Eric Sembrat's Test Bonanza

Abstract

For over a century, researchers have been investigating collective cognition, in which a group of individuals together processes information and acts as a single cognitive unit. However, we still know little about the circumstances under which groups achieve better (or worse) decisions than individuals. To address this question my research applies concepts and methods from psychology to both individuals and groups in order to directly compare their cognitive abilities.

As model systems I use house-hunting by the ant Temnothorax rugatulus, and navigation by the homing pigeon, Columba livia. My work has shown that 1) rational group decisions can emerge from interactions among irrational individuals, 2) groups can process more information and thus are less prone to cognitive overload than individuals, and 3) groups show more precise discrimination than individuals, but individuals make better decisions than groups for easy discrimination tasks.

Furthermore, my recent research has shown, for the first time in a non-human species, that group decisions are influenced by past experiences and improve progressively, a phenomenon known as cumulative cultural evolution. By combining empirical data and models I elucidate the emergent processes of collective cognition and suggest how and when groups (fail to) achieve higher cognitive performance than individuals.  

Fluid turbulence is one of the greatest unsolved problems of classical physics (and the subject of a million dollar mathematical (Millenium) challenge).  Centuries of research--including Leonardo da Vinci’s observations of “la turbolenza” and the best efforts of numerous physicists (Heisenberg, Kelvin, Rayleigh, Sommerfeld, ...)--have failed to yield a tractable predictive theory.  However, recent theoretical and computational advances have successfully linked recurring transient patterns (coherent structures) within turbulence to unstable solutions of the equations governing fluid flow (the Navier-Stokes equations).  The solutions describing coherent structures provide a geometrical structure that guides the evolution of turbulence.   We describe laboratory experiments where the geometry of key coherent structures is identified and harnessed to construct a roadmap to forecast the behavior of weakly turbulent flows.

BIO

Michael F. Schatz is a professor and associate chair for the introductory physics program in the School of Physics at the Georgia Institute of Technology.   In 1991, Schatz received his PhD in physics at the University of Texas, Austin; he joined the faculty of Georgia Tech in 1996.  Schatz conducts research in both experimental nonlinear dynamics and physics education. He is currently a Director of the Hands-on Research in Complex Systems Schools at the International Centre for Theoretical Physics (ICTP) in Trieste, Italy.  He is a recipient of the Cottrell Scholars Award and a Fellow of the American Physical Society.

The 2016 confirmation of Einstein's prediction of gravitational waves has put the spotlight back on the importance of curvature for the physics of the universe. While the ability of mass to curve our space has fueled the imagination of many, it is by far not the only instance of warped spaces being important for physics: The materials science of the very small scale -the science of nanostructures and nanoengineering- is one of them. In fact, often these 'small' spaces are very strongly curved, far from what mathematicians call 'Euclidean'; for example two parallel lines may no longer only meet at infinity. Bizarre and exotic spaces with very unusual properties. Until recently, many of these complex spaces defied most people's imagination, but Virtual Reality technology has now been developed to help us immerse in them. Prof Sabetta Matsumoto will take us on a tour -enabled by the latest in Virtual Reality technology- into the innate beauty and mystery of some spaces, such as the cross between a Euclidean straight line and Poincare's hyperbolic plane made popular by Escher's artwork. Real-world applications or technological uses of these mathematical insights may seem to be light-years off, but don't worry, the real world will catch up with the imagination faster than we think.

Lecture begins at 6:30 PM and stay after the talk for a Virtual Reality Demo at 7:30 PM!

Abstract:

Almost all animals must move to survive and reproduce. But it's rarely easy to move in the wild - animals typically face multiple challenges when trying to get from A to B. This talk will present loosely connected work centered on how gait, gait regulation, and gait adaptation in many-legged animals can give insight into several interesting aspects of locomotor biology.

Work showing that dog gait dynamics can be parsimoniously predicted using symmetry considerations, but only after including a constant phase shift between fore and hind limbs, that in turn depends on dog aspect ratio, will be presented. Subsequent work examining dogs walking on rough terrain, mice being perturbed externally via earthquakes and internally via muscle stimulation, and spiders being perturbed through autotomy (self-amputation), will be discussed.

The collective results suggest that gait regulation is gait specific (walking regulation varies from trotting regulation), that the gait specific variance may reflect static vs dynamic constraints, and that gait adaptation can occur on the temporal dimension (e.g. limping), and not just in phase relationships, or spatial changes.

Abstract

Gene therapy holds the promise of treatment of numerous diseases including many types of cancer, cardiovascular diseases and genetic disorders. Even though methods for gene delivery have been an active research area since early 90s, no gene therapeutic agents have been FDA approved for use in humans. The progress in gene therapy has been hindered by lack of safe, predictable, and reliable methods for packaging, delivery, and transport of genetic material.

Efficient wrapping or packaging of DNA is a critical part enabling gene delivery, where nucleic acids are transported across cell membranes with the help of transfection vectors such as proteins, cationic dendrimers or nanoparticles. Because DNA/RNA transfection is dependent on the size, shape, and surface properties of the DNA/RNA-vector complex, control over assembly structure is critical for creating effective transfection agents. Evolving nanomaterials to the clinic requires optimization, which is prohibitively expensive, and a mechanistic understanding of carriers-NA interactions, which remains unknown.  Our group attempts to advance tailored materials gene delivery by a multiscale optimization employing all-atom molecular dynamics (MD) simulations, leveraging machine learning algorithms and employing dissipative particle dynamics (DPD) simulations.

In this talk, I’ll discuss two avenues for designing nanomaterials for gene delivery: the design of ligand functionalized inorganic nanoparticles and self-assembling DNA-based nanomaterials. We employed atomistic molecular dynamics simulations to understand the binding of nucleic acids to monolayer-protected gold nanoparticles. Results from these simulations were analyzed to determine modes of DNA and RNA bending with nanoparticles. These results allowed us to determine the training data for machine learning algorithms and design novel ligands capable of controlled wrapping of NA around NP.

The information from MD simulations was used to parameterize and developed a DPD-based model, which allows for large-scale simulations of self-assembling polyelectrolytes materials and their morphological response to the changes in salt concentration and applied this method for the prediction of responsive morphologies of DNA-based micelles and gels. Our results will enable design of more efficient gene delivery systems with enhanced biocompatibility and selectivity.

ABSTRACT

The Materials Genome Initiative (MGI) has heralded a sea change in the philosophy of materials design. Here, we highlight the importance of computational data generation and screening, targeted synthesis and characterization, polymer fingerprinting, machine-learning prediction models, and the creation of an online Polymer Informatics platform (https://www.polymergenome.org) to guide ongoing and future polymer discovery and design. We lay special emphasis on the fingerprinting of polymers in terms of their genome or constituent atomic and molecular fragments, an idea that pays homage to the pioneers of the human genome project who identified the basic building blocks of the human DNA. By scoping the polymer genome, we present an essential roadmap for the design of polymer dielectrics, and provide future perspectives and directions for expansions to other polymer subclasses and properties.

Abstract

Some of the main challenges in nanoscience and nanotechnology are the reproducibility and uniformity of the properties of nanostructures. These are crucial for the advancement of this field from the fundamental point of view and future applications. The study of metallic nanoparticles passivated with thiolate ligands has rapidly grown over the last two decades. Synthesis methods have allowed obtaining nanoparticles with a precise number of atoms, which ensures the uniformity and reproducibility of their physical properties.

A large number of atomically exact ligand-protected nanoclusters (NCs) have been fabricated using gold, silver and both species. Here, I analyze the criteria that have been used to explain the experimental stability of metal NCs with a precise number of atoms protected with thiolate ligands. Based on experimental evidence, I show that these criteria are not enough to explain stability and thus, I propose to explore in detail the role of the ligand composition on the stabilization of these NC, which has not been adequately considered.

Additionally, these NCs are good candidates to be used to increase the sensitivity of chiroptical techniques, such as circular dichroism (CD) to characterize chiral molecules. CD signals are tiny in intensity limiting the analysis to samples with large concentrations of chiral molecules, and large enantiomeric excess, being this latter difficult and expensive to achieve. Thus, the increment of the detection limit in chiral spectroscopies would have a significant impact in pure and applied sciences.

It was found that optical activity increases when some chiral molecules are adsorbed in metallic NPs. However, CD signals might significantly differ from that measured of the molecules alone. Although significant, CD response of molecules adsorbed in metallic NPs is yet poorly understood. Here, I want to respond to some open questions. In particular, we investigate how the intrinsic chirality of the molecule, the number of molecules, the specie of the metallic NC, and the induced chiral arrangement of the molecules affect the optical activity in metallic NCs.

Abstract

What is the use of half a wing? More generally, how do novel structures and behaviors evolve? Controlled aerial behaviors in ant workers and other insects of the tropical rain forest canopy demonstrate directed gliding in the complete absence of wings. Importantly, tree-dwelling bristletails (the sister group to the winged insects) also exhibit aerial righting responses and directed gliding while falling. Ontogenetic, biomechanical, paleontological, and phylogenetic analyses suggest that controlled aerial behaviors preceded the origin of wings in vertebrates as well, indicating functional aerodynamics for only partially feathered structures and for rudimentary flapping kinematics.

Use of a robotic Archaeopteryx similarly demonstrates biomechanical functionality of the intermediate winged condition, consistent with arboreal and gravitationally assisted origins of flight in all volant taxa.  I will also present in this talk recent work on aerial maneuverability in hummingbirds, describing a variety of experimental perturbations to elicit extreme examples of flight control (e.g., flight through apertures, in heavy rain, in high turbulence, and at high elevations).