The fundamental: BI and AI
Artificial Enhancement of biological intelligence 

Objectives

Previously, we briefly touch based on artificial intelligence and covered the problem people face when it comes to learning and applying AI. We also covered the key objectives of this course. In this section, our goal is to cover the fundamentals of biological intelligence (BI) and Artificial intelligence (AI). By the end of this section you'll have a firm grasp of the following:​​

• History of AI 

• BI vs AI

• Take away

• Next 

​​

​History

​

We'll oversimplify the fascinating history of AI by highlighting the key dates and names we think you should know:

​

1950 || Alan Turing proposed the Turing Test in his paper "Computing Machinery and Intelligence," introducing the idea of machine intelligence.

​

1956 || John McCarthy led the Dartmouth Conference where the term Artificial Intelligence (AI) was coined and considered one of the founding events of AI.

​

1950s-1960s || Allen Newell and Herbert A. Simon introduced early AI programs, such as the Logic Theorist and General Problem Solver (GPS), and showed that machines could solve logical problems.

​

1966 || Joseph Weizenbaum developed ELIZA, an early chatbot which demonstrated natural language processing (NLP).

​​

Therefore, the 1950s through the 1960s mark the birth of artificial intelligence. However, From 1970s through  the 1980s, the field of artificial intelligence was threatened due to funding cuts and skepticism. Weathering the storm of extension, from 1990s through 2000s we witnessed machine learning and the AI boom. Lastly, from the 2010s to now, we've witnessed the rise of AI application in various field and day-to-day activities:​

​

• 2011: IBM’s Watson won Jeopardy!, demonstrating AI’s capabilities in natural language understanding.

​

• 2012: The deep learning revolution took off when AlexNet, a neural network, won the ImageNet competition, greatly advancing computer vision.

​

• 2016: Google DeepMind’s AlphaGo defeated human Go champion Lee Sedol, marking a major achievement in AI strategy and reinforcement learning.

​

• 2020s: AI systems like OpenAI’s GPT-3 (2020) and ChatGPT (2022, based on GPT-3.5) revolutionized natural language processing, while GPT-4 (2023) further enhanced AI capabilities.

​

Needless, to say that the field has gone through a lot, it's still standing and is likely here to stay. Interested in learning more about the history of AI?

​

Watch this presentation from MIT's Chancellor for Academic Advancement, Prof. W. Eric Grimson, "Artificial Intelligence - Past, Present, Future"

​

That concludes it for our AI history. Moving on we'll be covering biological intelligence and artificial intelligence. 

​

​Biological intelligence (BI) & 

Artificial intelligence

​

​

Data Collection → 2. Data Processing (Cleaning, Transforming, Analyzing) → 3. Machine Learning / Deep Learning (Model Training, Evaluation) → 4. Data Storage (Databases, Model Repositories, Cloud Storage)

​

Fundamentals of AI (Artificial Intelligence) Explained

Artificial Intelligence (AI) is a broad field of computer science that focuses on creating machines capable of performing tasks that typically require human intelligence. These tasks include reasoning, learning, problem-solving, perception, language understanding, and decision-making. Below, we break down the core principles and concepts that form the foundation of AI.

1️⃣ Types of AI

AI can be categorized into three main types based on its capabilities:

a) Narrow AI (Weak AI)

• Definition: AI that is designed and trained to perform a specific task or a narrow set of tasks.

• Example: Virtual assistants like Siri or Alexa, image recognition software, or recommendation algorithms.

• Limitations: Narrow AI is highly specialized and cannot perform tasks outside of its programmed function.

b) General AI (Strong AI)

• Definition: AI that has the ability to understand, learn, and apply knowledge across a broad range of tasks, much like a human.

• Example: Theoretical AI that could perform any intellectual task that a human being can.

• Current Status: General AI is still in the research phase and does not yet exist.

c) Superintelligent AI

• Definition: AI that surpasses human intelligence in all aspects, including creativity, problem-solving, and decision-making.

• Future Vision: This is a hypothetical concept that could revolutionize humanity, but it remains speculative.

2️⃣ Core Concepts in AI

a) Machine Learning (ML)

• Definition: A subset of AI that focuses on building systems that can learn from data and improve over time without being explicitly programmed.

• Key Idea: Instead of following hardcoded rules, machine learning systems analyze patterns in data and make predictions or decisions based on those patterns.

• Types of ML:

  • Supervised Learning: The model is trained on labeled data (input-output pairs).

  • Unsupervised Learning: The model finds hidden patterns or structures in unlabeled data.

  • Reinforcement Learning: The system learns by interacting with an environment and receiving feedback in the form of rewards or penalties.

b) Deep Learning (DL)

• Definition: A subset of machine learning that uses neural networks with many layers (deep networks) to model complex patterns in large datasets.

• Key Idea: Deep learning can automatically extract features from data (like images or speech) and learn very intricate relationships, making it ideal for tasks like image recognition or natural language processing.

• Example: Convolutional Neural Networks (CNNs) for image classification, Recurrent Neural Networks (RNNs) for speech recognition.

c) Natural Language Processing (NLP)

• Definition: A field of AI that focuses on enabling machines to understand, interpret, and generate human language.

• Key Idea: NLP involves various tasks, such as text analysis, language translation, sentiment analysis, and conversational agents like chatbots.

• Applications: Chatbots, language translation tools, sentiment analysis in social media, voice assistants.

d) Computer Vision

• Definition: A field of AI that enables machines to interpret and understand visual information from the world, such as images or videos.

• Key Idea: It involves tasks like object detection, image segmentation, and facial recognition, allowing machines to “see” and process visual data.

• Applications: Self-driving cars, facial recognition systems, medical image analysis.

3️⃣ Key Techniques in AI

a) Neural Networks

• Definition: A type of machine learning model inspired by the human brain's structure, consisting of layers of interconnected "neurons" (nodes).

• Key Idea: Neural networks learn by adjusting the strength of connections between neurons during training, which allows them to make predictions based on data.

• Application: Used in deep learning for tasks like speech recognition, image classification, and natural language processing.

b) Reinforcement Learning

• Definition: A type of learning where an agent learns by interacting with an environment and receiving feedback in the form of rewards or punishments.

• Key Idea: The agent's goal is to maximize its cumulative reward over time by taking actions that lead to favorable outcomes.

• Application: Used in applications like robotics, game playing (e.g., AlphaGo), and autonomous driving.

c) Decision Trees and Random Forests

• Decision Trees: A model that makes decisions by splitting data into branches based on feature values. It's easy to understand and interpret.

• Random Forests: A collection of decision trees that improves accuracy by combining the predictions of multiple trees to make a final decision.

• Application: Widely used for classification tasks like predicting customer churn, medical diagnosis, and fraud detection.

4️⃣ Data in AI

a) Importance of Data

• Data as the Fuel for AI: AI systems, especially machine learning models, rely heavily on data to learn and make decisions. The more data an AI model has, the better it can learn patterns and improve over time.

• Data Quality: High-quality, clean, and relevant data is essential for building effective AI models. Poor data can lead to inaccurate predictions and biased outcomes.

b) Data Processing and Preprocessing

• Data Cleaning: Involves removing noise, missing values, and irrelevant information to ensure the model learns from accurate data.

• Feature Engineering: The process of selecting, modifying, or creating new features from raw data that help the model learn better.

• Normalization/Standardization: Rescaling the data to ensure that all features have a similar range and scale.

5️⃣ AI Applications

AI has diverse applications across multiple fields, including:

• Healthcare: AI is used for diagnosing diseases, analyzing medical images, and predicting patient outcomes.

• Finance: AI helps detect fraud, manage investments, and optimize trading strategies.

• Retail: AI enhances personalized recommendations, demand forecasting, and inventory management.

• Automotive: Self-driving cars use AI to navigate roads, avoid obstacles, and follow traffic rules.

• Entertainment: AI is used in content recommendation systems like Netflix or Spotify.

6️⃣ Ethical Considerations in AI

As AI technologies grow, there are ethical concerns regarding their use, including:

• Bias in AI: If training data is biased, AI models can make biased decisions, leading to unfair outcomes.

• Privacy: AI systems that use personal data must ensure privacy and comply with data protection laws.

• Job Displacement: AI automation may replace certain jobs, raising questions about the future of work and its impact on employment.

Conclusion

AI is a vast and multifaceted field with the potential to revolutionize various industries and improve everyday life. The fundamental concepts of AI—machine learning, deep learning, natural language processing, and more—are shaping the way we interact with technology. Understanding these basics provides a solid foundation for anyone interested in learning more about how AI works and its impact on society.

​

We believe that most people struggle with AI primarily because they don't understand BI or don't think about BI

​But before, lets dive into what is AI 

What to expect in this course, real life examples and a business side of AI and opportunities 

A lot of parallels in AD is based off of biological species way of processing the world 

Problem with understanding AI

This is a challenging fit primarily because of the way AI is 

MIT's effort to incorporate AI in every facet of its institution (embed youtube video here from MIT talk)  

Data Processing in AI

Data processing is a critical step in artificial intelligence (AI), as AI systems rely on large volumes of data to learn, make predictions, and perform tasks. The process involves several key stages to transform raw data into meaningful insights that AI models can use effectively.

1️⃣ Stages of Data Processing in AI

1. Data Collection

🔹 Gathering raw data from various sources (e.g., sensors, databases, websites, IoT devices, user interactions).
🔹 Examples: Social media feeds, medical records, satellite images, financial transactions.

2. Data Cleaning & Preprocessing

🔹 Removing noise, duplicates, and missing values.
🔹 Handling inconsistent or incorrect data.
🔹 Normalizing and standardizing data (e.g., scaling numerical values).
🔹 Example: In a customer database, filling in missing age values using averages.

3. Data Transformation & Feature Engineering

🔹 Converting raw data into useful formats.
🔹 Extracting important features (e.g., converting text to numerical vectors for NLP).
🔹 Reducing dimensionality to improve efficiency (e.g., PCA, autoencoders).
🔹 Example: Transforming timestamps into time-of-day categories for sales prediction.

4. Data Splitting (Training, Validation, Testing)

🔹 Dividing data into sets to train and evaluate AI models.
🔹 Typical split: 70% Training / 15% Validation / 15% Testing.
🔹 Example: A facial recognition model learns on 70% of images and is tested on the remaining 30%.

5. Data Storage & Management

🔹 Organizing data in databases, data lakes, or cloud storage for scalability.
🔹 Using distributed storage systems for big data (e.g., Hadoop, AWS S3).
🔹 Example: Storing healthcare records securely for machine learning analysis.

6. Data Processing for Model Training

🔹 Feeding structured data into AI models (e.g., deep learning networks, decision trees).
🔹 Running algorithms to learn patterns and relationships.
🔹 Optimizing models using gradient descent, backpropagation, and hyperparameter tuning.
🔹 Example: Training an AI to detect fraudulent credit card transactions based on spending behavior.

7. Data Inference & Decision Making

🔹 Using the trained model to make real-time predictions.
🔹 Deploying AI systems in real-world applications.
🔹 Example: AI-powered chatbots providing instant customer support.

2️⃣ Data Processing Techniques in AI

🔹 Batch Processing – Processing large amounts of data in chunks (e.g., analyzing logs overnight).
🔹 Real-Time Processing – Processing data as it arrives (e.g., fraud detection in banking).
🔹 Stream Processing – Handling continuous data flows (e.g., monitoring social media trends).

3️⃣ Tools & Technologies for Data Processing in AI

🔹 Big Data Tools – Hadoop, Spark, Apache Flink
🔹 Databases – SQL, NoSQL (MongoDB, Firebase)
🔹 Data Cleaning Tools – Pandas (Python), OpenRefine
🔹 Cloud Platforms – AWS, Google Cloud, Azure
🔹 AI & ML Frameworks – TensorFlow, PyTorch, Scikit-learn

Conclusion

Efficient data processing is the foundation of successful AI models. Without clean, structured, and well-managed data, even the most advanced AI algorithms cannot perform accurately. As AI continues to evolve, faster, more intelligent data processing techniques will drive innovation in fields like healthcare, finance, and autonomous systems.

Would you like more details on any stage of data processing? 🚀

​

Key Aspects of AI in Terms of Data, Data Processing, and Storage

Artificial Intelligence (AI) is data-driven, meaning its success heavily depends on how data is collected, processed, and stored. These three aspects—data, data processing, and storage—form the foundation of AI systems, enabling them to learn, make decisions, and improve over time.

1️⃣ Data in AI

Types of Data in AI

AI relies on vast amounts of data, which can be categorized into:

🔹 Structured Data – Organized data in databases (e.g., tables with rows and columns, customer records).
🔹 Unstructured Data – Raw data without a predefined format (e.g., images, videos, emails, social media posts).
🔹 Semi-Structured Data – A mix of structured and unstructured data (e.g., JSON, XML files).

Sources of Data in AI

🔹 Sensors & IoT Devices – Real-time data from smart devices.
🔹 Databases & Data Warehouses – Stored historical data for analysis.
🔹 Web & Social Media – User-generated content like tweets, posts, and blogs.
🔹 Enterprise Systems – Business transactions, customer interactions, and supply chain data.

Challenges with AI Data

❌ Data Quality Issues – Incomplete, inconsistent, or biased data can affect AI accuracy.
❌ Data Privacy & Security – AI must comply with regulations (e.g., GDPR, CCPA) to protect personal information.
❌ Data Labeling – Supervised learning requires labeled data, which can be time-consuming and expensive.

2️⃣ Data Processing in AI

Steps in AI Data Processing

1️⃣ Data Collection – Gathering raw data from multiple sources.
2️⃣ Data Cleaning & Preprocessing – Removing noise, handling missing values, and standardizing data.
3️⃣ Data Transformation – Converting raw data into a structured format, feature extraction, and encoding.
4️⃣ Data Splitting – Dividing data into training, validation, and test sets.
5️⃣ Model Training & Learning – AI learns from the processed data using algorithms (e.g., neural networks).
6️⃣ Inference & Decision Making – AI applies learned patterns to new data for predictions.

Data Processing Techniques

🔹 Batch Processing – Large volumes of data are processed periodically (e.g., financial reporting).
🔹 Real-Time Processing – AI processes data as it arrives (e.g., fraud detection in banking).
🔹 Stream Processing – Continuous data flow processing (e.g., social media monitoring).

Tools for AI Data Processing

🔹 Big Data Frameworks – Apache Spark, Hadoop.
🔹 Machine Learning Libraries – TensorFlow, PyTorch, Scikit-learn.
🔹 Data Cleaning & Transformation – Pandas, OpenRefine.

3️⃣ Data Storage in AI

Types of Data Storage for AI

🔹 Databases – SQL (e.g., MySQL, PostgreSQL) and NoSQL (e.g., MongoDB, Firebase).
🔹 Data Lakes – Store vast amounts of structured & unstructured data (e.g., Amazon S3, Azure Data Lake).
🔹 Cloud Storage – Scalable and accessible storage for AI workloads (e.g., Google Cloud Storage, AWS S3).
🔹 Distributed Storage – AI requires high-speed, scalable storage systems like Hadoop Distributed File System (HDFS).

Storage Challenges in AI

❌ Scalability – Large AI models require massive storage solutions.
❌ Latency – High-speed access to data is crucial for real-time AI applications.
❌ Security – Protecting AI data from cyber threats and unauthorized access.

Conclusion

AI depends on high-quality data, efficient data processing, and scalable storage solutions to function effectively. Optimizing these aspects ensures AI models perform accurately, securely, and efficiently in real-world applications.

Would you like a deeper dive into any specific area? 🚀

​

In AI, processing generally comes before storage, but they are closely interconnected in a cycle of data flow and system functionality. Here's how it works:

1️⃣ Data Processing Comes First

Data processing involves cleaning, transforming, analyzing, and organizing data before it can be stored or used in an AI system. Without this preprocessing and transformation, data cannot be effectively used by AI models.

Why Processing Comes First:

• Cleaning & Filtering: Raw data may be noisy, incomplete, or irrelevant, and needs to be cleaned before storage.

• Feature Engineering: Processing data into a format that AI models can use (e.g., normalizing values, extracting key features).

• Real-Time Processing: For real-time AI applications (e.g., fraud detection), data is processed first, often in memory, before being stored.

Once processed, data is either temporarily stored (e.g., in cache, RAM, or real-time buffers) or stored in databases or data lakes for future use.

2️⃣ Data Storage Comes After Processing

Data storage typically occurs after processing, but it may involve storing intermediate processed data or final outputs from AI systems.

Why Storage Happens After Processing:

• Efficient Organization: Processed data is organized in a structured way that allows for easy querying and retrieval.

• Reduced Redundancy: Storing only clean and processed data avoids redundancy and optimizes storage capacity.

• Long-Term Storage: Once data is processed, it is stored in databases, cloud storage, or data lakes for future use, sharing, or further analysis.

Flow Summary:

1. Data Collection → 2. Data Processing (Cleaning, Transforming, Analyzing) → 3. Data Storage (Databases, Data Lakes, Cloud)

However, in some real-time AI systems, both processing and storage may happen in parallel (e.g., buffering real-time data before it's processed and stored).

So, while processing typically happens first to prepare data for analysis, both aspects are intertwined in a continuous data cycle.

​

​

In the Flow Summary you provided, Machine Learning (ML) and Deep Learning (DL) come into play after Data Processing and before Data Storage, depending on the nature of the task. Here's a detailed breakdown of where ML and DL fit into the workflow:

1️⃣ Data Collection

• Data is gathered from various sources like sensors, databases, websites, etc.

2️⃣ Data Processing (Cleaning, Transforming, Analyzing)

• Data Cleaning: Removes noise, missing values, or irrelevant data.

• Data Transformation: Converts data into a structured format suitable for analysis (e.g., feature extraction).

• Data Analysis: Identifies patterns and relationships in the data that are useful for modeling.

3️⃣ Machine Learning and Deep Learning

After Data Processing, the next step is to apply Machine Learning (ML) or Deep Learning (DL) algorithms. This is where the AI model is trained and learns from the processed data.

• Machine Learning: ML is a subset of AI that involves training algorithms on processed data to make predictions or decisions without being explicitly programmed. Common techniques include:

  • Supervised Learning (e.g., classification, regression).

  • Unsupervised Learning (e.g., clustering, anomaly detection).

  • Reinforcement Learning (e.g., training an agent through rewards and punishments).

• Deep Learning: A subset of ML, deep learning uses neural networks with many layers (hence "deep") to analyze large amounts of data. It is especially effective for unstructured data like images, audio, and text. DL models, like Convolutional Neural Networks (CNNs) for image recognition or Recurrent Neural Networks (RNNs) for time-series data, learn features automatically from the data.

ML/DL Tasks:

• Model Training: This is where the AI learns from the data. Models are trained using training data and validated using validation data.

• Model Evaluation: After training, the model's performance is evaluated using test data to ensure it generalizes well to unseen data.

4️⃣ Data Storage

Once the machine learning or deep learning model is trained and evaluated, the trained model and the processed data may be stored for future use or deployment.

• Trained Models: Stored for use in production environments (e.g., cloud storage, model repositories).

• Processed Data: Depending on the application, the processed data used to train the model may be stored for future analysis or updates.

Flow Summary Including ML/DL:

1. Data Collection → 2. Data Processing (Cleaning, Transforming, Analyzing) → 3. Machine Learning / Deep Learning (Model Training, Evaluation) → 4. Data Storage (Databases, Model Repositories, Cloud Storage)

Conclusion

• Machine Learning and Deep Learning come after data processing as part of the step where the system "learns" from the data.

• The learned models are stored for future use, and the performance of the models can be continually improved as new data is collected, leading to an ongoing cycle of data processing, learning, and model storage.

Let me know if you'd like more details on any specific part of this flow!