Learning through techniques as Machine Learning (ML) and Deep Learning is now possible. Thanks to the development of Artificial Intelligence. [1] They have gained appeal and are some hot topics, with applications in practically every sector of human endeavours. Among these are, to name a few, Voice Processing [2], Computer Vision [3], Cyber Security [4], and Medical Diagnosis [5]. Despite being influenced by the biological brain, the learning and memory encoding processes in humans are not directly tied to these algorithms. These artificial neural networks' (ANNs) [6] learning processes for changing masses and biases are standardized on optimization strategies and the minimizing of loss and error functions. As larger pool of nascent data is fed into the system, the ANNs currently use a huge count subjecting hyperparameters [7] that are fixed via some ad hoc approach for better prediction. These synaptic alterations are based primarily on facts and lack or have little solid theoretical support. Additionally, these methods need a huge quantity of training data to be able to accurately forecast or classify the distribution of the target classes.
ANNs have achieved great success, but when it comes to doing tasks like natural language processing [8], they fall well short of the human intellect. Researchers are concentrating on creating biologically inspired algorithms and architectures in order to utilize the remarkable learning capabilities of the Homo sapiens’ brain alongside better understandability of the same. This is being done in relation to memory encoding and learning. One of the brain's most intriguing traits is its capacity for "Chaos" – the phenomenon whereby straightforward deterministic nonlinear systems exhibit complex unexpected and random – like behavior. Electroencephalogram (EEG) signals [9] are known to have chaotic dynamics [10]. A neural system's sensitivity to little changes in internal functioning characteristics aids in producing the optimal response to various influences. This characteristic resembles the chaotic systems' dynamical characteristics. Furthermore, it is evident that the brain is constantly switching between several states rather than returning to homeostasis after a transient. For this reason, it is hypothesized that the brain can display a variety of behaviors, including periodicity in orbits, weak nature of chaotic dynamics, as well as strong nature of chaos, depending on the functional parameters of the neurons. Cerebral networks, that are made up of trillions of neurons, exhibit chaotic activity, but so is the scenario for the dynamics of individual neurons at both the cellular as well as subcellular levels. These neurons' ability to build impulse trains is what allows the brain to transmit and store information. When various ions pass across the axonal membrane and affect the voltage across it, action potentials or impulses are produced. For the communication bridging the ion passages and the axonal membrane, Huxley and Hodgkin initially put forth a dynamic system's model that is able to produce accurate action potentials [11]. Later, it was suggested to use its streamlined counterparts, for instance, the Hindmarsh-Rose [12] and the Fitzhugh-Nagumo model [13][14]. These models all display chaotic behavior.
Recurrent neural networks [15][16] are one type of artificial neural network that exhibits chaotic dynamics; however, as far as we are aware, none of these proposed architectures thus far subjecting classification tasks show chaos at the details of individuality of neurons. Though, other chaotic neuron models have been proposed as a theoretical description for memory encoding inside our brain.
One of these models is the Aihara model [17], that has been applied to cognitive tasks in the network's erratic periodic orbits [18]. Freeman, Kuzma, and colleague developed chaotic simulations that were motivated by the mammalian sensory pathway to demonstrate the process of memorizing scents [19–21]. Chaos in neural networks has also been studied by Tsuda and others. Globally coupled chaotic maps' dynamical properties have been studied by Kaneko, who hypothesized that these networks would be able to handle biological data.
Generalized Luröth Series (GLS) 1D chaotic map neurons make up ChaosNet, an artificial neural network (ANN) [22]. This network can learn from a small number of training examples to perform classification tasks. To utilize some of the best characteristics of biological neural networks, ChaosNet was developed. It has been demonstrated that, while using significantly fewer training samples than traditional ANNs, it can perform difficult classification tasks on par with or better than conventional ANNs.
ChaosNet, which was inspired by biological neurons, uses a property similar to the "spike-count rate" of the firing of chaotic neurons as a neural code for learning. Additionally, the network can exhibit a hierarchical architecture that can incorporate information as it is transmitted to deeper, higher levels of the network. Generalized Luröth Series, or GLS, is a piecewise linear 1D chaotic map that represents the neuron that we suggest. Examples of GLS include the well-known Tent map, Binary map, and its skewed relatives. The sorts of GLS neurons that are employed in ChaosNet are
$${T}_{Skew-Binary}\left(x\right)=\left\{\begin{array}{cc}\frac{x}{b}& 0\le x<b\\ \frac{\left(x-b\right)}{\left(1-b\right)}& b\le x<1\end{array}\right.$$
and
$${T}_{Skew-Tent}\left(x\right)=\left\{\begin{array}{cc}\frac{x}{b}& 0\le x<b\\ \frac{\left(1-b\right)}{\left(1-b\right)}& b\le x<1\end{array}\right.$$
Figure 2. The architecture of ChaosNet [24] Luroth neural networks for purposes relating to classification. 𝐶1, 𝐶2,. . ., 𝐶𝑛 are the unit dimensional GLS neurons. Each neuron initially exhibits 𝑞 units of normalized neuronal activity. The input to the network, or the normalized collection of stimuli, is denoted by the {𝑥𝑖}𝑛𝑖=1. When a GLS neuron 𝐶𝑖's chaotic activity value 𝐴𝑖(𝑡), starting from initial neural activity (𝑞), reaches the 𝜀 -neighborhood of stimulus, it stops firing chaotically. This neuron has a “firing time” of 𝑁𝑖 ms. 𝐴𝑖(𝑡) contains topological transitivity symbolic sequence feature 𝑝𝑖. This feature is extracted from 𝐴𝑖(𝑡) of the 𝐶𝑖's GLS-neuron.
A cryptocurrency [25], often called as a crypto-currency or just a "crypto," is a sort of digital money that is supported or maintained by no single central body, such as a bank or government. It is a decentralized means of verifying that the parties to a transaction genuinely have the funds they claim to have, eliminating the need for traditional middlemen like banks when money is being transferred between two businesses. Digital ledgers, which are computerized databases that use safe encryption to protect transaction records, regulate the production of new currencies, and confirm ownership transfers, are used to keep individual coin ownership records. Cryptocurrency is typically not authorized by a centralized unit and isn't exist in tangible form like paper money. In contrast to digital currencies managed by a central bank, cryptocurrency usually employs decentralized control (CBDC). When a cryptocurrency [26] is coined, generated in anticipation of issuance, or released by a single issuer, it is considered centralized. When utilized with decentralized governance, each cryptocurrency uses distributed ledger technology, generally a blockchain, which acts as a public database of financial transactions. Currency, commodities, and stocks are traditional asset classes and macroeconomic indicators with moderate sensitivity to cryptocurrency returns.
Financial or personal gain is the intended outcome of cryptocurrency fraud, which is a dishonest behavior in the cryptocurrency business. By convincing their unwitting victims to take an action, such as clicking on a link or disclosing personal information, scammers, and hackers on the internet hope to make some quick cash.
For cryptocurrency scams, criminals frequently try to gain access to a victim's digital wallet in order to steal their cryptocurrency assets. Typically, they will ask you to connect your wallet to a bogus website or deceive you into giving them access to your wallet's private keys. Cryptocurrency Fraudulences can be of many types, like,
Phishing: Even though fraudsters are nothing new, individuals continue to fall for them every day. A malicious hyperlink in an inbox or a bogus website that occasionally uncannily resembles its genuine counterpart can both be used in phishing scams. Your personal information, such as your internet passwords or the private keys to your crypto wallet, may be stolen using the link or website.
Middle Man attacks: Man-in-the-middle assaults are a technique used by con artists to obtain your personal information, much to phishing scams. To access your bitcoin wallet or private account information, a guy will disrupt a Wi-Fi session on a broad network as opposed to doing so through links. Use a VPN to secure your data while depositing cryptocurrency to avoid this.
Investors’ Scam: Investment managers that offer to help you make significant improvements on your portfolio may be posing as fraudsters. These dishonest people will entice customers to transmit them cryptocurrencies and may even promise to increase its worth by 50 times. Forbes Advisor does caution, though, that "if you comply with their demands, kiss goodbye to your cryptocurrency." With this scam, the con artist is probably deceiving several people, taking their cryptocurrency with them, and then vanishing.
Pump & Dump: This is true for both regular stock markets and cryptocurrency marketplaces. When a coin launches, its owners sell all their holdings, which is known as a pump and dump strategy. As a result, the price reaches an erroneous peak before dropping sharply after the initial public offering is over. False statements made about a project that cause a lot of hype can make these tactics worse.