Skip to main content

Enterprise Technology Terminology: B

149 results · page 4 of 8

  • Behavior-Based Access Control

    Behavior-based access control is an access control approach that uses observed behavior and contextual activity to adjust permissions and authorization decisions in real time, helping enterprises enforce least privilege, support zero trust architectures, and improve detection of anomalous or high-risk access.

  • Behavior Tree Engine

    Behavior Tree Engine is a software runtime component that interprets and executes behavior trees as a formal model for decision-making, enabling structured, observable control logic for autonomous systems, robotics, and interactive agents in enterprise and industrial environments.

  • Belief–Desire–Intention Model

    Belief–Desire–Intention Model is a formal agent architecture that represents an agent’s beliefs, desires, and intentions to capture goal-directed, explainable decision behavior, used in enterprise multiagent systems, automation, and decision-support applications to structure goals, assumptions, and commitments in software agents.

  • Bell State

    Bell state is a maximally entangled quantum state of two qubits that forms a standard basis for describing two-particle entanglement and supports protocols such as quantum teleportation, quantum key distribution, and entanglement-based quantum networking in enterprise and research environments.

  • Bell State Distribution

    Bell state distribution is the process of creating and delivering maximally entangled two-qubit Bell states between remote endpoints over quantum channels, enabling entanglement-based quantum communication, networking, and cryptographic protocols that enterprises can incorporate into long-term secure communication and quantum network architectures.

  • Benchmark Dataset

    Benchmark dataset is a standardized collection of data, labels, and evaluation protocols used to train, test, and compare algorithms or systems under reproducible conditions, which supports objective model evaluation, procurement decisions, and governance in enterprise and technical environments.

  • Benchmark Harness

    Benchmark harness is a configurable framework that automates running and measuring performance benchmarks for systems, applications, or algorithms, enabling controlled, repeatable tests and comparable metrics that support enterprise performance engineering, capacity planning, and procurement or architecture decision-making.

  • Benchmark Suite

    Benchmark suite is a standardized collection of performance tests and workloads used by enterprises to measure and compare systems, hardware, software, or services under defined conditions, supporting objective evaluation, capacity planning, and procurement decisions in technical and architectural contexts.

  • Bent Pipe Architecture

    Bent pipe architecture is a satellite communication design where the satellite functions as a transparent relay, only amplifying and frequency-shifting signals, while all routing, security, and higher-layer processing occur in ground networks that enterprises and service providers control.

  • Beta

    Beta is a pre-release phase in the software lifecycle where near-final builds are tested by selected or public users under realistic conditions to uncover defects, validate performance and security controls, and provide structured feedback before full enterprise or market-wide deployment.

  • Bfloat16 Format

    Bfloat16 format is a 16-bit floating-point representation with the same 8-bit exponent as 32-bit floats but a shorter mantissa, used in enterprise AI and high-performance computing to reduce memory and compute cost while maintaining broad numeric range.

  • Bias and Fairness Audit

    Bias and fairness audit is a structured assessment of data, models, and decision processes to detect and quantify disparate outcomes across groups and to evaluate whether algorithmic systems align with defined fairness, legal, and governance requirements in enterprise environments.

  • Bias Audit Process

    Bias audit process is a structured methodology organizations use to test automated systems and datasets for unfair or disparate outcomes across protected groups, providing documented evidence for governance, compliance, and ongoing control of algorithmic and AI-driven decision-making.

  • Bias Detection Engine

    Bias Detection Engine is a software or algorithmic component that evaluates data sets, machine learning models, and decision workflows for statistical bias and fairness deviations, supporting enterprise governance, regulatory compliance, and formal risk management of automated and AI-supported decisions.

  • Bias Detection System

    Bias detection system is a framework of tools and processes that analyzes datasets and algorithmic models to measure group-level disparities, support fairness evaluations, and provide auditable evidence for governance, compliance, and risk management in enterprise machine learning and automated decision systems.

  • Bias Exploitation Attack

    Bias exploitation attack is a deliberate method of manipulating an AI or machine learning system by targeting its preexisting statistical or data-induced biases to trigger skewed, unfair, or policy-violating outputs, creating governance, compliance, and operational risks for enterprises.

  • Bias Mitigation

    Bias mitigation is the set of processes, techniques, and governance practices organizations use to detect, reduce, and manage unwanted bias in data, algorithms, and models, supporting regulatory compliance, documented fairness criteria, and reliable operation of enterprise AI and analytics systems.

  • Bias Mitigation Strategy

    Bias mitigation strategy is a structured approach that enterprises use to detect, measure, and reduce unwanted bias in data-driven or AI-enabled systems, supporting governance, compliance, and repeatable procedures for assessing and managing group-level disparities in model behavior and decisions.

  • Bidirectional Encoder Representations from Transformers

    Bidirectional Encoder Representations from Transformers (BERT) is a transformer-based language representation model that generates contextual text embeddings for diverse natural language processing tasks, relevant to enterprises for automating search, classification, and text understanding within analytics, customer operations, and knowledge management systems.

  • Big Data

    Big data is data collections whose size, complexity, and speed exceed the capabilities of traditional databases, requiring distributed storage and processing. It matters in enterprises because it underpins large-scale analytics, machine learning, and data-driven operations under governed, secure architectures.