Find research datasets worth reusing

opencc-zeroJun 2024View details →

Dataset: Global X Artificial Intelligence & Technology ETF (AIQ) Stock Performance

This dataset provides historical stock market performance data for specific companies. It enables users to analyze and understand the past trends and fluctuations in stock prices over time. This information can be utilized for various purposes such as investment analysis, financial research, and market trend forecasting.

opencc-zeroJun 2024View details →

Dataset: Themes Generative Artificial Intelligence ETF (WISE) Stock Performance

opencc-zeroJun 2024View details →

Dataset: First Trust Nasdaq Artificial Intelligence and Robotics ETF (ROBT) Stock Performance

opencc-by-4.0Jul 2024View details →

Data from an empirical study on generative artificial intelligence painting

This is an empirical study on generative artificial intelligence painting, including research methods, basic data of 17 subjects and their AI art painting, and quantitative data after their data analysis.

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 20. Overview about which Binding Mechanisms Work at what Hierarchical Levels and Development Stages of the Brain

As a result of our research, in [60], a solution to the binding problem for perception was suggested by combining the already existing binding hypotheses in a conclusive way, supplementing them with other insights about the perceptual system of the brain, and translating them into a technically implementable model. It was demonstrated via computational simulations that different binding mechanisms proposed in literature are not mutually inclusive. On the contrary! At different hierarchical levels and in different development stages, different binding mechanisms are acting in perception. An overview about these circumstances is given in Figure 20. A detailed description can be found in.

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 19. AnyLogic Implementation of Decision-Making Architechture in an Virtual Autonomous Agent Environment

Figure 19 shows a screenshot of the test implementation. The picture in the middle shows the modules and interfaces of the decision-making architecture which were realized by so-called “active objects” and “ports”. On the left side, the implemented modules are listed. In the right lower corner of the figure, the agents and the virtual environment are displayed. The environment comprises different “objects” (food sources, obstacles, predators, other agents, etc.). In order to survive, the agents have to access food sources. However, the accessing of food sources bears difficulties and risks.

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 16. Architecture for Affective Situation Assessment of Perceptual Images (Internal Connections between Emotions are not Depicted for Better Clarity of the Graphic)

Based on the concept of affective neuro-symbols, a model was developed according to which emotions can be represented by affective neuro-symbolic networks (see right half of Figure 16, referred to as architecture of “internal perception” in contrast to the “external perception” architecture of the left half of Figure 16, which has already been presented in Section 4.2). The individual affective neuro-symbols (depicted as circles) represent different emotions (fear, anger, guilt, joy, rage, panic, love, happiness, etc.).

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 13. Implementation of Neuro-Symbols and their Communication in AnyLogic

Figure 13 and Figure 14 show screenshots of the model implementation in AnyLogic. Figure 13a shows how individual neuro-symbols were implemented. Neuro-symbols are realized by socalled active objects with an input port and an output port via which information is exchanged with other elements. Additionally, variables are used for calculating the activation of the neuro-symbols (not depicted) and for storing properties of neuro-symbols (e.g., the location property). Timers and state charts serve for processing information that arrives in a certain time window or in a certain temporal succession at the input port. Whenever new input information arrives at the input port, the activation degree of the neuro-symbol is recalculated and checked against the threshold value. Based on this, the neuro-symbol is either activated or deactivated and the corresponding information is sent via the output port by using “message objects” (see Figure 13b).

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 17. Autonomous Decision-Making Architecture

An overview of the decision-making architecture is presented in Figure 17. The architecture was guided by two core concepts. The first core concept is that human intelligence bases on a combination of low-level and high-level mechanisms. Low-level mechanisms are mainly predefined. They are not in all situations completely accurate but have the advantage of being fast. High-level mechanisms are not predefined and thus slower but more accurate. The second core concept concerns the use of so-called emotions as mechanism for the evaluation of information

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 15. Input Sources of an Affective Neuro-Symbol Representing an Emotion

Based on the descriptions given above and the concept of neuro-symbolic information processing outlined in Section 4.2, so-called “affective neuro-symbols” were defined for the affective situation assessment architecture (see Figure 15). These affective neuro-symbols can principally receive information from four different sources: (1) body states, (2) objects and events perceived in the environment (external perception), (3) from other emotions and (4) cognitive (reasoning) processes. An input from one of these sources can in certain circumstances already be sufficient to activate an affective neuro-symbol. Different sources can either have an exhibitory or inhibitory effect on the activation of an affective neuro-symbol.

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 14. Implementation of Overall Architecture of the Perception Model in AnyLogic 4.3

To perform complex functions, individual neuro-symbols are then connected to networks. Figure 14 shows a screenshot of the AnyLogic implementation of the overall system at the beginning of the learning phase. The lowest neuro-symbolic levels receive the direct sensor information as input. The higher neuro-symbolic levels are originally not interconnected amongst each other. Instead, they are connected to so-called “learning ports”, which additionally receive control information needed for the supervised learning process. Details about the multi-stage multilevel learning process can be found in.

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 12. Affinities and Differences of Neuro-Symbolic Networks in Comparison to Classical Neural Networks

After having briefly illustrated the basic function principle of neuro-symbolic networks, this section aims at reviewing their affinities and differences to standard neural networks like for example multi-layer perceptrons (MLPs) [58]. A summary of these affinities and differences is given in Figure 12. The affinities concern certain functions of individual nodes of the networks. In both cases, weighted input information is summed up and an activation function is applied to this sum. In both cases, the individual nodes are interconnected to form networks. Much larger than the number of affinities between neuro-symbolic networks and neural network is however the number of differences. The first difference consists in the application domain. Neuro-symbolic networks have so far mainly been applied for complex, large-scale sensor data processing of multimodal data – an application which can so far barely be handled by neural networks.

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 11. Activated Neuro-Symbols for Detecting that a Person walks around in the Room

With the example of Figure 11, also the function of feedback connections can be explained. According to the existing feedforward connections, the neuro-symbol “object stands” would be activated together with the neuro-symbol “object moves” whenever the neuro-symbols “motion” and “object moves” are active, because it is activated by a subset of the neuro-symbols that activate the neuro-symbol “object moves”. This activation would however be undesired in this concrete case. For this reason, an inhibitory feedback connection exists from the neuro-symbol “object moves” to the neuro-symbol “object stands” that inhibits the activation of the neuro-symbol “object stands”.

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 10. Meeting Room Equipped with Different Sensors

To illustrate the basic working principle of a perceptual neuro-symbolic network in a concrete application, a simplified, concrete example is given in the following. In this example, office meeting room is equipped with different sensors (tactile floor sensors, motion detectors, light barriers, a door contact sensor, a camera, and a microphone) as sketched in Figure 10.

BRAIN Journal-Brain-Like Artificial Intelligence for Automation-Figure 18. Different Modules involved in the Autonomous Decision-Making Process

The basic functioning of this architecture is now described in the following step by step using Figure 18a-f, where always the relevant modules of the model are highlighted for better comprehension.

opencc-by-4.0Dec 2014View details →

BRAIN Journal-On the Idea of a New Artificial Intelligence Based Optimization Algorithm Inspired From the Nature of Vortex-Figure 1. Working mechanism of the VOA.

As it can be seen from the algorithm steps, the VOA employs simple equations. It is an advantage that the algorithm can be formed and applied within optimization problems whereas alternative algorithms may contain some complex solution steps (This situation may be also an disadvantage for the VOA when it is applied in more difficult optimization problems but while the world is transformed into a ‘strong simplicity’, the VOA may be a practical solution approach). The working mechanism of the VOA can be visualized briefly as like in Figure 1.

Artificial Intelligence and the Future of Smart Cities-Figure 2. Traditional growth model vs. adapted growth model Source: Adapted after Purdy & Daugherty, 2016

The use of AI is not limited to smart buildings or transportation. It covers a wide range of application from medical diagnosis, to robot control and virtual assistance scientific tools. Nowadays, AI can be encountered in many services such as: cars speech recognitions, industrial robots, intelligent vacuum cleaners or fridges and so further. It can also be used in smart homes which permits by using hundreds or even thousands of sensors to provide services according to our preferences such as: ambient assisted living, energy saving etc. According to Skouby et al. (2014), AI also can be utilized in smart homes by adding personalized features in form of context awareness which allows AI to move beyond automation level. These authors designed a four-layer pyramid which encases the ICTs based infrastructure for future smart cites (Figure 3).

Artificial Intelligence and the Future of Smart Cities-Table 1. Smart city complex system factors

Figure 2 underlines the importance of traditional production factors such as capital and labour in achieving and driving growth which arises when either stock of capital or labour increase or they are more effectively used.  Total factor productivity (TFP) represents the growth enhanced by the use of technology and innovation. Besides these traditional factors, Purdy and Daugherty (2016) consider that AI can be seen as a new production factor, a capital-labour hybrid that will lead to significant growth opportunities. This is due to the advancement made in AI, which allows nowadays to replicate some labour activities at a greater and faster scale than humans (e.g. virtual text assistance, self-learning machines) (Purdy and Daugherty, 2016).

Artificial Intelligence and the Future of Smart Cities-Figure 1. Smart people, smart ICTs and smart cities

According to other authors “water, sewer, transportation, electricity, telecommunications, housing, healthcare, education — all of these functions—will have to be built from the ground up” (Glasmeier & Christopherson, 2015). This search is facilitated by the evolution of ICTs in general and of AI in particular. AI offers possibilities to replace the human being in complex and dangerous activities. But, smart cities start from smart human capital (Shapiro, 2006; Holland, 2008), because only smart people can create smart ICTs equipped with AI (Figure 1). These people and technologies will solve, by creativity and cooperation, problems associated with urban agglomerations, pollution, the depletion of some natural resources etc.

Artificial Intelligence and the Future of Smart Cities-Figure 9. AI influence on the environment

The participants consider that AI development will increase the energy consumption and e- waste (M=3.60, SD=1.03), but will improve also the level of citizens’ information on the environmental changes (M=3.60, SD=.49). The contribution to CO2 emissions is on the fourth places (M=3.40, SD=.49), followed by the attracting of the community members to environmental actions (M=3.00, SD=.64). In the analysis of the statically significant differences by gender, female participants scored significantly higher (M=4.00, SD=.64) than male participants (M=3.63, SD=.77) in the case of the information of citizens on the environmental changes (M=3.80, SD=.75 vs. M=3.72, SD=.75) and the attracting of community members to environmental actions (M=3.00, SD=.90 vs. M=2.63, SD=.88). The analysis on age category, the results revealed that the 26-30 age group scored the highest at both questions (Figure 9).