Unpredictability of AI: On the Impossibility of Accurately Predicting All Actions of a Smarter Agent

2020 ◽  
Vol 07 (01) ◽  
pp. 109-118
Author(s):  
Roman V. Yampolskiy
Keyword(s):  
Intelligent System ◽  
Impossibility Result ◽  
The Impact ◽  
Ai Safety

The young field of AI Safety is still in the process of identifying its challenges and limitations. In this paper, we formally describe one such impossibility result, namely Unpredictability of AI. We prove that it is impossible to precisely and consistently predict what specific actions a smarter-than-human intelligent system will take to achieve its objectives, even if we know the terminal goals of the system. In conclusion, the impact of Unpredictability on AI Safety is discussed.

scholarly journals Knocking and Listening: Learning Mechanical Impulse Response for Understanding Surface Characteristics

Sensors ◽  
2020 ◽  
Vol 20 (2) ◽  
pp. 369 ◽  
Author(s):  
Semin Ryu ◽  
Seung-Chan Kim
Keyword(s):  
Intelligent System ◽  
Surface Characteristics ◽  
Machine Learning Techniques ◽  
Test Accuracy ◽  
Learning Approaches ◽  
Well Test ◽  
Mechanical Responses ◽  
Learning Techniques ◽  
The Impact ◽  
Mechanical Impulse

Inspired by spiders that can generate and sense vibrations to obtain information regarding a substrate, we propose an intelligent system that can recognize the type of surface being touched by knocking the surface and listening to the vibrations. Hence, we developed a system that is equipped with an electromagnetic hammer for hitting the ground and an accelerometer for measuring the mechanical responses induced by the impact. We investigate the feasibility of sensing 10 different daily surfaces through various machine-learning techniques including recent deep-learning approaches. Although some test surfaces are similar, experimental results show that our system can recognize 10 different surfaces remarkably well (test accuracy of 98.66%). In addition, our results without directly hitting the surface (internal impact) exhibited considerably high test accuracy (97.51%). Finally, we conclude this paper with the limitations and future directions of the study.

scholarly journals Intelligent Energy Management Strategy for Automated Office Buildings

Energies ◽  
2019 ◽  
Vol 12 (22) ◽  
pp. 4326 ◽  
Author(s):  
Simplice Igor Noubissie Tientcheu ◽  
Shyama P. Chowdhury ◽  
Thomas O. Olwal
Keyword(s):  
Energy Consumption ◽  
Intelligent System ◽  
Weather Condition ◽  
Electrical Energy ◽  
Integrated System ◽  
Office Buildings ◽  
Photovoltaic System ◽  
Energy Management Strategy ◽  
Significant Saving ◽  
The Impact

The increasing demand to reduce the high consumption of end-use energy in office buildings framed the objective of this work, which was to design an intelligent system management that could be utilized to minimize office buildings’ energy consumption from the national electricity grid. Heating, Ventilation and Air Conditioning (HVAC) and lighting are the two main consumers of electricity in office buildings. Advanced automation and control systems for buildings and their components have been developed by researchers to achieve low energy consumption in office buildings without considering integrating the load consumed and the Photovoltaic system (PV) input to the controller. This study investigated the use of PV to power the HVAC and lighting equipped with a suitable control strategy to improve energy saving within a building, especially in office buildings where there are reports of high misuse of electricity. The intelligent system was modelled using occupant activities, weather condition changes, load consumed and PV energy changes, as input to the control system of lighting and HVAC. The model was verified and tested using specialized simulation tools (Simulink®) and was subsequently used to investigate the impact of an integrated system on energy consumption, based on three scenarios. In addition, the direct impact on reduced energy cost was also analysed. The first scenario was tested in simulation of four offices building in a civil building in South Africa of a single occupant’s activities, weather conditions, temperature and the simulation resulted in savings of HVAC energy and lighting energy of 13% and 29%, respectively. In the second scenario, the four offices were tested in simulation due to the loads’ management plus temperature and occupancy and it resulted in a saving of 20% of HVAC energy and 29% of lighting electrical energy. The third scenario, which tested integrating PV energy (thus, the approach utilized) with the above-mentioned scenarios, resulted in, respectively, 64% and 73% of HVAC energy and lighting electrical energy saved. This saving was greater than that of the first two scenarios. The results of the system developed demonstrated that the loads’ control and the PV integration combined with the occupancy, weather and temperature control, could lead to a significant saving of energy within office buildings.

scholarly journals A New Maintenance Optimization Model Based on Three-Stage Time Delay for Series Intelligent System with Intermediate Buffer

2021 ◽  
Vol 2021 ◽  
pp. 1-19
Author(s):  
Xiaolei Lv ◽  
Qinming Liu ◽  
Zhinan Li ◽  
Yifan Dong ◽  
Tangbin Xia ◽  
...  
Keyword(s):  
Time Delay ◽  
Preventive Maintenance ◽  
Intelligent System ◽  
Buffer Stock ◽  
Normal Operation ◽  
Series System ◽  
Model Based ◽  
The Impact ◽  
Series Systems ◽  
Maintenance Model

For the maintenance problem of intelligent series system with buffer stock, a preventive maintenance model based on the threestage time delay theory is proposed. Firstly, the intelligent series system is decomposed into n − 1 virtual series systems by using approximate decomposition method. The impact factor is introduced to establish the failure rate and maintenance rate model of each virtual machine. Secondly, a preventive maintenance model based on the three-stage time delay theory is proposed for each virtual series system. The machine state from normal operation to failure stage is divided into three steps: initial defect, serious defect, and fault, and different distribution functions are defined in different stages to simulate the degradation process of the machine. Based on the three-stage time delay theory, the machine cost ratio model was established by taking the machine monitoring time and buffer stock as decision variables and the minimum unit time cost rate as objective function. Finally, the rationality and validity of the model are verified by an example analysis, which provides a basis for the maintenance of the intelligent series system.

scholarly journals Understanding the Impact of Neural Variations and Random Connections on Inference

2021 ◽  
Vol 15 ◽  
Author(s):  
Yuan Zeng ◽  
Zubayer Ibne Ferdous ◽  
Weixiang Zhang ◽  
Mufan Xu ◽  
Anlan Yu ◽  
...  
Keyword(s):  
Neural Network ◽  
Intelligent System ◽  
Design Method ◽  
Random Network ◽  
Network Connectivity ◽  
Processing Technique ◽  
Model Parameters ◽  
Perceptron Algorithm ◽  
The Impact

Recent research suggests that in vitro neural networks created from dissociated neurons may be used for computing and performing machine learning tasks. To develop a better artificial intelligent system, a hybrid bio-silicon computer is worth exploring, but its performance is still inferior to that of a silicon-based computer. One reason may be that a living neural network has many intrinsic properties, such as random network connectivity, high network sparsity, and large neural and synaptic variability. These properties may lead to new design considerations, and existing algorithms need to be adjusted for living neural network implementation. This work investigates the impact of neural variations and random connections on inference with learning algorithms. A two-layer hybrid bio-silicon platform is constructed and a five-step design method is proposed for the fast development of living neural network algorithms. Neural variations and dynamics are verified by fitting model parameters with biological experimental results. Random connections are generated under different connection probabilities to vary network sparsity. A multi-layer perceptron algorithm is tested with biological constraints on the MNIST dataset. The results show that a reasonable inference accuracy can be achieved despite the presence of neural variations and random network connections. A new adaptive pre-processing technique is proposed to ensure good learning accuracy with different living neural network sparsity.

scholarly journals ПРИНЦИП ЦІЛІСНОЇ ОРГАНІЗАЦІЇ ІНТЕЛЕКТУАЛЬНИХ СИСТЕМ

2019 ◽  
pp. 4-16
Author(s):  
Серій Ілліч Доценко
Keyword(s):  
Control System ◽  
Intelligent Systems ◽  
Systems Approach ◽  
Intelligent System ◽  
Control Object ◽  
The Other ◽  
Methodological Basis ◽  
Other Hand ◽  
Intellectual System ◽  
The Impact

The antinomy of the division of the intellectual system into parts has been formed, namely: the intellectual system is an organized whole, which is formed from at least two parts; for an intelligent system, as an organized whole, it is impossible to divide into a controlling part (control system) and a part of which is controlled. It has been established that the antinomy of dividing an intelligent system into parts is generated by the fact that, traditionally, the control system and the control object are considered separately. Therefore, it is considered the system, and not an organized whole. The role of the theory of functional systems in the development of cybernetic systems as intellectual systems is defined. This theory is the basis for the development of intelligent systems A. V. Chechkinim, K. A. Pupkov, and other authors. On the other hand, M. I. Meltzer develops the theory of dialogue systems for managing production enterprises, the basis of which is the mathematical theory of systems. It is shown that the functional representation architectures for these systems are similar. The similarity is determined on the basis of the task approach. On the one hand, there is a mutual non-recognition of the results of scientific schools of physical and technical cybernetics, and on the other hand, there is a similarity of the results obtained. It has been established that the methodological basis of the holistic approach is the task approach to the formation of a solving system, developed in the theory of dialogue management of production. To do this, it is necessary to include the “Activity to get the result” block in the solving system in order to turn it into an intellectual system. The methodological basis of a systems approach is a functional approach to the formation of systems. The main lesson of the classical cybernetics crisis, regarding the organizational principle for two parts of an organized whole, is to establish a dialectical unity of concepts in the form of a “general” concept and a “concrete” concept for problem-solving results in the control system and control object. Thus, a dialectically organized whole is formed. The article also analyzes the impact of the study of intelligent systems on the development of the methodological foundations of the Industry 4.0 platform. The next task that needs to be solved is the formation of the principle of functional self-organization, which is the basis for the formation of a mechanism for ensuring consistency between the results of solving problems in parts of a dialectically organized whole

Intelligent integrated system of football teaching and training

2021 ◽  
pp. 1-9
Author(s):  
Jin Wang ◽  
Jianxiong Wang ◽  
Sijie Tan
Keyword(s):  
Task Allocation ◽  
Intelligent System ◽  
Fuzzy Theory ◽  
Integrated System ◽  
Assessment System ◽  
Control Mode ◽  
Robot Soccer ◽  
Technical Problems ◽  
The Impact ◽  
And Training

Football is a popular sport all over the world, and it is also an important part to show the comprehensive strength of the country. With the development of football in China, how to effectively carry out football teaching and training has become a hot topic. At present, the international mainstream method is to use robot soccer simulation training to simulate the real game scene, analyze and study the game tactics. However, due to the late start of artificial intelligence in China, there are still technical problems in the field of robot soccer, such as insufficient recognition and unreasonable task allocation. In order to solve these problems, this paper proposes an intelligent soccer teaching and training system based on fuzzy theory. In this paper, the control mode of the intelligent system is further optimized by combining the classical algorithm of fuzzy mathematics with PID control. In the design of football training model, this paper puts forward an innovative game situation assessment system, which can better analyze the impact of environmental factors. In the aspect of control circuit, this paper adopts the mainstream LM2678 step-down circuit, which has good stability and has been widely used in intelligent control system. In the final experimental analysis, this paper uses the current representative basic decision-making system as the comparative object, through a number of experiments including technical indicators, system characteristics, etc. Analysis of the data shows that the intelligent system based on fuzzy theory has better task allocation ability than the traditional way, and obviously improves the comprehensive performance of the system. In the simulation competition, the system in this paper has made outstanding achievements, which further shows the superiority of the system.

scholarly journals Research on the Construction of Cold Chain Logistics Intelligent System Based on 5G Ubiquitous Internet of Things

2021 ◽  
Vol 2021 ◽  
pp. 1-11
Author(s):  
Qi Hao Han
Keyword(s):  
Internet Of Things ◽  
Intelligent System ◽  
Hierarchical Control ◽  
Scientific Basis ◽  
Cold Chain ◽  
Agricultural Product ◽  
The Internet ◽  
Gps System ◽  
The Impact ◽  
The Internet Of Things

In the research on the construction of cold chain logistics intelligent system based on 5G ubiquitous Internet of Things as an emerging technology, the Internet of Things is penetrating into the logistics field with its technical advantages and changing the original industry form. According to the development status of cold chain logistics, this paper constructs a set of cold chain logistics intelligent system by using RFID technology, sensing equipment, GPS system, fuzzy PID hierarchical control, and other Internet of Things technologies and analyzes the impact of the cost sharing ratio of Internet of Things on the income of wholesalers, retailers, and supply chain. The example analysis shows that when retailers share all the adoption costs of the Internet of Things, the benefit of the whole fresh agricultural product cold chain logistics is the largest. This study provides a scientific basis for logistics decision-making for wholesalers and retailers and has important application value for improving the logistics efficiency of the whole fresh agricultural product cold chain.

scholarly journals Artificial intelligent system for multimedia services in smart home environments

2021 ◽  
Author(s):  
Albert Rego ◽  
Pedro Luis González Ramírez ◽  
Jose M. Jimenez ◽  
Jaime Lloret
Keyword(s):  
Deep Learning ◽  
Smart Home ◽  
Intelligent System ◽  
Multimedia Services ◽  
Home Environments ◽  
Artificial Intelligent ◽  
New Applications ◽  
The Impact ◽  
Deep Learning Model ◽  
Smart Home Environments

AbstractInternet of Things (IoT) has introduced new applications and environments. Smart Home provides new ways of communication and service consumption. In addition, Artificial Intelligence (AI) and deep learning have improved different services and tasks by automatizing them. In this field, reinforcement learning (RL) provides an unsupervised way to learn from the environment. In this paper, a new intelligent system based on RL and deep learning is proposed for Smart Home environments to guarantee good levels of QoE, focused on multimedia services. This system is aimed to reduce the impact on user experience when the classifying system achieves a low accuracy. The experiments performed show that the deep learning model proposed achieves better accuracy than the KNN algorithm and that the RL system increases the QoE of the user up to 3.8 on a scale of 10.

scholarly journals Detection of damaged seeds in laboratory evaluation of precision planter using impact acoustics and artificial neural networks

2012 ◽  
Vol 1 (2) ◽  
pp. 67 ◽  
Author(s):  
Hadi Karimi ◽  
Hossein Navid ◽  
Asghar Mahmoudi
Keyword(s):  
Neural Network ◽  
Intelligent System ◽  
Acoustic Signals ◽  
Laboratory Evaluation ◽  
Ann Model ◽  
Feed Forward Neural Network ◽  
Ann Models ◽  
Artificial Neural ◽  
Seed Metering Device ◽  
The Impact

In the present study, feasibility of laboratory detection of damaged seeds in precision planters caused by malfunction of seed metering device was investigated. An acoustic-based intelligent system was developed for detection of damaged pelleted tomato seeds. To improve the Artificial Neural Network (ANN) models a total of 2000 seeds sound signals, 1000 samples for damaged seeds and 1000 for undamaged ones were recorded. When seed metering device drove out seeds, the ejected seeds were impacted to steel plate, and their acoustic signals were recorded from the impact. The bounced seeds lied on the running grease belt. In each stage of experiments, damaged seeds were determined manually in grease belt and related damaged seed sound signals were designated. Achieved acoustic signals, were processed and potential features were extracted from the analysis of sound signals in time and frequency domains. The method is based on feature generation by Fast Fourier Transform (FFT), feature selection by statistical methods and classification by Multilayer Feed forward Neural Network. Features such as amplitude, phase and power spectrum of sound signals were computed through a 1024-point FFT. By using statistical factors (maximum, minimum, median, mean and variance) for each vector of data, feature vector was reduced to 15 factors. In developing the ANN models, several ANN architectures, each having different numbers of neurons in hidden layer, were evaluated. The best model was chosen after a number of evaluations based on minimizing the mean square error (MSE), correct detection rate (CDR) and correlation coefficient (r). Selected ANN, 15-17-2 was configured for classification. CDR of the proposed ANN model for undamaged and damaged seeds was 99.49 and 100 respectively. MSE of the system was found to be 0.0109.

Sign in / Sign up

Export Citation Format

Share Document