scholarly journals Molecular convolutional neural networks with DNA regulatory circuits

2021 ◽  
Author(s):  
Hao Pei ◽  
Xiewei Xiong ◽  
Tong Zhu ◽  
Yun Zhu ◽  
Mengyao Cao ◽  
...  
Keyword(s):  
Neural Network ◽  
Computation Time ◽  
Great Promise ◽  
Hierarchical Networks ◽  
Computing Power ◽  
Synthetic Dna ◽  
Noisy Information ◽  
Regulatory Circuits ◽  
Biochemical Systems ◽  
Intelligent Behavior

Abstract Complex biomolecular circuits enable cells with intelligent behavior for survival before neural brains evolved. Synthesized DNA circuits in liquid phase developed as computational hardware can perform neural-network-like computation that harness the collective properties of complex biochemical systems, however the scaling up in complexity remains challenging to support more powerful computation. we present a systematic molecular implementation of the convolutional neural network (ConvNet) algorithm with synthetic DNA regulatory circuits based on a simple DNA switching gate architecture. We experimentally demonstrated that a DNA-based ConvNet based on shared-weight architecture of a 3×6 sized kernel can simultaneously implement parallel multiply-accumulate (MAC) operations for 144 bits inputs and recognize patterns up to 8 categories autonomously. Furthermore, it can connect with another DNA circuits to construct hierarchical networks, which can recognize patterns up to 32 categories with a two-step classification approach of performing coarse classification on language (Arabic numerals, Chinese oracles, English alphabets and Greek alphabets) and then classifying them into specific handwritten symbols. With a simple cyclic freeze/thaw approach, we can decrease computation time from hours to minutes. Our approach shows great promise in the realization of high computing power molecular computer with ability to classify complex and noisy information.

Developing an Artificial Intelligence Project in your Radiology Department

2020 ◽  
Vol 2 ◽  
pp. 58-61 ◽  
Author(s):  
Syed Junaid ◽  
Asad Saeed ◽  
Zeili Yang ◽  
Thomas Micic ◽  
Rajesh Botchu
Keyword(s):  
Neural Network ◽  
Artificial Intelligence ◽  
Health Care ◽  
Deep Learning ◽  
Learning Algorithms ◽  
Radiology Department ◽  
Short Article ◽  
Computing Power ◽  
Road Map ◽  
Key Concepts

The advances in deep learning algorithms, exponential computing power, and availability of digital patient data like never before have led to the wave of interest and investment in artificial intelligence in health care. No radiology conference is complete without a substantial dedication to AI. Many radiology departments are keen to get involved but are unsure of where and how to begin. This short article provides a simple road map to aid departments to get involved with the technology, demystify key concepts, and pique an interest in the field. We have broken down the journey into seven steps; problem, team, data, kit, neural network, validation, and governance.

scholarly journals Embedded Deep Learning Prototyping Approach for Cyber-Physical Systems: Smart LIDAR Case Study

2021 ◽  
Vol 10 (1) ◽  
pp. 18
Author(s):  
Quentin Cabanes ◽  
Benaoumeur Senouci ◽  
Amar Ramdane-Cherif
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Computation Time ◽  
Physical World ◽  
Cyber Physical Systems ◽  
Physical Systems ◽  
Research Technology ◽  
Standard Design ◽  
Hardware Architectures ◽  
Voxel Grid

Cyber-Physical Systems (CPSs) are a mature research technology topic that deals with Artificial Intelligence (AI) and Embedded Systems (ES). They interact with the physical world via sensors/actuators to solve problems in several applications (robotics, transportation, health, etc.). These CPSs deal with data analysis, which need powerful algorithms combined with robust hardware architectures. On one hand, Deep Learning (DL) is proposed as the main solution algorithm. On the other hand, the standard design and prototyping methodologies for ES are not adapted to modern DL-based CPS. In this paper, we investigate AI design for CPS around embedded DL. The main contribution of this work is threefold: (1) We define an embedded DL methodology based on a Multi-CPU/FPGA platform. (2) We propose a new hardware design architecture of a Neural Network Processor (NNP) for DL algorithms. The computation time of a feed forward sequence is estimated to 23 ns for each parameter. (3) We validate the proposed methodology and the DL-based NNP using a smart LIDAR application use-case. The input of our NNP is a voxel grid hardware computed from 3D point cloud. Finally, the results show that our NNP is able to process Dense Neural Network (DNN) architecture without bias.

scholarly journals Fpga Implementation of Precise Convolutional Neural Network for Extreme Learning Machine

2020 ◽  
Vol 9 (8) ◽  
pp. 470-480
Keyword(s):  
Neural Network ◽  
Receptive Field ◽  
Convolutional Neural Network ◽  
Computation Time ◽  
Field Approach ◽  
Pixel Array ◽  
Gradient Based ◽  
Feed Forward Neural Networks ◽  
Learning Machine ◽  
Hidden Layer

Feed-forward neural networks can be trained based on a gradient-descent based backpropagation algorithm. But, these algorithms require more computation time. Extreme Learning Machines (ELM’s) are time-efficient, and they are less complicated than the conventional gradient-based algorithm. In previous years, an SRAM based convolutional neural network using a receptive – field Approach was proposed. This neural network was used as an encoder for the ELM algorithm and was implemented on FPGA. But, this neural network used an inaccurate 3-stage pipelined parallel adder. Hence, this neural network generates imprecise stimuli to the hidden layer neurons. This paper presents an implementation of precise convolutional neural network for encoding in the ELM algorithm based on the receptive - field approach at the hardware level. In the third stage of the pipelined parallel adder, instead of approximating the output by using one 2-input 15-bit adder, one 4-input 14-bit adder is used. Also, an additional weighted pixel array block is used. This weighted pixel array improves the accuracy of generating 128 weighted pixels. This neural network was simulated using ModelSim-Altera 10.1d and synthesized using Quartus II 13.0 sp1. This neural network is implemented on Cyclone V FPGA and used for pattern recognition applications. Although this design consumes slightly more hardware resources, this design is more accurate compared to previously existing encoders

scholarly journals Comparison of the efficiency of neural network algorithms in recognition and classification problems

2020 ◽  
Vol 224 ◽  
pp. 01025
Author(s):  
Alexey Beskopylny ◽  
Alexandr Lyapin ◽  
Nikita Beskopylny ◽  
Elena Kadomtseva
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Computation Time ◽  
Classification Problems ◽  
Handwritten Digit Recognition ◽  
Network Algorithms ◽  
Digit Recognition ◽  
Labour Intensity ◽  
Handwritten Digit ◽  
Accuracy Of Results

The article is devoted to the problem of comparing the effectiveness of feedforward (FF) and convolutional neural networks (CNN) algorithms in the problems of handwritten digit recognition and classification. In recent years, the attention of many researchers to the FF and CNN algorithms has given rise to many hybrid models focused on solving specific problems. At the same time, the efficiency of each algorithm in terms of accuracy and labour intensity remains unclear. It is shown that in classical problems, FFs can have advantages over CNN in terms of labour intensity with the same accuracy of results. Using the handwritten digits data from the MNIST database as an example, it is shown that FF algorithms provide greater accuracy and require less computation time than CNN.

scholarly journals Hybrid Imperialistic Competitive Algorithm Incorporated with Hopfield Neural Network for Robust 3 Satisfiability Logic Programming

2019 ◽  
Vol 8 (2) ◽  
pp. 144 ◽  
Author(s):  
Vigneshwer Kathirvel ◽  
Mohd. Asyraf Mansor ◽  
Mohd Shareduwan Mohd Kasihmuddin ◽  
Saratha Sathasivam
Keyword(s):  
Neural Network ◽  
Logic Programming ◽  
Computation Time ◽  
Imperialist Competitive Algorithm ◽  
Hopfield Neural Network ◽  
Absolute Error ◽  
Training Algorithm ◽  
Competitive Algorithm ◽  
Bayesian Criterion ◽  
Expected Outcome

Imperialist Competitive algorithm (ICA) is a robust training algorithm inspired by the socio-politically motivated strategy. This paper focuses on utilizing a hybridized ICA with Hopfield Neural Network on a 3-Satisfiability (3-SAT) logic programming. Eventually the performance of the proposed algorithm will be compared to other 2 algorithms, which are HNN-3SATES (ES) and HNN-3SATGA (GA). The performance shall be evaluated with the Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Sum of Squares Error (SSE), Schwarz Bayesian Criterion (SBC), Global Minima Ratio and Computation Time (CPU time). The expected outcome will portray that the IC algorithm will outperform the other two algorithms in doing 3-SAT logic programming.

Combined Trajectory Planning and Tracking for Autonomous Vehicles on Deformable Terrains

2020 ◽  
Author(s):  
James Dallas ◽  
Yifan Weng ◽  
Tulga Ersal
Keyword(s):  
Neural Network ◽  
Trajectory Planning ◽  
Autonomous Vehicles ◽  
Performance Metrics ◽  
Soft Soil ◽  
Computation Time ◽  
Operating Conditions ◽  
The Neural Network ◽  
Vehicle Response ◽  
The Impact

Abstract In this work, a novel combined trajectory planner and tracking controller is developed for autonomous vehicles operating on off-road deformable terrains. Common approaches to trajectory planning and tracking often rely on model-dependent schemes, which utilize a simplified model to predict the impact of control inputs to future vehicle response. However, in an off-road context and especially on deformable terrains, accurately modeling the vehicle response for predictive purposes can be challenging due to the complexity of the tire-terrain interaction and limitations of state-of-the-art terramechanics models in terms of operating conditions, computation time, and continuous differentiability. To address this challenge and improve vehicle safety and performance through more accurate prediction of the plant response, in this paper, a nonlinear model predictive control framework is presented that accounts for terrain deformability explicitly using a neural network terramechanics model for deformable terrains. The utility of the proposed scheme is demonstrated on high fidelity simulations for a notional lightweight military vehicle on soft soil. It is shown that the neural network based controller can outperform a baseline Pacejka model based scheme by improving on performance metrics associated with the cost function. In more severe maneuvers, the neural network based controller can achieve sufficient fidelity as compared to the plant to complete maneuvers that lead to failure for the Pacejka based controller. Finally, it is demonstrated that the proposed framework is conducive to real-time implementability.

Neural Network-Based Metaheuristic Parameterization with Application to the Vehicle Matching Problem in Ride-Hailing Services

2019 ◽  
Vol 2673 (10) ◽  
pp. 311-320
Author(s):  
Arslan Ali Syed ◽  
Irina Gaponova ◽  
Klaus Bogenberger
Keyword(s):  
Neural Network ◽  
Optimization Problems ◽  
Optimal Solution ◽  
Computation Time ◽  
Problem Instance ◽  
Neighborhood Search ◽  
Matching Problem ◽  
Solution Quality ◽  
Actual Problem ◽  
Parameter Values

The majority of transportation problems include optimizing some sort of cost function. These optimization problems are often NP-hard and have an exponential increase in computation time with the increase in the model size. The problem of matching vehicles to passenger requests in ride hailing (RH) contexts typically falls into this category.Metaheuristics are often utilized for such problems with the aim of finding a global optimal solution. However, such algorithms usually include lots of parameters that need to be tuned to obtain a good performance. Typically multiple simulations are run on diverse small size problems and the parameters values that perform the best on average are chosen for subsequent larger simulations.In contrast to the above approach, we propose training a neural network to predict the parameter values that work the best for an instance of the given problem. We show that various features, based on the problem instance and shareability graph statistics, can be used to predict the solution quality of a matching problem in RH services. Consequently, the values corresponding to the best predicted solution can be selected for the actual problem. We study the effectiveness of above described approach for the static assignment of vehicles to passengers in RH services. We utilized the DriveNow data from Bavarian Motor Works (BMW) for generating passenger requests inside Munich, and for the metaheuristic, we used a large neighborhood search (LNS) algorithm combined with a shareability graph.

scholarly journals A Biologically Inspired Framework for the Intelligent Control of Mechatronic Systems and Its Application to a Micro Diving Agent

2018 ◽  
Vol 2018 ◽  
pp. 1-16 ◽  
Author(s):  
Wallace M. Bessa ◽  
Gerrit Brinkmann ◽  
Daniel A. Duecker ◽  
Edwin Kreuzer ◽  
Eugen Solowjow
Keyword(s):  
Neural Network ◽  
Sliding Mode ◽  
Weight Vector ◽  
Mechatronic Systems ◽  
Biologically Inspired ◽  
Control Approach ◽  
The Neural Network ◽  
Control Scheme ◽  
Intelligent Behavior ◽  
Intelligent Controllers

Mechatronic systems are becoming an intrinsic part of our daily life, and the adopted control approach in turn plays an essential role in the emulation of the intelligent behavior. In this paper, a framework for the development of intelligent controllers is proposed. We highlight that robustness, prediction, adaptation, and learning, which may be considered the most fundamental traits of all intelligent biological systems, should be taken into account within the project of the control scheme. Hence, the proposed framework is based on the fusion of a nonlinear control scheme with computational intelligence and also allows mechatronic systems to be able to make reasonable predictions about its dynamic behavior, adapt itself to changes in the plant, learn by interacting with the environment, and be robust to both structured and unstructured uncertainties. In order to illustrate the implementation of the control law within the proposed framework, a new intelligent depth controller is designed for a microdiving agent. On this basis, sliding mode control is combined with an adaptive neural network to provide the basic intelligent features. Online learning by minimizing a composite error signal, instead of supervised off-line training, is adopted to update the weight vector of the neural network. The boundedness and convergence properties of all closed-loop signals are proved using a Lyapunov-like stability analysis. Numerical simulations and experimental results obtained with the microdiving agent demonstrate the efficacy of the proposed approach and its suitableness for both stabilization and trajectory tracking problems.

scholarly journals Neural Network Identification and Control of a Parametrically Excited Structural Dynamic Model of an F-15 Tail Section

2000 ◽  
Vol 7 (6) ◽  
pp. 355-361 ◽  
Author(s):  
Ayman A. El-Badawy ◽  
Ali H. Nayfeh ◽  
Hugh Van Landingham
Keyword(s):  
Neural Network ◽  
Computer Simulation ◽  
Dynamic Model ◽  
Parametric Excitation ◽  
Vibration Suppression ◽  
Great Promise ◽  
Structural Dynamic ◽  
Neural Network Controller ◽  
Structural Dynamic Model ◽  
Computer Simulation Studies

We investigated the design of a neural-network-based adaptive control system for a smart structural dynamic model of the twin tails of an F-15 tail section. A neural network controller was developed and tested in computer simulation for active vibration suppression of the model subjected to parametric excitation. First, an emulator neural network was trained to represent the structure to be controlled and thus used in predicting the future responses of the model. Second, a neurocontroller to determine the necessary control action on the structure was developed. The control was implemented through the application of a smart material actuator. A strain gauge sensor was assumed to be on each tail. Results from computer-simulation studies have shown great promise for control of the vibration of the twin tails under parametric excitation using artificial neural networks.

scholarly journals Design Space Exploration for YOLO Neural Network Accelerator

Electronics ◽  
2020 ◽  
Vol 9 (11) ◽  
pp. 1921
Author(s):  
Hongmin Huang ◽  
Zihao Liu ◽  
Taosheng Chen ◽  
Xianghong Hu ◽  
Qiming Zhang ◽  
...  
Keyword(s):  
Neural Network ◽  
Design Space Exploration ◽  
Design Space ◽  
Space Exploration ◽  
Computation Time ◽  
Memory Bandwidth ◽  
Block Transmission ◽  
Transmission Strategy ◽  
The Neural Network ◽  
The Given

The You Only Look Once (YOLO) neural network has great advantages and extensive applications in computer vision. The convolutional layers are the most important part of the neural network and take up most of the computation time. Improving the efficiency of the convolution operations can greatly increase the speed of the neural network. Field programmable gate arrays (FPGAs) have been widely used in accelerators for convolutional neural networks (CNNs) thanks to their configurability and parallel computing. This paper proposes a design space exploration for the YOLO neural network based on FPGA. A data block transmission strategy is proposed and a multiply and accumulate (MAC) design, which consists of two 14 × 14 processing element (PE) matrices, is designed. The PE matrices are configurable for different CNNs according to the given required functions. In order to take full advantage of the limited logical resources and the memory bandwidth on the given FPGA device and to simultaneously achieve the best performance, an improved roofline model is used to evaluate the hardware design to balance the computing throughput and the memory bandwidth requirement. The accelerator achieves 41.99 giga operations per second (GOPS) and consumes 7.50 W running at the frequency of 100 MHz on the Xilinx ZC706 board.

Sign in / Sign up

Export Citation Format

Share Document