EXPLORE / 06
Terms index
Specialized AI and software-development terms with concise explanations
Entity records
4516 records · 37/113
- 1441
Multi-source Log Correlation
A data analysis technique that links discrete log events from various systems (auth, network, apps) using timestamps, user IDs,
Website pending - 1442
Computer Worm
A type of malicious software that spreads automatically over networks, runs independently without host programs, and consumes sy
Website pending - 1443
AI-driven Materials Discovery
Using ML models to predict molecular properties or generate candidate structures, significantly shortening the cycle from theore
Website pending - 1444
Model Catalog
A centralized metadata repository for managing pretrained ML models, supporting versioning, search, and direct deployment invoca
Website pending - 1445
Neuron Acceleration
Hardware acceleration based on AWS Trainium/Inferentia chips, mapping DL models to custom silicon via compilers to boost inferen
Website pending - 1446
Sequence Extrapolation
A technique inferring future elements or states based on historical patterns and trends within a known data sequence.
Website pending - 1447
Finite State Machine (FSM)
A computational model consisting of finite states and transition rules that switches between states based on inputs to execute s
Website pending - 1448
SIMD Instructions
Single Instruction Multiple Data architecture allowing CPUs to perform identical operations on multiple data elements per cycle,
Website pending - 1449
RaBitQ Quantization
A stochastic binary quantization method for high-dimensional vectors that balances storage compression and retrieval accuracy vi
Website pending - 1450
Approximate Nearest Neighbor (ANN)
A technique for rapidly finding similar points in high-dimensional space by trading slight accuracy loss for significant speed g
Website pending - 1451
PIPE Financing
Private Investment in Public Equity, used by listed companies to raise capital quickly, often involving share dilution or specif
Website pending - 1452
Authentication Vulnerability
A flaw or misconfiguration in the identity verification process that allows unauthorized users to bypass security mechanisms and
Website pending - 1453
Permission Boundaries
Technical policies defining the scope of actions an AI agent can execute, limiting resource access to reduce risks from autonomo
Website pending - 1454
RAD
A methodology accelerating software delivery through iterative prototyping and visual component reuse rather than linear waterfa
Website pending - 1455
Generative Agent
An autonomous system that plans steps, invokes tools, and generates code or content based on natural language instructions.
Website pending - 1456
Output Token
The smallest unit of text generated by an LLM, commonly used to measure actual content volume and computational resource consump
Website pending - 1457
Multi-view Geometric Constraints
Leveraging correspondences across different viewpoints and camera parameters to constrain object scale, position, and shape in 3
Website pending - 1458
Physics Engine Validation
Using physics engines to simulate motion, collision, and contact during 3D reconstruction to ensure virtual scenes are both visu
Website pending - 1459
Always-On Visual Perception
A technical state where devices continuously capture and process visual data via sensors to understand environmental context in
Website pending - 1460
SLAM
Technology enabling mobile robots or cameras to simultaneously construct maps and determine their position in unknown environmen
Website pending - 1461
World Coordinate System
A fixed global reference frame used to unify local observations from different timestamps or viewpoints (e.g., camera coordinate
Website pending - 1462
MANO Model
A low-dimensional statistical model based on linear blend skinning that describes hand pose, shape, and orientation with few par
Website pending - 1463
Balanced Accuracy
The arithmetic mean of recall per class, used to address class imbalance and better reflect model performance on minority classe
Website pending - 1464
3D Stacked DRAM
Technology vertically stacking DRAM layers directly above logic chips like GPUs/NPUs to shorten data paths, thereby increasing b
Website pending - 1465
ROI Optimization
A strategy or algorithmic process that analyzes return-on-investment data to automatically adjust resource allocation for maximu
Website pending - 1466
Degrees of Freedom (DoF)
The number of independent parameters defining a mechanical structure's configuration; higher DoF in hands enables complex joint
Website pending - 1467
Pyodide
A distribution of CPython compiled to WebAssembly, allowing Python and its scientific libraries to run directly in browsers or N
Website pending - 1468
High-Speed Switch
Core data center networking device connecting large-scale GPU clusters, relying on high-performance PCBs to handle massive data
Website pending - 1469
Printed Circuit Board (PCB)
A board that mechanically supports and electrically connects components; AI servers require high-layer, low-loss PCBs for high-s
Website pending - 1470
Compute-in-Memory (CIM)
Embedding computing logic directly into memory arrays or adjacent cells to reduce data movement overhead and break the von Neuma
Website pending - 1471
Resistive Random Access Memory (RRAM)
A non-volatile memory that stores data via resistance state changes, offering high density and low power potential for compute-i
Website pending - 1472
Whisper
A Transformer-based automatic speech recognition model developed by OpenAI, capable of multilingual transcription and translatio
Website pending - 1473
gRPC
A high-performance open-source RPC framework based on HTTP/2 using Protocol Buffers, supporting bidirectional streaming for low-
Website pending - 1474
Algorithmic Auditing
The systematic process of testing and evaluating automated decision systems for discrimination, bias, or ethical violations, oft
Website pending - 1475
Active Noise Cancellation
Algorithmic suppression of background noise to enhance target voice clarity, improving speech capture quality in noisy environme
Website pending - 1476
System Throughput
Total tokens processed by the entire inference system per unit time, reflecting overall capacity under high concurrency.
Website pending - 1477
Agentic Inference
An inference paradigm optimized for autonomous agents executing multi-step tasks, tool calls, and long-context interactions.
Website pending - 1478
AIOps
An operations paradigm using ML algorithms to analyze logs, metrics, and traces for fault prediction, root cause analysis, and a
Website pending - 1479
Voltage Regulator Module (VRM)
Power conversion component providing stable low-voltage high-current supply to processors, impacting chip efficiency and stabili
Website pending - 1480
System-in-Package (SiP)
Integrating multiple ICs or passive components into a single package to reduce footprint and enhance performance.
Website pending