44 of the Computer Science glossary’s terms, each with the part of the syllabus where you first meet it.
A link between a parent node and one of its children.
Processing data on or near the device that produces it, at the "edge" of the network, in place of sending it all to a distant cloud data centre.
A device at the edge of a network, close to where data is produced, such as a phone, camera, sensor or car, that runs machine learning models locally instead of sending data to a remote server.
A way of choosing the number of clusters k by plotting the within-cluster sum of squares against k and picking the value at the "elbow", where further increases in k give only small reductions.
Unwanted electrical noise induced in a transmission medium by nearby electrical equipment or radio signals, which can corrupt the data being carried.
Discarded electrical and electronic equipment, such as computers, phones and servers, which contains valuable and toxic materials and is often exported to countries with weaker recycling controls.
A single value stored in an array or list, identified by its index.
A selection structure that tests a series of conditions in order and runs only the block belonging to the first condition that is true, with an optional final else block for when none is true.
Scanning incoming email for spam, phishing and malware, and blocking, quarantining or labelling suspicious messages before they reach users.
A feature selection strategy in which the learning algorithm itself selects or down-weights features while the model is being trained.
Flash storage packaged with its controller in a single chip that is soldered permanently onto a device's circuit board.
A computer system with a dedicated function that is built into a larger mechanical or electronic device, usually with limited memory, processing power and user interface.
The bundling of an object's data and the methods that operate on that data into a single unit, the class.
The process in which each layer of the TCP/IP model, on the sending host, wraps the data from the layer above with its own header (and sometimes a trailer), with the reverse process, de-encapsulation, removing each layer's header on the receiving host.
A network protocol that encrypts the data it carries and authenticates the other party, so that intercepted traffic cannot be read or altered undetected.
The process of converting readable data into an unreadable form using an algorithm and a key, so that only someone with the correct key can convert it back (decryption).
A secret value, usually a long string of bits, that controls how an encryption algorithm transforms plaintext into ciphertext and back.
The condition reached when the file pointer has passed the last character of a file, so that no data is left to read.
The queue operation that adds a new item at the rear of a queue.
A model that combines the predictions of several different models in order to predict more accurately than any of them alone.
A person, object, place, event or concept about which data is stored.
The rule that every record has a primary key value that is not null and is unique within its table, so that every record can be identified.
A diagram that models the data of a system as entities, the attributes of each entity and the relationships between entities, including how many instances of each entity may take part in each relationship.
Everything outside the agent that it interacts with: it receives the agent's actions and responds with a new situation and a reward.
The effect of a technology on the natural environment over its life cycle, including the energy and water used to make and run it, the greenhouse gases emitted and the waste produced when it is discarded.
One complete pass of the whole training data set through a learning algorithm during training.
An action-selection rule in which an agent explores by choosing a random action with probability ε and otherwise exploits by choosing the action with the highest estimated value.
Fairness in how the benefits, costs and risks of a technology are distributed, taking into account that different groups start from different positions and may need different support to benefit equally.
Invalid test data, of the wrong type or clearly outside the valid range, that a program should reject without crashing.
The difference between the set point and the measured value of the controlled quantity, which the controller uses to decide its next output.
The family of standards (IEEE 802.3) for cabled local area networks, which defines the cables, connectors, signals and frame format that devices use to exchange data.
The study of what is right and wrong in human conduct, applied in computing to the choices made when designing, building and using technology and to their effects on people and the environment.
The straight-line distance between two points, the square root of the sum of the squared differences between their coordinates.
A guarantee that if no new updates are made to a data item, all copies of it will become identical after some time, although reads in the meantime may return an older value.
An event during the running of a program that disrupts the normal flow of instructions, represented by an object that records the type of error and a message.
A technique in which code that might fail is placed in a block that is monitored for exceptions, and handler code runs if a matching exception occurs, so the program can recover instead of stopping.
A file of machine code produced by translation that the operating system can load and run directly on one platform, without the source code or the translator.
The degree to which the reasons for a model's individual outputs can be described in terms a person can understand, such as which inputs most influenced a decision.
In reinforcement learning, choosing the action with the highest estimated reward based on what the agent has learned so far.
In reinforcement learning, choosing an action that is not currently believed to be the best, in order to gather information about its reward.
A complexity in which the number of steps roughly doubles each time the input size increases by one.
The extent to which a product can be extended with new features or adapted to new requirements without major changes to its existing design and code.
The process that copies data from source systems (extract), cleans and converts it into one consistent format (transform), and adds it to the data warehouse (load).
Using a model to predict values outside the range of the data it was fitted to, as opposed to interpolation, which predicts within that range.