Deep Learning Models: Powerful AI Techniques Explained

Deep Learning Models – An incredibly young and boundless field. It’s like a rebellious teenager in AI—constantly evolving, changing views almost daily, and refusing to be boxed into rigid theories. One day, an idea is revolutionary; the next, it’s outdated.

Deep Learning is not just strengthening AI; it’s practically the backbone of all AI-driven innovations today. Its execution is so successful that even AI skeptics are raising their eyebrows (and possibly reconsidering their career choices).

The principles of deep learning are sprinting ahead of time. Given enough computing power and storage (because let’s be honest, DL eats resources like a teenager eats snacks), it can analyze massive amounts of data. Whether it’s recognizing patterns, extracting features, or decoding the mysteries hidden in images, videos, speech, or social media—DL does it all.

But here’s the fun part—just when we think we’ve got DL all figured out, it surprises us. One breakthrough, and the game changes again. It’s unpredictable, exciting, and, if you ask me, one of the best rollercoasters in the tech world. The future? Well, I’d say buckle up. Deep Learning isn’t just evolving; it’s rewriting the rules of AI as we speak.

Imagine an AI that understands sarcasm, predicts the next viral meme, or even helps doctors detect diseases earlier than ever—this is the power of deep learning. And trust me, we’re only scratching the surface.

• Part-1: Deep Learning (DL) – Introduction to Basics
• Part-2 : The ABC of Deep Learning – A New Frontier in The Digital Age

Deep Learning – Outlook

Deep learning is used with too much ease to predict the unpredictable. In our opinion “We all are so busy in creating artificial intelligence by using a combination of non-bio neural networks and natural intelligence rather than exploring what we have in hand.

AILabPage defines Deep learning is “Undeniably a mind-blowing synchronisation technique applied on the bases of 3 foundation pillars large data, computing power, skills (enriched algorithms) and experience which practically has no limits“.

• Learning Through Representations: Deep learning uncovers hidden data patterns, refining decision-making like a detective solving mysteries—only with math, not magnifying glasses.
• Popularity of Neural Networks: Neural networks are the rockstars of AI, mimicking human thinking (minus the coffee breaks) and revolutionizing everything from chatbots to self-driving cars.
• AI Breakthrough: Deep learning isn’t just innovation—it’s the driving force behind AI’s ability to learn, adapt, and make decisions, marking a true technological renaissance.
• Human vs. AI Learning: Humans rely on memory and emotions; AI learns through patterns and numbers. We have nostalgia, AI has backpropagation—different paths, same goal.
• Structured vs. Adaptive Learning: While humans learn through trial and error, deep learning models process vast data efficiently—no procrastination, no forgetting, just learning.
• Shaping the Future: Deep learning is redefining industries, driving automation, and making AI smarter—whether for medical breakthroughs or just recognizing cats online.

Hoping that there would be no time when we need to do the reverse i.e. where will use Artificial Intelligence to create Natural Intelligence in future. So in nutshell, we need to be careful not to translate any of our research or machine learning into human experiences.

“I think people need to understand that deep learning is making a lot of things, behind the scenes, much-better” – Sir Geoffrey Hinton

Introduction – Deep Learning?

Deep learning, also called as a subset of machine learning which is a specialist with an extremely complex skill set in order to achieve far more better results from the same data set. It purely on the basis of NI (Natural Intelligence) mechanics of the biological neuron system. It has a complex skill set because of the methods it uses for training i.e. learning in deep learning is based on “learning data representations” rather than “task-specific algorithms,” which is the case for other methods.

In short, it’s okay to say that deep learning models, which are essentially mathematical, broadly fall into two categories: Neural Network-Based Models and Non-Neural Network-Based Models. Additionally, hybrid models combining both approaches also exist, as outlined below.

• Neural Network-Based Models – Includes CNNs, RNNs, transformers, and generative models, designed for tasks like image recognition, NLP, and sequential data processing.
• Non-Neural Network-Based Models – Covers probabilistic, evolutionary, and sparse models that focus on optimization, uncertainty modeling, and computational efficiency.
• Hybrid & Emerging Approaches – Combines neural and non-neural techniques (e.g., neuroevolution, Bayesian deep learning) to enhance AI adaptability and performance.

Neural models leverage architectures like CNNs, RNNs, and transformers, while non-neural approaches use probabilistic, evolutionary, and efficient strategies to optimize AI applications across domains like vision, NLP, and optimization.

Neural Network-Based Models

At AILabPage, we’ve seen how Neural Network-based Mathematical Learning Models leveraging architectures like CNNs, RNNs, and GNNs, these models go beyond pattern recognition—they tackle complex mathematical problems head-on. Whether it’s optimizing functions, solving differential equations, or estimating uncertainties, they integrate data-driven learning with fundamental mathematical principles, unlocking new possibilities in science, engineering, and beyond.

🤖 Feedforward Neural Networks (FNNs) – Concept of Feedforward Learning

Feedforward Neural Networks (FNNs) are the simplest form of artificial neural networks, where information moves in only one direction—from the input layer through hidden layers to the output layer, without looping back.

Key Architectures

• MLP (Multilayer Perceptron): A classic FNN with multiple layers of neurons using activation functions like ReLU, Sigmoid, or Tanh, enabling hierarchical feature learning and complex pattern recognition in deep learning models.
• ELM (Extreme Learning Machines): A specialized FNN that trains extremely fast by randomly initializing weights and analytically solving output weights.
• RBFN (Radial Basis Function Networks): Uses radial basis functions as activation functions, particularly useful in pattern recognition and function approximation.

Use Cases

• Classification: Image recognition, spam detection, and speech classification.
• Regression: Stock price prediction, weather forecasting, and real estate valuation.

These networks are the foundation of deep learning and are primarily used for supervised learning tasks.

🖼 Convolutional Neural Networks (CNNs) – Revolutionized Computer Vision

Convolutional Neural Networks (CNNs) have reshaped how machines perceive and interpret visual data, much like how human eyes process the world. These networks extract hierarchical features—edges, textures, patterns, and complex objects—using specialized convolutional layers.

Unlike traditional image processing, CNNs automatically learn feature representations without manual engineering, making them a backbone of modern AI-driven vision tasks.

Key Architectures & Evolution

• LeNet – The first CNN, designed for digit recognition.
• AlexNet – Sparked the deep learning revolution in 2012, winning the ImageNet challenge.
• VGG – Introduced deeper networks with smaller convolutional filters.
• ResNet – Solved the vanishing gradient problem with residual connections.
• EfficientNet – Optimized for accuracy and efficiency with compound scaling.
• MobileNet – Designed for mobile and edge devices with lightweight architecture.

Real-World Applications of CNNs

• Image Recognition: Powering everything from face recognition to autonomous vehicles.
• Object Detection: Used in surveillance, robotics, and industrial automation.
• Medical Imaging: Identifying tumors, fractures, and diseases like diabetic retinopathy.

CNNs are not just about pixels—they decode the visual world, making AI truly see and understand!. From the early days of LeNet (the pioneer) to ultra-deep architectures like ResNet and efficient lightweight models like MobileNet, CNNs have continued to push the boundaries of image understanding.

🔁 Recurrent Neural Networks (RNNs) – Sequence modeling and recurrent structures

Unlike traditional neural networks that process data independently, Recurrent Neural Networks (RNNs) introduce a memory mechanism, making them ideal for sequential data like speech, time-series, and natural language. RNNs learn from past inputs to influence future outputs, mimicking human cognitive processes—like remembering words in a sentence before predicting the next one.

However, vanilla RNNs suffer from vanishing gradients, limiting their ability to capture long-range dependencies. To overcome this, architectures like LSTM (Long Short-Term Memory) and GRU (Gated Recurrent Units) were developed, enhancing the ability to retain information over long sequences.

Key Variants of RNNs

• Vanilla RNN – Basic recurrent structure, struggles with long-term dependencies.
• LSTM (Long Short-Term Memory) – Solves vanishing gradients using memory cells & gates.
• GRU (Gated Recurrent Unit) – A simpler alternative to LSTM, reducing computational load.
• Bi-RNN (Bidirectional RNN) – Processes data both forward and backward for deeper context.

Why RNNs Are Being Replaced by Transformers

• Parallelization – Transformers process sequences all at once, unlike RNNs’ step-by-step approach.
• Long-Term Dependencies – RNNs struggle with distant relationships, while self-attention in Transformers efficiently captures them.
• Computational Efficiency – Transformers leverage parallel computation, making them faster and more scalable.

That said, RNNs still hold value in areas where sequence-by-sequence processing is intuitive and effective!

Real-World Applications of RNNs

🔹 Speech Recognition – Powering voice assistants and real-time transcription.
🔹 Time-Series Forecasting – Predicting stock prices, weather patterns, and sensor data.
🔹 Language Modeling – Used in chatbots, sentiment analysis, and text generation.

While Transformers dominate NLP today, RNNs paved the way for sequential learning and remain relevant in many domains.

📊 Graph Neural Networks (GNNs) – Graph-based learning

Graph Neural Networks (GNNs) are transforming AI by enabling deep learning on graph-structured data. Unlike traditional neural networks that process data in grids or sequences, GNNs leverage the power of relationships, making them ideal for tasks where connections matter as much as individual data points.

What is Graph-Based Learning?

Graph-based learning is a paradigm where AI models analyze entities (nodes) and their relationships (edges) rather than treating data as isolated units. This approach is crucial for domains where interactions define meaning—think social networks, molecular structures, and recommendation systems. GNNs learn from these connections, capturing both local and global patterns within complex networks.

Popular GNN Models – The Building Blocks

• GCN (Graph Convolutional Networks) – Extends convolution operations to graphs, allowing models to aggregate information from neighbors, making it great for semi-supervised learning.
• GAT (Graph Attention Networks) – Uses an attention mechanism to weigh node importance dynamically, improving performance on diverse graph structures.
• GraphSAGE – Generates embeddings for unseen nodes by sampling and aggregating neighboring nodes, making it scalable for large graphs.

Use Cases – Where GNNs Shine

• Social Network Analysis – Understanding user interactions, detecting communities, and improving recommendation algorithms.
• Drug Discovery – Predicting molecular properties, accelerating new drug development, and enhancing biomedical research.
• Fraud Detection – Identifying suspicious transaction patterns by analyzing relationships in financial networks.

GNNs represent a shift in AI thinking—from isolated data points to interconnected intelligence. As real-world problems often exist in networks, GNNs are unlocking new ways to model and understand the world.

 🔍 Attention Based Models – Transformers evolution from RNNs

The evolution of AI has been a journey of continuous refinement, and transformer-based models represent one of the biggest leaps forward. These models have redefined how AI understands and generates language, processes images, and even bridges multiple modalities.

From RNNs to Attention-Based Models – A Paradigm Shift

Recurrent Neural Networks (RNNs) once dominated sequence processing, but their limitations—like struggling with long-range dependencies—paved the way for something more powerful: attention-based models. Transformers, introduced with the attention mechanism, revolutionized deep learning by enabling parallel processing and capturing contextual relationships more effectively. This shift made models more efficient, scalable, and capable of handling vast amounts of data.

Key Architectures – The Pillars of Modern AI

Applications – Beyond Words to Multimodal AI

Transformer-based models power a wide spectrum of AI applications:
* NLP – From chatbots to machine translation, these models enable human-like interactions.
* Image Processing – Vision Transformers are redefining computer vision, from medical imaging to autonomous systems.
* Multimodal AI – The fusion of text, vision, and audio processing is unlocking new possibilities, from AI-generated videos to intelligent assistants that “see and talk.”

The transformer era is here to stay, pushing AI beyond structured tasks into creative, context-aware intelligence. As these models continue to evolve, they will reshape how humans and AI interact, bringing smarter, more intuitive experiences to the world.

🎭 Generative Models – Introduction to Generative AI

Generative AI has come a long way, redefining creativity, automation, and problem-solving. These models are not just about generating images or text—they embody the power of AI to understand, learn, and recreate complex patterns. Let’s explore some key types of generative models shaping the future.

Autoencoders & VAEs – Learning Features, Compressing Data

Autoencoders and Variational Autoencoders (VAEs) serve as the foundation for feature learning and data compression. By encoding data into a lower-dimensional space and then reconstructing it, they help extract meaningful representations. VAEs take this further by adding a probabilistic twist, making them excellent for tasks like denoising, anomaly detection, and generating new but coherent variations of existing data.

GANs – The Driving Force Behind Deepfakes & Image Synthesis

Generative Adversarial Networks (GANs) introduced the world to AI-driven creativity. From generating ultra-realistic faces to enhancing image quality, GANs operate through a clever adversarial process—where two networks, the generator and discriminator, compete to refine outputs. While they have powered deepfake technology, they also contribute to artistic creations, medical imaging, and game design.

Diffusion Models – The Future of Generative AI

Diffusion models have emerged as the next breakthrough in generative AI. Techniques like Stable Diffusion and DALL-E have set new benchmarks in text-to-image synthesis, bringing photorealistic results with remarkable flexibility. These models work by progressively refining noise into structured outputs, leading to highly detailed and controlled image generation. Their potential goes beyond visuals—expect innovations in video generation, molecule design, and even AI-driven art.

As we move forward, generative AI is evolving from simple pattern replication to genuine creativity. It’s an exciting space where technology meets imagination, and the best is yet to come!

🏆 Reinforcement Learning (RL) Models -Deep Learning in decision-making and control

Deep Reinforcement Learning (DRL) is where deep learning meets decision-making, enabling AI to learn through trial and error—just like humans. By interacting with an environment, receiving rewards, and optimizing strategies, DRL models drive breakthroughs in automation, control, and intelligent decision-making.

Deep Learning in Decision-Making & Control

RL isn’t just about making choices—it’s about learning optimal actions in dynamic environments. Unlike supervised learning, which relies on labeled data, RL agents explore, experiment, and improve over time. With deep learning, these agents can handle complex, high-dimensional problems, making RL a key player in AI-driven autonomy.

Popular DRL Models – Powering Intelligent Systems

• Deep Q-Networks (DQN) – Uses deep learning to approximate Q-values, making it capable of playing video games at superhuman levels.
• PPO (Proximal Policy Optimization) – A stable and efficient policy gradient method, widely used in robotics and real-world decision-making tasks.
• AlphaZero – A revolutionary self-learning system that mastered chess, Go, and shogi through pure reinforcement learning, without human input.

Real-World Applications – Where RL Takes the Lead

• Robotics – Training robots to adapt, optimize movements, and perform complex tasks.
• Autonomous Driving – Teaching self-driving cars to navigate safely in dynamic environments.
• AI Agents – Enhancing digital assistants, game-playing AI, and autonomous decision-makers.

Reinforcement Learning is shaping the future of AI, creating systems that don’t just follow rules but learn, adapt, and improve over time. The journey from trial and error to mastery is what makes DRL an exciting and powerful force in artificial intelligence.

Specialized & Emerging Neural Networks

AI is evolving beyond traditional architectures, with specialized and emerging neural networks pushing the boundaries of intelligence. At AILabPage, we love diving deep into AI’s evolving landscape, where cutting-edge models are reshaping the way we think about learning, reasoning, and adaptability.

Here’s a hands-on breakdown of some of the most fascinating advancements in neural network-based AI.

Each of these models represents a leap forward in AI’s evolution to tackle unique challenges, from adaptive real-time learning to mimicking how the human brain processes information. As AI advances, these specialized networks pave the way for smarter, more efficient, and adaptable models, unlocking new possibilities in artificial intelligence and beyond.

Non-Neural Network-Based Models in Deep Learning

These models embrace probability, logic, evolution, and physics to tackle challenges where traditional neural networks may fall short. At AILabPage, our hands-on work shows that probabilistic models help in uncertainty estimation, symbolic AI brings structured reasoning, and physics-based methods solve complex simulations efficiently. Evolutionary algorithms drive optimization, pushing AI towards adaptability and efficiency. As we explore the deep tech space, the fusion of NN and Non-NN approaches is unlocking smarter, more interpretable AI solutions—ones that think beyond raw data and embrace true intelligence.

Probabilistic Models – Non-Neural Networks

Utilize probability distributions to model uncertainty, operating independently of neural networks.

Hidden Markov Models (HMMs) – Modeling Sequential Data

• How It Works: Models sequences by assuming a Markov process, where the probability of each state depends only on the previous state.
• Key Feature: Captures temporal dependencies in sequential data.
• Applications:
  • Speech Recognition: Transcribing spoken language into text.
  • NLP: Part-of-speech tagging, named entity recognition.
  • Bioinformatics: DNA sequence analysis, protein structure prediction.

Hidden Markov Models (HMMs) are a cornerstone of probabilistic modeling, designed to handle sequential data where the underlying states are not directly observable. By leveraging probabilities and transitions between states, HMMs excel in tasks where temporal patterns and dependencies are critical.

What is Sequential Modeling? – Sequential modeling is a paradigm where AI systems analyze data points in a sequence, capturing dependencies between past and future states. This approach is essential for domains where the order of data matters—think speech recognition, natural language processing (NLP), and time-series forecasting. HMMs learn from these sequences, enabling them to predict future states or classify sequences based on observed patterns.

Gaussian Mixture Models (GMMs) – Density Estimation

Gaussian Mixture Models (GMMs) are probabilistic models used for density estimation and clustering. Unlike traditional clustering methods, GMMs assume that data points are generated from a mixture of several Gaussian distributions, making them ideal for capturing complex, multi-modal data distributions.

• How It Works: Represents data as a mixture of multiple Gaussian distributions, each with its own mean and variance.
• Key Feature: Flexible modeling of complex, multi-modal data.
• Applications:
  • Clustering: Grouping similar data points into clusters.
  • Anomaly Detection: Identifying outliers in data.
  • Density Estimation: Modeling the probability distribution of data.

What is Density Estimation? – Density estimation is the process of modeling the probability distribution of data. It’s crucial for understanding the underlying structure of data, identifying clusters, and making probabilistic predictions. GMMs are particularly useful when data cannot be explained by a single Gaussian distribution, allowing for more flexible and accurate modeling.

Use Cases – Where Probabilistic Models Shine

Why Probabilistic Models Matter

Probabilistic models like HMMs and GMMs represent a shift in AI thinking—from deterministic predictions to probabilistic reasoning. They provide a framework for understanding uncertainty, capturing dependencies, and modeling complex data distributions. As real-world problems often involve uncertainty and sequential data, probabilistic models are unlocking new ways to analyze and interpret the world.

Other Advanced Non-Neural Network-Based Deep Learning Models

Comparison: Neural Network vs. Non-NN Models

Comparison: Neural Network vs. Non-NN Models

At AILabPage, we often find ourselves at the crossroads of choosing between Neural Network (NN)-based models and Non-Neural Network (non-NN) models. Both have unique strengths, and understanding when to use each is key to solving real-world problems effectively. Let’s dive into the nuances, trade-offs, and magic of hybrid approaches.

When to Use NN-Based Models vs. Non-NN-Based Models

Neural Network (NN) Models are your go-to when:

• Data is Complex and High-Dimensional: Think images, audio, or text. NNs excel at capturing intricate patterns in unstructured data.
• Accuracy is Paramount: For tasks like image recognition, language translation, or game-playing AI, NNs often deliver state-of-the-art performance.
• Adaptability is Needed: NNs can learn and adapt from data, making them ideal for dynamic environments like autonomous driving or real-time recommendations.

Non-NN Models shine when:

• Interpretability is Critical: In fields like healthcare or finance, understanding why a model made a decision is as important as the decision itself.
• Data is Structured and Small: For tabular data or small datasets, traditional models like decision trees or SVMs often outperform NNs.
• Computational Resources are Limited: Non-NN models are lightweight and efficient, making them perfect for edge devices or real-time systems.

Performance Trade-Offs: Accuracy, Complexity, and Interpretability

Hybrid Models – Combining Neural Networks + Symbolic AI

At AILabPage, we’re big fans of hybrid models—the best of both worlds. By combining the pattern recognition power of Neural Networks with the logical reasoning of Symbolic AI, hybrid models are redefining what’s possible in AI.

Why Hybrid Models?

• Interpretability Meets Power: Neural Networks handle the heavy lifting of pattern recognition, while Symbolic AI provides logical reasoning and explainability.
• Robustness: Hybrid models are less prone to overfitting and can generalize better across diverse tasks.
• Real-World Applications:
  • Healthcare: Diagnosing diseases with both data-driven insights and expert rules.
  • Autonomous Systems: Combining perception (NNs) with decision-making (Symbolic AI) for safer navigation.
  • Explainable AI: Building trust by providing clear, logical explanations for AI decisions.

Comparison of Neural Network vs. Non-NN Models

Why This Matters

At AILabPage, we believe that no single model fits all problems. The choice between NN and Non-NN models depends on the task, the data, and the desired outcome. While NNs push the boundaries of what’s possible, Non-NN models keep us grounded with their simplicity and interpretability. And when we combine the two? That’s where the real magic happens.

Optimization and Efficiency Improvements in Deep Learning

At AILabPage, we believe that the future of AI lies not just in building smarter models but also in making them faster, lighter, and more efficient. Optimization and efficiency improvements are the unsung heroes of deep learning, enabling us to push boundaries while keeping resources in check. Let’s explore some of the key techniques driving this transformation.

Hyperparameter Optimization: Fine-Tuning for Excellence

Hyperparameters are the backbone of any deep learning model, and finding the right combination can feel like searching for a needle in a haystack. Thankfully, we have powerful tools to guide us:

• Grid Search: A systematic approach to explore every possible combination—simple but effective.
• Bayesian Optimization: A smarter way to navigate the hyperparameter space, balancing exploration and exploitation.
• Genetic Algorithms: Inspired by natural selection, these algorithms evolve the best hyperparameters over time.

These methods help us fine-tune models to perfection, ensuring they perform at their best without endless trial and error.

Efficient AI Techniques: Doing More with Less

As models grow larger and more complex, efficiency becomes critical. Here’s how we’re tackling the challenge:

• Model Pruning: Trimming the fat by removing unnecessary neurons or connections, making models leaner without sacrificing accuracy.
• Quantization: Reducing the precision of model weights and activations, speeding up inference and reducing memory usage.
• Knowledge Distillation: Transferring knowledge from a large, complex model to a smaller, faster one—like a mentor passing wisdom to a protégé.

These techniques allow us to deploy powerful AI even on resource-constrained devices, bringing intelligence to the edge.

Neuromorphic Computing: The Hardware Revolution

Efficiency isn’t just about software—it’s also about rethinking hardware. Neuromorphic computing takes inspiration from the human brain to create AI hardware that’s faster, more energy-efficient, and capable of real-time learning. Innovations like Intel Loihi are paving the way for a new era of AI, where hardware and software work in harmony to deliver unprecedented performance.

Why This Matters

Optimization and efficiency improvements are more than just technical advancements—they’re about making AI accessible, sustainable, and scalable. Whether it’s fine-tuning hyperparameters, pruning models, or innovating hardware, these efforts ensure that AI can thrive in the real world, solving problems faster and smarter than ever before.

Future of Deep Learning Model Architectures

The future of deep learning lies in smarter, more efficient, and ethical architectures. From Mixture-of-Experts to Quantum AI, these innovations are redefining how models learn, reason, and scale, paving the way for transformative advancements in AI.

As deep learning evolves, hybrid models, self-supervised learning, and ethical AI are shaping the next frontier. These architectures promise scalability, adaptability, and fairness, ensuring AI systems are not only powerful but also responsible and inclusive.

Conclusion – Deep learning is a fast-moving field, constantly evolving to tackle diverse data types—images, text, graphs, and audio—each demanding specialized models. Some prioritize speed and efficiency, while others focus on accuracy, interpretability, or adaptability. At AILabPage, our hands-on approach explores these innovations, from CNNs excelling in vision tasks to transformers redefining NLP, GNNs unlocking graph-based insights, and diffusion models pushing generative AI forward. The real challenge lies in selecting the right model for the right problem—balancing power, precision, and practicality. As AI models continue to evolve, bridging efficiency with reasoning, the future promises even more breakthroughs, making AI not just smarter but also more intuitive and impactful in real-world applications. The journey has just begun!

—

Points to Note:

All credits if any remains on the original contributor only. We have covered all basics around the Generative Neural Network. Though often such tasks struggle to find the best companion between CNNs and RNNs algorithms to look for information.

Books + Other material Referred

• Research through open internet, news portals, white papers and imparted knowledge via live conferences & lectures.
• Lab and hands-on experience of  @AILabPage (Self-taught learners group) members.
• NIPS 2016 Tutorial: GANs
• Unsupervised Representation Learning with Deep Convolutional GAN

Feedback & Further Question

Do you need more details or have any questions on topics such as technology (including conventional architecture, machine learning, and deep learning), advanced data analysis (such as data science or big data), blockchain, theoretical physics, or photography? Please feel free to ask your question either by leaving a comment or by sending us an  via email. I will do my utmost to offer a response that meets your needs and expectations.

========================== About the Author ===================

Read about Author at : About Me

Thank you all, for spending your time reading this post. Please share your opinion / comments / critics / agreements or disagreement. Remark for more details about posts, subjects and relevance please read the disclaimer.

FacebookPage    ContactMe      Twitter

=================================================================