Today for AI

EXPLORE / 06

Terms index

Specialized AI and software-development terms with concise explanations

Entity records

4600 records · 77/115

  1. 3041

    Node-based workflow

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

    Website pending
  2. 3042

    Inference Acceleration

    Reducing model prediction time via algorithmic or hardware optimization

    Website pending
  3. 3043

    Low-resource Learning

    Machine learning approaches when labeled data is extremely limited.

    Website pending
  4. 3044

    MLP-Mixer

    Pure MLP architecture processing image patches and channel information using only multilayer perceptrons

    Website pending
  5. 3045

    Real-world Use Cases

    Scenarios where AI is applied to solve specific problems in actual business contexts

    Website pending
  6. 3046

    LLM Benchmarking

    Process of evaluating model performance using standardized datasets and metrics

    Website pending
  7. 3047

    Time Series Forecasting

    Machine learning task predicting future values based on historical data points

    Website pending
  8. 3048

    Patching

    Splitting long sequences into fixed-length sub-sequences to enhance local feature extraction

    Website pending
  9. 3049

    Inference Chip

    Hardware accelerators specifically designed to execute neural network forward propagation

    Website pending
  10. 3050

    Messages API

    A standardized conversational interface typically used to manage multi-turn dialogue context

    Website pending
  11. 3051

    Model Merging

    Combining weights from multiple fine-tuned models to retain their respective capabilities

    Website pending
  12. 3052

    Matryoshka Embeddings

    A technique allowing truncation of vector dimensions at inference to balance accuracy and storage cost

    Website pending
  13. 3053

    Text Generation

    Process where models produce coherent text based on prompts

    Website pending
  14. 3054

    Joint Reasoning

    Cognitive process deriving conclusions by combining information from different modalities

    Website pending
  15. 3055

    GPU

    Graphics Processing Unit, widely used for deep learning training and inference due to strong parallel computing capabilities.

    Website pending
  16. 3056

    Crowdsourced Evaluation

    Using many ordinary users to subjectively score or select preferences among model outputs.

    Website pending
  17. 3057

    Low-Rank Projection

    Mapping high-dimensional parameter matrices to a lower-dimensional subspace to reduce computation

    Website pending
  18. 3058

    NPU

    Hardware accelerator optimized for neural network operations

    Website pending
  19. 3059

    Binary Quantization

    Extreme compression limiting weights or activations to -1 and +1

    Website pending
  20. 3060

    Vector Quantization

    Mapping continuous high-dimensional vectors to discrete codebook indices

    Website pending
  21. 3061

    Vision Foundation Models

    Large-scale pre-trained general visual understanding models

    Website pending
  22. 3062

    GPU Inference

    The process of using GPUs to accelerate the forward pass prediction in neural networks.

    Website pending
  23. 3063

    Serverless Computing

    A cloud computing execution model where the provider dynamically manages allocation of machine resources.

    Website pending
  24. 3064

    Open Source Governance

    Mechanisms managing decision rights, contribution processes, and community norms in open source projects

    Website pending
  25. 3065

    AI Regulation

    Legal norms established by governments or institutions for AI development and application

    Website pending
  26. 3066

    Dataset Viewer

    Tool interface for online previewing and exploring ML datasets

    Website pending
  27. 3067

    Text2SQL

    Technology converting natural language queries into structured SQL statements

    Website pending
  28. 3068

    Self-Alignment

    Technique where a model uses its own generated data for preference optimization or instruction tuning

    Website pending
  29. 3069

    Speaker Diarization

    Technology that distinguishes different speakers in audio

    Website pending
  30. 3070

    Key-Value Cache

    Stores historical key and value vectors from attention layers to accelerate inference

    Website pending
  31. 3071

    GPU Acceleration

    Using parallel processing power of graphics cards to speed up deep learning training and inference

    Website pending
  32. 3072

    Pretrained Language Model

    Model that acquires general language understanding via self-supervised learning on massive unlabeled text

    Website pending
  33. 3073

    Inference Accelerator

    Hardware chips optimized specifically for machine learning inference tasks

    Website pending
  34. 3074

    Non-Player Character (NPC)

    Game characters controlled by the computer rather than players, often serving as narrative or interaction vehicles

    Website pending
  35. 3075

    Model Arena

    A mechanism ranking multiple models' output quality via user blind-test voting.

    Website pending
  36. 3076

    Model Containerization

    Packaging models with dependencies for standardized deployment

    Website pending
  37. 3077

    FSDP

    PyTorch native strategy sharding parameters, gradients, and optimizer states across devices

    Website pending
  38. 3078

    Data Quality

    Metrics assessing accuracy, completeness, consistency, and usability of datasets

    Website pending
  39. 3079

    Document VQA

    A task that combines image and text information to answer questions about document content

    Website pending
  40. 3080

    Open LLM

    Large language models with publicly available weights allowing free modification and use

    Website pending