How to boil an egg
Today I asked ChatGPT to explain to a robot how to boil an egg. After a brief pause, the Natural Language Processing model replied.
Here is a step-by-step explanation for a robot on how to boil an egg:
- Fill a saucepan with enough water to cover the eggs by about one inch.
- Place the saucepan on a stove and turn on the heat to high.
- Use a temperature sensor to monitor the water temperature.
- Once the water reaches a boiling point (212°F or 100°C), carefully place the eggs in the boiling water using a mechanical arm or gripper.
- Set a timer for 4-9 minutes depending on the desired level of doneness (soft-boiled 4-5 minutes, medium-boiled 6-7 minutes, hard-boiled 8-9 minutes).
- Once the timer goes off, use a mechanical arm or gripper to remove the eggs from the boiling water and place them into a bowl filled with cold water.
- Use a timer to let the eggs sit in the cold water for a few minutes, then remove the eggs from the water.
- Gently crack the shells and peel them off, starting from the large end.
Note: It is important for the robot to have a mechanism for handling hot objects, such as heat-resistant gloves or tongs, to ensure the safety of the eggs and the robot.
Impressive as it certainly sounds, the description above is still too vague. Suppose we had a robot sophisticated enough to handle the operations described before. Will it be able to boil an egg? As humans, we understand perfectly the context and materials involved in the process. But what exactly is an egg? We have no doubt that a chicken’s egg will do the job, but so would a duck’s or goose’s egg. What about turtle’s eggs? They are indeed collected and eaten in certain parts of the world. Also, how does the stove operate? Depending on whether it uses gas, electricity or wood, there are additional constraints on the material the saucepan is made of. For instance, one made of aluminium won’t work on a modern induction stovetop.
These may sound like unnecessary complications to an otherwise simple and straightforward process. But even if these robots are not yet among us, something similar already exists. They are swarms of programs sieving the internet for content and meaning. We call them by the somewhat sinister name of “agents”. These agents interact with the Semantic Web, a technology that represents information in a manner that makes it easier for computers to understand and process.
One of the most apparent and defining services built on top of semantic web technologies is recommendation systems, such as those used by Netflix or Amazon, which make accurate recommendations based on the relationships between users, items, and context.
Ontologies are a crucial component of the semantic web, providing a common vocabulary and a shared understanding of concepts and relationships in a particular knowledge domain. In the semantic web context, an ontology is a machine-readable representation of knowledge expressed as a set of classes (or concepts) with relations operating between them. Ontologies have been used to unify the representation of gene and gene product attributes in molecular biology (Gene Ontology / GO); to describe things that are of interest in financial business applications and the ways that those things can relate to one another (Financial Industry Business Ontology / FIBO); or to enable interoperability between devices from different providers and among various activity sectors in the Internet of Things (Smart Applications REFerence ontology / SAREF), just to name a few examples.
In the field of applied sciences, the Elementary Multiperspective Material Ontology (EMMO) is a foundation (or top-level) ontology providing a common starting point for defining domain-specific and application ontologies. The EMMO is based on analytical philosophy and scientific principles. In particular, real world objects are represented in EMMO by different ways (perspectives) of understanding them. Perspectives are an expression of reductionism (i.e. objects are made of sub-objects) and epistemological pluralism (i.e. objects are always defined according to the perspective of an interpreter). Furthermore, the way in which objects relate to each other is founded on principles of mereology (parthood) and causality.
To demonstrate the flexibility and expressivity of the EMMO, let us try to describe the process of boiling an egg using two different perspectives. From a strict process engineering point of view, it makes sense to use the Reductionistic perspective to decompose the process of boiling an egg (the “whole”) into steps that are causally connected in space and time. The process can be expressed as a workflow of causally-connected events (tiles), thus defining a beginning, the intermediate steps, and an end. At the same time, properties can be attached to each of the objects by using the Semiotic perspective. That way, the process can be fully characterised in terms of any known or observed physical quantities such as the mass of the objects, their temperature throughout the process, the inertia moment of the egg before and after being boiled, etc. To describe the materials involved, the Physicalistic perspective allows looking at the egg as a complex material. The egg is encased in a solid crust made of an inorganic salt embedded in an organic matrix. Boiling the egg then transforms the microscopic structure of the proteins in the egg’s white and yolk, turning them from a gel to a solid. These descriptions fulfil different purposes and are by no means the only possible ones. For example, another (Holistic) perspective would be to consider the egg’s role as food.
The power of ontologies resides in their ability to capture different requirements and levels of detail. On top of that, EMMO offers foundational rules based on physical principles and aims to provide a unified framework that supports the pluralisms of our observations and understanding and enables interoperability between the many different applications of material science.
Goldbeck Consulting is part of the development team of the EMMO. It is working on European projects ranging from connecting data sources to manufacturing processes to describing and deploying materials modelling workflows into open platforms.