AI terms starting with L.

Label

लेबल

The correct answer attached to a training example that the model tries to predict.

Labeling (Annotation)

लेबलिङ / टिप्पणी

Manually tagging data with correct answers so models can learn from it.

Language Model

भाषा मोडेल

A model that predicts likely sequences of words to understand or generate text.

Large Language Model (LLM)

ठूलो भाषा मोडेल

A very large neural network trained on massive text to understand and generate language.

Latency

ल्याटेन्सी / विलम्ब

The delay between sending a request to a model and receiving its response.

Latent Dirichlet Allocation (LDA)

लेटेन्ट डिरिक्ले एलोकेसन

A statistical method for discovering topics across a set of documents.

Latent Space

लेटेन्ट स्पेस / गुप्त स्थान

A compressed internal representation where a model encodes the essential features of data.

Layer

तह / लेयर

A group of neurons at the same depth in a network; data flows from layer to layer.

Layer Normalization

लेयर नर्मलाइजेसन

Normalizing values across features within a layer to stabilize training.

Lead Scoring

लिड स्कोरिङ

Ranking potential customers by how likely they are to convert into buyers.

Learning Curve

सिकाइ वक्र

A plot showing how a model's performance changes as it sees more training data.

Learning Rate

लर्निङ रेट / सिकाइ दर

A setting that controls how big each step is when a model updates its parameters.

Lemmatization

लेमाटाइजेसन

Reducing words to their dictionary base form using grammar, e.g. 'better' to 'good'.

LiDAR

लिडार

A sensor that measures distance with laser light to build 3D maps of surroundings.

Linear Regression

रेखीय रिग्रेसन

A model that fits a straight line to predict a numeric value from input features.

Load Balancing

लोड ब्यालेन्सिङ

Distributing requests across multiple servers to keep a system fast and reliable.

Logging

लगिङ

Recording events and messages from a running system for debugging and analysis.

Logic

तर्कशास्त्र

The formal study of valid reasoning, a foundation of early AI systems.

Logistic Regression

लजिस्टिक रिग्रेसन

A model used for classification that outputs the probability of a category.

Long Context

लामो सन्दर्भ

The capacity of a model to process very large amounts of input text at once.

LoRA

लोरा (LoRA)

A lightweight fine-tuning method that adapts a large model by training small added layers.

Loss Function

लस फंक्सन / हानि फंक्सन

A measure of how far a model's predictions are from the correct answers.

Low-Code / No-Code

लो-कोड / नो-कोड

Platforms that let people build software with little or no traditional programming.

LSTM

एलएसटीएम

A type of RNN that can remember information over long sequences using memory gates.