ISO 9241-11 Definition of Usability

Signal detection is the process of identifying potential causal relationships between a drug and an adverse event from large datasets, typically spontaneous reporting systems. It uses statistical methods, known as disproportionality analysis, to find drug-event combinations reported more frequently than expected. A common measure is the Reporting Odds Ratio (ROR), a value greater than one suggesting a potential signal that requires further investigation.

Jakob Nielsen, a prominent usability consultant in UI and webdesign mainly, defined usability through five quality components: Learnability (how easy is it for users to accomplish basic tasks the first time?), Efficiency (how quickly can they perform tasks once learned?), Memorability (can users reestablish proficiency after a period of not using it?), Errors (how many errors do users make?), and Satisfaction (how pleasant is it to use?).

In fair machine learning, impossibility theorems demonstrate that it is mathematically impossible for an algorithm to satisfy multiple, seemingly intuitive fairness criteria simultaneously, except in trivial cases. For example, an algorithm cannot generally satisfy both demographic parity (equal positive rates across groups) and equalized odds (equal true positive and false positive rates across groups) if the base rates differ between groups.

RAID (acronym of Redundant Array of Independent Disks) is a data storage virtualization technology that combines multiple physical disk drives into one or more logical units for data redundancy, performance improvement, or both. Different RAID levels provide varying balances of reliability, availability, performance, and capacity. For example, RAID 1 mirrors data across two disks, while RAID 5 stripes data and parity across at least three disks.

Agile project management is an iterative approach to delivering a project throughout its life cycle. It breaks down large projects into smaller, manageable tasks completed in short iterations or 'sprints.' This allows for frequent reassessment, adaptation of plans, and flexibility in response to change. It prioritizes customer collaboration, working software, and responding to change over comprehensive documentation and rigid plans.

Algorithmic bias mitigation techniques are categorized into three main stages relative to the model training process. Pre-processing methods modify the training data itself (e.g., reweighing, resampling). In-processing methods incorporate fairness constraints directly into the model's learning algorithm. Post-processing methods adjust the model's predictions after they have been made to improve fairness.

Algorithmic confounding occurs when a proxy variable used by an algorithm is correlated with a protected attribute (like race or gender) and also with the outcome of interest. The algorithm may inadvertently learn to discriminate based on the protected attribute by using the proxy, even if the protected attribute itself is explicitly excluded from the model's input data.