Punjab PSEB SS Computer Application flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Binary number system

A base-2 positional number system using only digits 0 and 1, where each place value is a power of 2, widely used because digital circuits have two stable states.

Two's complement

A signed binary representation formed by inverting all bits of a number and adding 1, allowing subtraction to be performed as addition and giving one unique zero.

Universal logic gates

NAND and NOR gates, because any Boolean function can be built using only NAND gates or only NOR gates, making them the basis of digital circuit design.

De Morgan's laws

Two Boolean identities: NOT(A AND B) equals (NOT A) OR (NOT B), and NOT(A OR B) equals (NOT A) AND (NOT B), used to simplify logic expressions.

Cache memory

A small, fast memory between the CPU and main memory that stores recently used data and instructions, reducing average access time compared with fetching from RAM.

Primary versus secondary storage

Primary storage such as RAM is volatile and directly accessed by the CPU; secondary storage such as HDD or SSD is non-volatile and retains data when power is off.

TCP versus UDP

TCP is connection-oriented, ordered and reliable with acknowledgements and retransmission; UDP is connectionless and faster but does not guarantee delivery or ordering.

SQL DDL and DML

DDL statements such as CREATE, ALTER and DROP define or change schema structure; DML statements such as INSERT, UPDATE, DELETE and SELECT manipulate the data inside tables.

Operating system functions

An operating system manages hardware resources, schedules processes, allocates memory, handles file systems and devices, and provides a user interface for running applications.

Artificial intelligence definition

The branch of computer science that builds systems performing tasks associated with human intelligence, such as learning, reasoning, perception, language understanding and decision making.

Supervised learning

A machine learning approach that trains a model on labelled input-output pairs so it can predict the correct output for new, previously unseen inputs.

Unsupervised learning

A machine learning approach that finds patterns, groupings or structure in unlabelled data, with clustering and dimensionality reduction as typical tasks.

Reinforcement learning

A learning paradigm where an agent takes actions in an environment and learns a policy that maximises cumulative reward through trial, error and feedback signals.

Classification versus regression

Classification predicts a discrete category such as spam or not spam; regression predicts a continuous numeric value such as price or temperature.

Overfitting

A modelling failure where a model learns training data noise and performs very well on training data but poorly on new, unseen test data.

Train-validation-test split

Data is divided so the training set fits parameters, the validation set tunes hyperparameters, and the test set provides a final unbiased performance estimate.

Confusion matrix

A table of true positives, true negatives, false positives and false negatives that summarises a classifier's predictions and supports accuracy, precision and recall calculation.

Precision and recall

Precision is true positives divided by predicted positives; recall is true positives divided by actual positives; F1 score is their harmonic mean.

Natural language processing

The AI subfield concerned with enabling computers to understand, interpret and generate human language, including tasks like translation, sentiment analysis and chatbots.

Turing Test

A test proposed by Alan Turing in which a human judge converses with a machine and a human; the machine passes if the judge cannot reliably identify it.

Expert system

An AI program that encodes human specialist knowledge as rules in a knowledge base and uses an inference engine to give advice or diagnoses.

Backpropagation

A training algorithm for neural networks that applies the chain rule to compute the gradient of the loss with respect to each weight, working backward from output to input.

Gradient descent

An optimisation method that repeatedly adjusts parameters in the direction that reduces the loss, using the update rule w ← w − η·∂L/∂w with learning rate η.

Convolutional neural network

A deep network using learnable filters with shared weights and pooling layers to detect local spatial features, widely applied to image recognition and computer vision.

Vanishing gradient problem

A training difficulty in deep or recurrent networks where gradients shrink toward zero during backpropagation, slowing or stopping learning in early layers.

LSTM and GRU

Gated recurrent architectures that use gate mechanisms to control information flow, allowing the network to retain relevant context over long sequences.

Self-attention

A transformer mechanism computing softmax(QKᵀ/√d_k)·V so each position in a sequence can weight information from all other positions in parallel.

Large language model

A neural network with many parameters trained on large text corpora to predict the next token, enabling text generation, translation, summarisation and question answering.

Model temperature

A sampling parameter that scales output logits; higher values increase randomness and diversity, lower values make generation more deterministic and focused.

Spreadsheet absolute reference

A cell reference written with dollar signs such as $A$1 that stays fixed when a formula is copied to another cell, unlike a relative reference.

Mail merge

A word processing feature that combines one main document with a data source to generate many personalised copies, such as letters or certificates.

Phishing

A fraudulent email, message or website that impersonates a trusted entity to trick users into revealing passwords, OTPs or financial details.