📣 Big news: LightlyStudio is now live! Try it for free.

A-Z of Machine Learning and Computer Vision Terms

A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
Z
  • This is some text inside of a div block.
  • This is some text inside of a div block.
  • This is some text inside of a div block.
  • This is some text inside of a div block.
  • This is some text inside of a div block.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
A
A
Activation Function
Activation Function
Active Learning
Active Learning
AI Agent
AI Agent
AI Assistants
AI Assistants
AI Assisted Labeling
AI Assisted Labeling
Algorithm
Algorithm
Anchor Boxes
Anchor Boxes
Anomaly Detection
Anomaly Detection
Artificial Intelligence (AI)
Artificial Intelligence (AI)
Attribute
Attribute
B
B
Backpropagation
Backpropagation
Backpropagation
Backpropagation
Bagging
Bagging
Batch
Batch
Batch Normalization
Batch Normalization
Bayesian Network
Bayesian Network
Bias
Bias
Big Data
Big Data
Binary Classification
Binary Classification
Blur
Blur
Boosting
Boosting
Bounding Box
Bounding Box
C
C
Calibration
Calibration
Calibration Curve
Calibration Curve
Canonical Correlation Analysis (CCA)
Canonical Correlation Analysis (CCA)
Case-Based Reasoning
Case-Based Reasoning
Chain of Thought (CoT)
Chain of Thought (CoT)
ChatGPT
ChatGPT
Chi-Squared Automatic Interaction Detection (CHAID)
Chi-Squared Automatic Interaction Detection (CHAID)
Class Boundary (Statistics & Machine Learning)
Class Boundary (Statistics & Machine Learning)
Class Imbalance
Class Imbalance
Clustering
Clustering
COCO
COCO
Collaborative Filtering
Collaborative Filtering
Computer Vision
Computer Vision
Computer Vision Model
Computer Vision Model
Concept Drift
Concept Drift
Conditional Random Field (CRF)
Conditional Random Field (CRF)
Confusion Matrix
Confusion Matrix
Constrained Clustering
Constrained Clustering
Contrastive Learning
Contrastive Learning
Convolutional Neural Network (CNN)
Convolutional Neural Network (CNN)
Cross-entropy Loss
Cross-entropy Loss
Cross-Validation
Cross-Validation
D
D
Data Approximation
Data Approximation
Data Augmentation
Data Augmentation
Data Drift
Data Drift
Data Error
Data Error
Data Mining
Data Mining
Data Operations
Data Operations
Data Pre-processing
Data Pre-processing
Data Quality
Data Quality
Dataset
Dataset
Decision Boundary
Decision Boundary
Decision List
Decision List
Decision Stump
Decision Stump
Decision Tree
Decision Tree
Deep Learning
Deep Learning
Deep Neural Networks
Deep Neural Networks
Deep Reinforcement Learning (DRL)
Deep Reinforcement Learning (DRL)
DICOM
DICOM
Dimensionality Reduction
Dimensionality Reduction
Dimensionality Reduction
Dimensionality Reduction
Dropout
Dropout
Dynamic and Event-Based Classifications
Dynamic and Event-Based Classifications
E
E
Edge Cases
Edge Cases
Edge Computing
Edge Computing
Edge Detection
Edge Detection
Elastic Net
Elastic Net
Embedding Spaces
Embedding Spaces
Ensemble Learning
Ensemble Learning
Epoch
Epoch
Expectation-Maximization Algorithm (EM)
Expectation-Maximization Algorithm (EM)
Extreme Learning Machine
Extreme Learning Machine
F
F
F1 Score
F1 Score
Factor Analysis
Factor Analysis
False Positive Rate
False Positive Rate
Feature
Feature
Feature Extraction
Feature Extraction
Feature Hashing
Feature Hashing
Feature Learning
Feature Learning
Feature Scaling
Feature Scaling
Feature Selection
Feature Selection
Feature Vector
Feature Vector
Fisher’s Linear Discriminant
Fisher’s Linear Discriminant
Foundation Models
Foundation Models
FP-Growth Algorithm
FP-Growth Algorithm
Frame Rate
Frame Rate
Frames Per Second (FPS)
Frames Per Second (FPS)
Fully Connected Layer
Fully Connected Layer
Fuzzy Logic
Fuzzy Logic
G
G
Generative Adversarial Network (GAN)
Generative Adversarial Network (GAN)
Generative Adversarial Networks
Generative Adversarial Networks
D

Data Quality

Data quality refers to the condition of a dataset with respect to factors such as accuracy, completeness, consistency, timeliness, and validity. High-quality data correctly represents the real-world construct it is intended to model and is fit for its intended use in decision-making or model training. In ML, poor data quality (e.g., mislabelled samples, noise, bias) can degrade model performance more than poor algorithm choice. Ensuring data quality may involve processes like validation rules (to catch out-of-range or illogical values), data cleaning, deduplication, and periodic audits. When combining datasets, maintaining consistent formats and definitions (data integrity) is key. Overall, trustworthy analytics and AI systems begin with high-quality, reliable data.

Further Reading

🔗  Research Paper📄  Blog Post📄  Blog Post

Explore Our Products

LightlyStudio

Data Curation, Labeling & QA

Get data insights and find the perfect
selection strategy, label & QA data.

Learn More

LightlyTrain

Self-Supervised Pretraining & Fine-Tuning

Leverage self-supervised learning to pretrain
and fine-tune models

Learn More

Ready to Get Started?

Experience the power of automated data curation with Lightly

Learn More
Clustering
Dimensionality Reduction
Pooling Layer
Activation Function
Hidden Layer
Random Forests
K-Nearest Neighbor (KNN)
Support Vector Machines (SVMs)
Knowledge Transfer
Loss Function
Cross-entropy Loss
Convolutional Neural Network (CNN)
Deep Reinforcement Learning (DRL)
Zero-Shot Learning
YOLO Object Detection
XGBoost
Type I Error (False Positive)
XAI (Explainable AI)
Variational Autoencoders
Weight Decay (L2 Regularization)
Weak Supervision
Variance (Model Variance)
Unsupervised Learning
Pandas and NumPy
Transformers (Transformer Networks)
Type II Error (False Negative)
True Positive Rate (TPR)
Triplet Loss
Tabular Data
Training Data
Transfer Learning
Text Generation Inference
Surrogate Model
Synthetic Data
Stream-Based Selective Sampling
Supervised Learning
PyTorch
Sliding Window Attention
Semi-supervised Learning
Sentiment Analysis
Segment Anything Model (SAM)
Sensitivity and Specificity of Machine Learning
Semantic Segmentation
Self-Supervised Learning
Selective Sampling
Scikit-Learn
Scale Imbalance
ROC (Receiver Operating Characteristic) Curve
Responsible AI
Reinforcement Learning
Regularization Algorithms
Regression (Regression Analysis)
Region Proposal Network (RPN)
Region-Based CNN (R-CNN)
Recurrent Neural Network (RNN)
Recall (Sensitivity or True Positive Rate)
Random Forest
RAG Architecture
Query Synthesis Methods
Query Strategy (Active Learning)
Quantum Machine Learning
Prompt Engineering
Prompt Injection
Prompt Chaining
Predictive Model Validation
Precision
PR AUC
Pose Estimation
Pooling
Pool-Based Sampling
Pixel
Perceptron
Principal Component Analysis
Pattern Recognition
Parameter-Efficient Fine-Tuning (Prefix-Tuning)
Panoptic Segmentation
PACS (Picture Archiving and Communication System)
Overfitting
One-Shot Learning
Outlier Detection
Optimization Algorithms
Optical Character Recognition (OCR)
Object Tracking
Object Localization
Object Recognition
Object Detection
Normalization
Noise
NIfTI
Neural Networks
Neural Style Transfer
Neural Architecture Search
Natural Language Processing (NLP)
Motion Estimation
Multi-Task Learning
Motion Detection
Model Validation
Model Parameters
Micro-Models
Medical Image Segmentation

Get the most out of your data

Lightly optimizes your model training by selecting the most valuable data, reducing redundancy and bias while improving performance.

Get in contact with us
Company & Product
Legal
Postal Address
Social
info@lightly.ai
Copyright © 2024 Lightly AG. All rights reserved.