Integrated Calendar

Neuromorphic computing designs have an important role in the modern ‘big data’ era, as they are suitable for processing large amount of information in short time, eliminating the von Neumann (VN) bottleneck. The neuromorphic hardware, taking its inspiration from the human brain, is designed to be used for artificial intelligence tasks via physical neural networks, such as speech or image recognition, bioinformatics, visual art processing and much more. The memristor (memory + resistor), is one of the promising building blocks for this hardware, as it mimics the behavior of a human synapse, and can be used as an analog non-volatile memory. The memristor has been proven as a viable memory element and has been used for constructing resistive random access memory (RRAM) as a replacement for current VN hardware. However, the mechanism of operation and the conducting bridge formation mechanisms in electrochemical metallization memristors still require further investigation. A planar single-nanowire (NW) based memristor is a good solution for elucidating the mechanism of operation, thanks to the high localization of switching events, allowing in-situ investigation as well as post-process analysis. Our group, which has developed the guided-growth approach to grow guided planar NWs on different substrates, has used this method to integrate guided epitaxial NWs into functional devices such as field-effect transistors (FETs), photodetectors and even address decoders. However, the guided-growth approach has not been used for creating memristors up to date.
In this work, I successfully synthesized guided NWs of two metal-oxides on flat and faceted sapphire substrates – ZnO and β-Ga2O3 were successfully grown in the VLS mechanism as surface guided NWs. I successfully grew planar guided β-Ga2O3 NWs on six different sapphire substrates, for the first time as far as we know. We characterized the newly grown β-Ga2O3 NWs with SEM, TEM, EDS and Raman spectroscopy. The monoclinic NWs grew along surprising directions on the flat sapphire surfaces and I demonstrated a new mode of growth – epitaxy favored growth on a faceted surface, when graphoepitaxy is also possible. I created electrochemical metallization memristors with the obtained NWs and successfully demonstrated the effect of resistive switching for β-Ga2O3 guided NW based devices. With the abovementioned achievements, we expanded the guided-growth approach on flat and faceted sapphire surfaces, and opened the opportunity for creating surface guided-NW based neuromorphic hardware.

Animals generate complex patterns of behavior across life that can be modified over days, months, or even years. Across these long timescales individuals within the same population may show stereotyped behaviors, but also unique behaviors that distinguish them from each other. How are long-term patterns of behavior organized and regulated across development? And what are the underlying processes that establish and modify individual-to-individual behavioral variation?
By utilizing parallel long-term behavioral monitoring at high spatiotemporal resolution of multiple C. elegans individuals across their complete development time we demonstrate temporal regulation of behavioral plasticity by neuromodulators across developmental stages, structuring shared and unique individual responses to early-life experiences. I will further describe our development of unsupervised analyses of individual biases across development based on locomotion trajectory and individual postures which uncovered a large spectrum of individuality types within the isogenic populations. Lastly, I will present preliminary results suggesting that specific neuronal pathways are required to robustly synchronize long-term behavior with development time.

Our organs and tissues are made of different cell types that communicate with each other in order to achieve joint functions. However, little is known about the universal principles of these interactions. For example, how do cell interactions maintain stable cell composition, spatial organization and collective division of labor in tissues?
And what is the role of these interactions in tissue-level diseases where the healthy balance in the tissue is disrupted such as excess scarring following injury known as fibrosis? In this talk, I will discuss my work in developing new theoretical frameworks that explore the collective behavior of cells that emerges from cell-cell interactions.
I will present work on the cell communication circuit that controls tissue repair following injury and how it may lead to fibrosis. I will discuss a new mathematical approach to explore how cell interactions can be used to provide symmetry breaking and optimal division of labor in tissues, and how this approach can help to interpret complex patterns in real high-dimensional data.
I will introduce a new concept in complex networks – network hyper-motifs, where we explore how small recurring patterns (network motifs) are integrated within large networks, and how these larger patterns (hyper-motifs) can give rise to emergent dynamic properties. Finally, I will conclude with future directions that are aimed at revealing principles that unify our understanding of different tissues.