Demystifying AI's Buzzwords

The world of AI has become incredibly murky with many different buzzwords used interchangeably. AI, Machine Learning, Generative AI, Robotics, and even Data Science are some of the more commonly mentioned ones.

Let’s look at some of these a bit more closely to differentiate them.

Artificial Intelligence (AI)

"AI" stands for Artificial Intelligence and is a subfield of Computer Science. AI refers to a range of capabilities that grant machines the ability to learn, adapt, and perform tasks that typically require human intelligence. True AI needs to exhibit two characteristics: the ability to learn from experience and the capability to perform tasks in complex environments without constant guidance from a user.

AI encompasses various techniques, with one of the most well-known being Machine Learning. Additionally, other techniques such as Computer Vision, Deep Learning, and Natural Language Processing play significant roles.

For instance, consider the familiar self-service checkout at your local grocery store. Currently, these systems rely on barcode scanning and weight recognition to identify items. Advanced computer vision technology could be implemented to further enhance the checkout process.

When you place an item on the self-checkout kiosk's scanning area, a camera captures an image of the item. Computer vision algorithms then analyse the image to identify the item based on its shape, size, and visual features. This information is then matched with a database containing product details and prices.

Furthermore, such a system can automatically count the items you've placed in the scanner area and adjust the total accordingly. Currently, this is done manually and relies on trust. Computer vision also has the potential to aid in theft prevention by detecting instances where items are placed in the bagging area without being scanned. It can compare the scanned items with what's observed in the bagging area to ensure that everything is accounted for. As for the current approach, the system alerts a weight discrepancy and generates an error. This necessitates calling the checkout staff for resolution.

Moreover, computer vision could monitor your interactions at the kiosk and detect when you require assistance, subsequently notifying the staff.

Machine Learning (ML)

Machine Learning is a specific technique within AI, making it a subset of the broader umbrella term, AI. It involves a method for computers to learn from data and enhance their performance over time.

For example, imagine you're using a language translation app on your phone. The app wants to accurately translate sentences from one language to another. Instead of programmers writing explicit rules for every possible translation, the app uses Machine Learning.

This involves a training phase, feeding the app with lots of sentence pairs, where each sentence is given in both the original language and the target language.

The ML algorithm then learns patterns and relationships in these sentence pairs and becomes better at predicting translations as it starts to recognise context, idioms, and nuances.

When you then type a sentence in your language, the ML model uses what it learned to predict the translation in real-time, making educated guesses based on patterns it recognized during training.

Artificial Generative AI (AGI)

The new buzzword in town. Similar to Machine Learning, AGI is another subset of the broader field of AI. AGI focuses on creating new content or data that is similar to what it has learned from existing examples.

In more commonly understood AI, systems are designed to follow predefined rules or patterns to achieve specific goals. For instance, an AI program might be trained to recognise cats in images by analysing thousands of labeled cat pictures and learning the distinctive features of a cat.

On the other hand, Artificial Generative AI goes beyond recognition and decision-making. It involves training models to generate new content that resembles something created by humans, like art, music, text, images, and even entire scenarios. This type of AI can be creative in nature and produce novel outputs that weren't explicitly programmed into it.

For example, Midjourney is an interesting example of how AI can be used to bridge the gap between text and visual content, creating unique and original images based on human-provided prompts.

Data Science

Data Science is a field in its own right that involves extracting knowledge and insights from data. It combines elements of statistics, computer science, and domain expertise to collect, process, analyse, and interpret large and complex datasets. Data Scientists use various tools and techniques to uncover patterns, trends, and valuable information hidden within data, helping organizations make informed decisions and predictions. They develop models, create visualisations, and apply advanced analytics to solve problems and discover actionable insights that drive business strategies, scientific research, and innovation.

The crossover between Data Science and AI lies in the fact that both fields heavily rely on each other to achieve their goals. Data Science provides the foundation for AI by supplying the necessary data for training and validation of AI models. AI, on the other hand, enhances Data Science by creating more advanced algorithms that can automatically learn from data to make predictions, recognise patterns, and take actions. To simplify, Data Science provides the training data many AI models are based on.

As an example of the interplay of Data Science and AI Data Collection: The hospital gathers data from electronic health records, including patients' medical history, lab results, age, weight, and other relevant factors. Data Scientists analyse this data to identify patterns and risk factors. They might discover that patients with a family history of diabetes, higher BMI, and certain lab results are more likely to develop the condition. Using this information, Data Scientists create a predictive model. This model uses machine learning algorithms to predict the likelihood of a patient developing diabetes based on their individual data.

The hospital can then use this model to identify patients who are at a higher risk of diabetes. With early intervention, patients at risk can make lifestyle changes or receive appropriate medical attention, potentially preventing the onset of diabetes or managing the condition more effectively.

Robotics

Robotics is the field of technology and engineering that focuses on designing, building, and programming robots. These are autonomous or semi-autonomous machines that can perform tasks in the physical world. Robots are typically equipped with sensors, actuators, and a central processing unit, allowing them to sense their environment, make decisions, and interact with objects or surroundings. They find applications in various industries, from manufacturing and healthcare to space exploration and household chores, where they can automate processes, assist humans, and perform tasks that may be dangerous or repetitive.

Here's an intriguing thought experiment when contemplating the realms of AI and robotics.

Consider a task that humans would typically deem simple, such as walking to a door and opening it. Now contrast this with a different task, such as engaging in a chess competition. While we might perceive the latter as incredibly challenging, the former seems merely mundane.

However, AI flips this perspective on its head. Programming a computer to achieve victory in a chess competition is comparatively straightforward, as chess operates within a rule-based system with well-defined parameters. Yet, programming a robot to approach a door, extend its reach to the doorknob, and then turn it becomes a remarkably intricate endeavour.

In a way, robotics is the ultimate challenge of AI since it requires a combination of virtually all areas of AI.

~~~

Through all these terms, we've seen the interconnectedness that threads them together, forming a rich tapestry of technological advancement. As we contemplate AI's journey, from rule-based strategies to generative creativity, from learning patterns to generating new outcomes, and from data extraction to real-world physical interaction, one thing becomes clear: AI's potential knows no bounds. In every corner of these terms, we find AI's transformative power reshaping industries, enhancing experiences, and pushing the boundaries of what's achievable.