Search

Analog reservoir computing (ARC) systems have attracted attention owing to their efficiency in processing temporal information. However, the distinct functionalities of the system components pose challenges for hardware implementation. Herein, we report a fully integrated ARC system that leverages material versatility of the ferroelectric-to-mixed phase boundary (MPB) hafnium zirconium oxides integrated onto indium-gallium-zinc oxide thin-film transistors (TFTs). MPB-based TFTs (MPBTFTs) with nonlinear short-term memory characteristics are utilized for physical reservoirs and artificial neuron, while nonvolatile ferroelectric TFTs mimic synaptic behavior for readout networks. Furthermore, double-gate configuration of MPBTFTs enhances reservoir state differentiation and state expansion for physical reservoir and processes both excitatory and inhibitory pulses for neuronal functionality with minimal hardware burden. The seamless integration of ARC components on a single wafer executes complex real-world time-series predictions with a low normalized root mean squared error of 0.28. The material-device co-optimization proposed in this study paves the way for the development of area- and energy-efficient ARC systems.

Mathematical models play an important role in biochemical engineering. For example, the models developed in the field of systems biology have been used to identify drug targets to treat pathogens such as Pseudomonas aeruginosa in biofilms. In addition, competitive binding models for chromatography processes have been developed to predict expanded bed breakthrough of a recombinant protein expressed in a high cell density fermentation. Numerical computing tools such as MATLAB can facilitate the development of these models in biochemical engineering. Compared to other computational tools and programming languages, such as C++ and Java, MATLAB is easier for students to learn since it has a large library of functions and toolboxes, including some developed in systems biology. Basic MATLAB modeling techniques have been introduced in advanced engineering courses to enhance student learning experiences. However, a nationwide survey conducted by Villanova University showed that only 28.5% of schools in United States provide training in MATLAB in graduate-level courses in biochemical engineering. Instructors in this survey also indicated they would have a strong interest in using MATLAB modules, if the modules were readily available, in courses such as systems biology. To address the need for MATLAB education in biochemical engineering, Villanova University has developed MATLAB modules and integrated them in the course CHE8663 Systems Biology (Fall semester 2012 & 2013). The teaching modules as well as the teaching strategies for MATLAB are provided here for a course that was offered to in class students as well as distance-learning (i.e., e-learning) students.

Lateral size effects of ferroelastic domain structures in epitaxial PbTiO3 thin films were investigated systematically with a viewpoint of misfit strain relaxation mechanism. The epitaxial PbTiO3 thin films were patterned into discrete islands and the effects of lateral dimension were analyzed by reciprocal space mapping using synchrotron x-ray diffraction as well as finite element simulation. As the lateral two-dimensional planar size decreases in the PbTiO3 patterns on MgO(001), some of the a domains turned into c domains due to the relaxed tensile strain. In the PbTiO3 patterns on Pt(001)/MgO(001), on the other hand, the formation of 90° domains is enhanced by the reduction in compressive misfit strain. As the pattern size decreases further to 100 nm, the untilted a domains arise due to the almost completely relaxed misfit strains. Equilibrium domain structures in the epitaxial thin films and discrete islands are also analyzed by the finite element simulation and found to be consistent with the experimental observation. These results manifest that the domain structure and evolution of the epitaxial thin films could be engineered by the control of misfit strain and its relaxation. [ABSTRACT FROM AUTHOR]

Equilibrium domain structures commonly observed in epitaxial Pb-based ferroelectric thin films are analyzed by the finite element method (FEM) using a commercial package, ABAQUS (Hibbit, Karlsson & Sorensen, Inc., 1080 Main Street, Pawtucket, RI 02860-4847). Structures of periodic 90° domains in epitaxial PbTiO[sub 3] thin films on cubic single crystalline substrates are analyzed as a function of decreasing temperature in order to simulate cooling process after the film deposition at elevated temperature (T[sub G]). The degree of c-axis orientation (α) is determined as a function of temperature below the Curie temperature and compared to the experimental results. It is then possible to calculate the magnitude of misfit strain during film growth and its relaxation due to dislocation generation. The effect of PZT composition on c-domain abundance is also simulated. FEM simulation is performed with the assumption that the major driving force for such domain formation is thermoelastic strains arising from the film-substrate interaction and the cubic-tetragonal phase transformation. The FEM analysis also suggests that initial misfit stress at T[sub G] is not fully relaxed and the unrelaxed misfit stress is inversely related to final c-domain abundance. [ABSTRACT FROM AUTHOR]

Traffic redundancy elimination (RE) is an attractive approach to improve the throughput in bandwidth-limited networks. While previous studies show that the RE is useful for improving the throughput in such networks, we observed that the RE would not be an effective solution in wireless networks. We found the TCP congestion control cannot take advantage of the RE, without knowing how the underlying RE system manipulates each TCP packet. In this letter, we present a novel technique called REboost to enable the TCP layer to be aware of the underlying RE system and improve the throughput. Our evaluation with a prototype shows that REboost significantly improves the throughput compared with the previous RE systems. [ABSTRACT FROM PUBLISHER]