Random Heterogeneous Neurochaos Learning Architecture for Data Classification

Remya Ajai A S,Nithin Nagaraj

2024-10-31

Abstract:Inspired by the human brain's structure and function, Artificial Neural Networks (ANN) were developed for data classification. However, existing Neural Networks, including Deep Neural Networks, do not mimic the brain's rich structure. They lack key features such as randomness and neuron heterogeneity, which are inherently chaotic in their firing behavior. Neurochaos Learning (NL), a chaos-based neural network, recently employed one-dimensional chaotic maps like Generalized Lüroth Series (GLS) and Logistic map as neurons. For the first time, we propose a random heterogeneous extension of NL, where various chaotic neurons are randomly placed in the input layer, mimicking the randomness and heterogeneous nature of human brain networks. We evaluated the performance of the newly proposed Random Heterogeneous Neurochaos Learning (RHNL) architectures combined with traditional Machine Learning (ML) methods. On public datasets, RHNL outperformed both homogeneous NL and fixed heterogeneous NL architectures in nearly all classification tasks. RHNL achieved high F1 scores on the Wine dataset (1.0), Bank Note Authentication dataset (0.99), Breast Cancer Wisconsin dataset (0.99), and Free Spoken Digit Dataset (FSDD) (0.98). These RHNL results are among the best in the literature for these datasets. We investigated RHNL performance on image datasets, where it outperformed stand-alone ML classifiers. In low training sample regimes, RHNL was the best among stand-alone ML. Our architecture bridges the gap between existing ANN architectures and the human brain's chaotic, random, and heterogeneous properties. We foresee the development of several novel learning algorithms centered around Random Heterogeneous Neurochaos Learning in the coming days.

Machine Learning

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that the existing neural networks cannot fully simulate the complexity of the human brain in terms of structure and function, especially lacking two key features: randomness and neuronal heterogeneity. Specifically:

1. **Limitations of existing neural networks**:
- Although existing artificial neural networks (ANN) and deep neural networks (DNN) perform well in many tasks, they do not fully imitate the rich structure of the human brain.
- These networks lack the randomness and heterogeneity that show chaotic behavior like human brain neurons.

2. **Introducing random heterogeneity**:
- The paper proposes a new architecture - Random Heterogeneous Neurochaos Learning (RHNL). By randomly placing different types of chaotic neurons (such as Logistic mapping and generalized Lüroth sequence mapping) in the input layer, it simulates the randomness and heterogeneity of the human brain neural network.

3. **Improving classification performance**:
- This research aims to verify whether RHNL can outperform the existing uniform and fixed - heterogeneity neurochaos learning architectures (NL) on multiple public datasets, especially in the low - sample training scenario.

4. **Combining with traditional machine learning methods**:
- RHNL is not only tested as an independent classifier but also used in combination with traditional machine - learning methods (such as SVM, kNN, etc.) to evaluate its performance in different tasks.

### Formula summary

- **Logistic mapping**:
\[
x_{z + 1}=r x_z(1 - x_z)
\]
where \(0.0\leq x_z < 1.0\) and the bifurcation parameter \(0 < r\leq4.0\).

- **Skew - tent mapping**:
\[
C_{\text{Skew - Tent}}(z)=\begin{cases}
\frac{z}{b},&0\leq z < b\\
\frac{1 - z}{1 - b},&b\leq z < 1
\end{cases}
\]
where \(z\in[0.0, 1.0)\) and \(0.0 < b < 1.0\).

- **Lyapunov exponent**:
\[
\lambda_G(x)=\lim_{k\rightarrow\infty}\frac{1}{k}\sum_{j = 0}^{k - 1}\ln|G'(x_j)|
\]

- **Energy calculation**:
\[
E_c=\sum_{t = 1}^M|c(t)|^2
\]
where \(M\) is the discharge time.

- **Shannon entropy**:
\[
H(s)=-\sum_{i = 1}^2p_i\log_2(p_i)
\]
where \(p_1\) and \(p_2\) are the probabilities of symbols 0 and 1 respectively.

### Conclusion

By introducing random - heterogeneity neurons, this paper aims to more closely simulate the complex structure of the human brain and has verified the effectiveness of this new architecture on multiple datasets, especially its superior performance in classification tasks.