Today for AI

EXPLORE / 06

Terms index

Specialized AI and software-development terms with concise explanations

Entity records

4598 records · 76/115

  1. 3001

    Jupyter Notebook

    An interactive computing environment supporting live code execution and visualization

    Website pending
  2. 3002

    ASIC

    Integrated circuits customized for specific applications, often used for high-performance computing tasks.

    Website pending
  3. 3003

    Graph Classification

    Task predicting labels for entire graph structures, common in molecular property prediction

    Website pending
  4. 3004

    Transformers

    Deep learning architecture based on attention mechanisms, widely used in NLP

    Website pending
  5. 3005

    Data Cleaning

    Process of correcting or deleting erroneous and redundant data before model training

    Website pending
  6. 3006

    Near-deduplication

    Identifying and removing highly similar but not identical data samples

    Website pending
  7. 3007

    Safetensors

    A secure, fast, and portable format for storing tensors

    Website pending
  8. 3008

    NormalFloat4

    A 4-bit data type designed for normally distributed weights

    Website pending
  9. 3009

    Inference Container

    A standardized software package pre-configured with model loading, API interfaces, and dependencies.

    Website pending
  10. 3010

    Topic Modeling

    A technique for discovering abstract topics from large collections of text

    Website pending
  11. 3011

    Word Embeddings

    A technique mapping words to continuous numerical vectors in NLP

    Website pending
  12. 3012

    SQL Query

    The standard method for retrieving and manipulating relational data using Structured Query Language.

    Website pending
  13. 3013

    Embedded Database

    A lightweight database system that runs within an application process without requiring a separate server.

    Website pending
  14. 3014

    Stable Diffusion

    A text-to-image generation algorithm based on latent diffusion models

    Website pending
  15. 3015

    Model Bias

    Systematic deviation of model outputs from objective facts or fairness

    Website pending
  16. 3016

    Transformers.js

    JavaScript library for running Hugging Face models in the browser

    Website pending
  17. 3017

    EU AI Act

    The world's first comprehensive legal framework proposed by the EU to regulate AI.

    Website pending
  18. 3018

    Fully Homomorphic Encryption (FHE)

    Encryption scheme allowing computation on ciphertexts to generate an encrypted result that decrypts correctly.

    Website pending
  19. 3019

    3D Asset Generation

    The process of automatically creating 3D models from text or images using AI algorithms

    Website pending
  20. 3020

    Audio Generation

    Technology synthesizing music, speech, etc., using deep learning models

    Website pending
  21. 3021

    GPTQ

    A post-training quantization algorithm based on second-order information, designed for generative pre-trained transformers.

    Website pending
  22. 3022

    MLOps

    Practices for continuous integration, delivery, and operations of ML systems

    Website pending
  23. 3023

    Controllable Generation

    Guiding model output to meet specific requirements using conditional inputs

    Website pending
  24. 3024

    Adapter

    A lightweight module used to fine-tune pretrained models without altering original weights

    Website pending
  25. 3025

    3D Gaussian Splatting

    A real-time rendering technique using anisotropic Gaussians to represent scenes

    Website pending
  26. 3026

    RL Fine-tuning

    Adjusting pre-trained model parameters using reward signals to optimize specific goals

    Website pending
  27. 3027

    DDPM

    A deep learning architecture that generates data by gradually adding and removing noise

    Website pending
  28. 3028

    Inference API

    Service interface allowing developers to call pretrained models for prediction via HTTP requests

    Website pending
  29. 3029

    SDXL

    Stable Diffusion XL, a high-resolution text-to-image generation model

    Website pending
  30. 3030

    Private Deployment

    Running models locally or in private clouds instead of public ones

    Website pending
  31. 3031

    Latent Consistency Model (LCM)

    A technique that distills diffusion models into few-step generators.

    Website pending
  32. 3032

    Discrete Reasoning

    Reasoning tasks involving numerical calculations, sorting, etc., requiring precise logical steps.

    Website pending
  33. 3033

    Cold Start

    Significant latency occurring when a system starts from scratch or after idle time due to unwarmed resources.

    Website pending
  34. 3034

    TensorRT-LLM

    An open-source framework by NVIDIA designed to accelerate large language model inference

    Website pending
  35. 3035

    Aspect-Based Sentiment Analysis

    Identifying sentiment polarity towards specific aspects or attributes in text

    Website pending
  36. 3036

    Sparse Activation

    A mechanism that activates only a subset of parameters or sub-networks during inference to reduce computation

    Website pending
  37. 3037

    Leaderboard

    A system that ranks and compares model performance based on specific benchmarks.

    Website pending
  38. 3038

    Text-to-Image Generation

    AI technology that automatically generates visual images from natural language descriptions.

    Website pending
  39. 3039

    Node-based workflow

    A visual programming method that builds complex AI processing pipelines by connecting functional nodes.

    Website pending
  40. 3040

    Inference Acceleration

    Reducing model prediction time via algorithmic or hardware optimization

    Website pending