Parallel multi-objective optimization approaches for protein encoding

One of the main challenges in synthetic biology lies in maximizing the expression levels of a protein by encoding it with multiple copies of the same gene. This task is often conducted under conflicting evaluation criteria, which motivates the formulation of protein encoding as a multi-objective optimization problem. Recent research reported significant results when adapting the artificial bee colony algorithm to address this problem. However, the length of proteins and the number of copies have a noticeable impact in the computational costs required to attain satisfying solutions. This work is aimed at proposing parallel bioinspired designs to tackle protein encoding in multiprocessor systems, considering different thread orchestration schemes to accelerate the optimization process while preserving the quality of results. Comparisons of solution quality with other approaches under three multi-objective quality metrics show that the proposed parallel method reaches significant quality in the encoded proteins. In addition, experimentation on six real-world proteins gives account of the benefits of applying asynchronous shared-memory schemes, attaining efficiencies of 92.11% in the most difficult stages of the algorithm and mean speedups of 33.28x on a 64-core server-grade system.

Low-Memory Indoor Positioning System for Standalone Embedded Hardware

As the proportion and importance of the indoor spaces in daily life are gradually increasing, spatial information and personal location information become more important in indoor spaces. In order to apply indoor positioning technologies in any places and for any targets inexpensively and easily, the system should utilize simple sensors and devices. In addition, due to the scalability, it is necessary to perform indoor positioning algorithms on the device itself, not on the server. In this paper, we construct standalone embedded hardware for performing the indoor positioning algorithm. We use the geomagnetic field for indoor localization, which does not require the installation of infrastructure and has more stable signal strength than RF RSS. In addition, we propose low-memory schemes based on the characteristics of the geomagnetic sensor measurement and convergence of the target’s estimated positions in order to implement indoor positioning algorithm to the hardware. We evaluate the performance in two testbeds: Hana Square (about 94 m × 26 m) and SK Future Hall (about 60 m × 38 m) indoor testbeds. We can reduce flash memory usage to 16.3% and 6.58% for each testbed and SRAM usage to 8.78% and 23.53% for each testbed with comparable localization accuracy to the system based on smart devices without low-memory schemes.

Building Attention and Edge Convolution Neural Networks for Bioactivity and Physical-Chemical Property Prediction

We introduce Attention and Edge Memory schemes to the existing Message Passing Neural Network framework for graph convolution, and benchmark our approaches against eight different physical-chemical and bioactivity datasets from the literature. We remove the need to introduce <i>a priori</i> knowledge of the task and chemical descriptor calculation by using only fundamental graph-derived properties. Our results consistently perform on-par with other state-of-the-art machine learning approaches, and set a new standard on sparse multi-task virtual screening targets. We also investigate model performance as a function of dataset preprocessing, and make some suggestions regarding hyperparameter selection.

Building Attention and Edge Convolution Neural Networks for Bioactivity and Physical-Chemical Property Prediction

We introduce Attention and Edge Memory schemes to the existing Message Passing Neural Network framework for graph convolution, and benchmark our approaches against eight different physical-chemical and bioactivity datasets from the literature. We remove the need to introduce <i>a priori</i> knowledge of the task and chemical descriptor calculation by using only fundamental graph-derived properties. Our results consistently perform on-par with other state-of-the-art machine learning approaches, and set a new standard on sparse multi-task virtual screening targets. We also investigate model performance as a function of dataset preprocessing, and make some suggestions regarding hyperparameter selection.

Building Attention and Edge Convolution Neural Networks for Bioactivity and Physical-Chemical Property Prediction

We introduce Attention and Edge Memory schemes to the existing Message Passing Neural Network framework for graph convolution, and benchmark our approaches against eight different physical-chemical and bioactivity datasets from the literature. We remove the need to introduce <i>a priori</i> knowledge of the task and chemical descriptor calculation by using only fundamental graph-derived properties. Our results consistently perform on-par with other state-of-the-art machine learning approaches, and set a new standard on sparse multi-task virtual screening targets. We also investigate model performance as a function of dataset preprocessing, and make some suggestions regarding hyperparameter selection.

Analysis and Simulations of Hybrid Memory Scheme Based on Memristors

The investigation of new memory schemes is significant for future generations of electronic devices. The purpose of this research is to present a detailed analysis of the processes in the memory elements of a memory section with memristors and isolating Metal Oxide Semiconductor (MOS) transistors. For the present analysis, a modified window function previously proposed by the author in another memristor model is used. The applied model is based on physical nonlinear current-voltage and state-voltage characteristics. It is suitable for illustration of the processes in the memristors for both writing and reading procedures. The memory scheme is simulated using a nonlinear drift model with an improved window function. The used model was previously adjusted according to the reference Pickett model. The memory circuit is analyzed for writing and reading information procedures. The memristor current-voltage relationship is compared to physical experimental characteristics and to results acquired by the use of basic window functions. A satisfactory coincidence between the corresponding results is established. For the used logical signals, the memory elements operate in a state near to hard-switching mode. It is confirmed that the memristor model with a modified window function applied here is suitable for investigating complex memristor circuits for a general operating mode.

Two Bi-Accelerator Improved with Memory Schemes for Solving Nonlinear Equations

The present paper is devoted to the improvement of theR-order convergence of with memory derivative free methods presented by Lotfi et al. (2014) without doing any new evaluation. To achieve this aim one more self-accelerating parameter is inserted, which is calculated with the help of Newton’s interpolatory polynomial. First theoretically it is proved that theR-order of convergence of the proposed schemes is increased from 6 to 7 and 12 to 14, respectively, without adding any extra evaluation. Smooth as well as nonsmooth examples are discussed to confirm theoretical result and superiority of the proposed schemes.