever since.
In the current paper, we advance the state-of-the-art of distributed small $k$-DS construction by presenting a randomized asynchronous CONGEST KT1 algorithm for this task that runs in $ ilde{O}( k^{2} )$ time and sends $ ilde{O}( n k )$ messages whp.
When $k leq polylog( n )$, we obtain the first asynchronous algorithm for a non-trivial distributed task whose communication cost is near-linear (in $n$).

The talk will be self contained.

This presentation will outline recent evolution of AI methodologies, focusing on the emergence of Diffusion Models of AI inspired by non-equilibrium statistical mechanics, Transformers, and Reinforcement Learning. These innovations are revolutionizing our approach to reduced, Lagrangian turbulence modeling and are instrumental in formulating and solving new challenges, such as swimming navigation in chaotic environments.

More generally, attendees will gain insights into the synergy between AI and natural sciences and understand how this symbiosis is shaping the future of scientific research. This comprehensive vision is relevant to theoretical physicists, applied mathematicians, and computer scientists alike.

Given that black holes are by definition not directly observable, physicists had to devise indirect techniques to test their reality. These efforts were recognized by 6 Nobel prizes in 2017 and 2020. The last three laureates included R. Penrose, cited for his famous singularity  theorem, a real and strikingly beautiful mathematical proof. That recognition illustrates the role of mathematics in testing the reality of physical objects by proving or disproving specific mathematical  conjectures. In my talk I will address the issue of the stability of Kerr black holes, another precise conjecture that can be decided by pure mathematical techniques.

The human species is on a great scientific quest — to understand the neural mechanisms of human (primate) intelligence.  Recent progress in multiple subfields of brain research suggests that key next steps in this quest will result from building real-world capable, systems-level network models that aim to abstract, emulate and explain the primate neural mechanisms underlying natural intelligent behavior.  In this talk, I will briefly outline the story of how neuroscience, cognitive science and computer science (“AI”) converged to create specific, image-computable, deep neural network models intended to appropriately abstract, emulate and explain the mechanisms of primate core visual object identification and categorization.  Based on a large body of primate neurophysiological and behavioral data, some of these network models are currently the leading (i.e. most accurate) scientific theories of the internal mechanisms of the primate ventral visual stream and how those mechanisms support the ability of humans and other primates to rapidly and accurately infer latent world content (e.g. object identity, position, pose, etc.) from the set of pixels (image) received under typical (brief) natural viewing.

While still far from complete, because these leading neuroscientific models are fully observable and machine-executable, they offer predictive and potential application power that the field’s prior conceptual models did not.   I will describe two recent examples from our team.  First, I will show that — relative to the gold standard of primate brains and minds — the leading models are both similarly sensitive to and similarly robust to adversarial attack at both their neural levels and their behavioral levels.  Second, I will describe initial empirical tests of the closely related possibility of using such models to design spatial patterns of light energy on the retina (i.e. customized, synthetic images) to precisely, and non-invasively modulate neuronal activity deep in the primate brain.  Consistent with model predictions, these tests reveal surprisingly strong and precise neural population effects.  Besides being a tool for neuroscience, we see this as an exciting new application avenue of potential human clinical benefit.

Microorganisms play a crucial role in the weathering processes that transform rock into soil through chemical and physical mechanisms essential for nutrient cycling, nitrogen fixation, carbon storage, and organic matter decomposition. This intricate relationship between microbial life and landscapes forms the backbone of ecosystem dynamics and biogeochemical processes. Microbes influence rock weathering and soil production, adapting to their surroundings and creating distinct communities across various landscapes. These complex interactions and feedback mechanisms are pivotal to understanding the co-evolution of microbial communities and landscapes over time. However, existing research on microbial contributions to weathering and soil production has predominantly focused on relatively short timescales and small spatial scales. Understanding the interplay between the evolution of microbial communities and their role in weathering processes over geomorphic timescales within transient landscapes is important for a more complete understanding of how landscapes evolve as well as the impact of geomorphic changes on microbial community establishment and evolution.
The main objective of this study is to elucidate the long-term dynamics of microbial communities and their role in weathering processes over millennial timescales. To achieve this, we focused on recently deglaciated basins in the Eastern Sierra Nevada, CA, examining bacterial community composition in three phases of the weathering process: exposed rock at the surface, saprolite—the weathered rock found beneath soil, and soil. Sampling along an elevational transect, we collected 25 samples of rock, soil, and saprolite, and evaluated their bacterial composition using 16S rRNA and metagenomic sequencing.
Results show that both soil and saprolite samples exhibited diverse and similar microbial communities, indicating a developmental relationship between these habitats despite distinct geochemical compositions. In contrast, rock habitats are less diverse, and their composition resembles those of young deglaciated landscapes. Our findings point to a link between microbial community composition and rock-to-soil weathering processes, suggesting that the majority of weathering processes occur within the soil column (saprolite and soil), with exposed rock maintaining a steady state.
The stability of these microbial communities over extended timescales suggests a potentially significant role for microbial weathering in landscape evolution. This finding underscores the importance of considering microbial contributions in future geomorphic studies, as they may play a key role in shaping the Earth's surface. Moving forward, we plan on coupling a long-term, landscape-scale geomorphic perspective with 'omics approaches from microbial ecology to comprehensively understand the complex relationships between microbial life and landscapes, ultimately advancing our knowledge of ecosystem dynamics and health.

We prove the set of growth-rates of subgroups of a rank~r free group is dense in [1,2r−1]. Our main technical contribution is a concentration result for the leading eigenvalue of the non-backtracking matrix in the configuration model.

 

Based on a joint work with Michali Louvaris and Dani Wise.