An educational crypto, finance, and AI vocabulary puzzle about blockchain participation, business costs, model architecture, and dataset metadata.
Blockchain Node Roles
Node roles differ in the history they retain, the data they verify, and their consensus duties.
Full NodeArchive NodeLight ClientValidator Node
- Full Node: A full node independently checks blocks and transactions against the rules of its blockchain.
- Archive Node: An archive node retains historical blockchain states so applications can query past state directly.
- Light Client: A light client uses compact chain information and proofs to verify selected blockchain data without processing everything a full node does.
- Validator Node: A validator node participates in a blockchain consensus protocol by proposing or attesting to blocks.
Cost Accounting
Cost accounting distinguishes costs by their behavior and how they relate to a product or activity.
Fixed CostVariable CostDirect CostIndirect Cost
- Fixed Cost: A fixed cost stays constant in total over a specified period and relevant range of activity.
- Variable Cost: A variable cost changes in total as the level of production or another activity changes.
- Direct Cost: A direct cost can be specifically traced to a product, service, or other cost object.
- Indirect Cost: An indirect cost supports multiple cost objects and is allocated rather than directly traced to just one.
Neural Network Layers
Different layers learn representations, combine features, and change spatial resolution.
Embedding LayerDense LayerConvolution LayerPooling Layer
- Embedding Layer: An embedding layer maps discrete input identifiers to learned numerical vectors.
- Dense Layer: A dense layer applies learned weights connecting each input feature to each output unit, typically with a bias.
- Convolution Layer: A convolution layer applies shared learned filters across local regions of an input to detect patterns.
- Pooling Layer: A pooling layer summarizes groups of nearby values, often reducing the spatial size of a feature map.
Data Catalog Fields
Catalog metadata describes who maintains data, what its fields mean, and how it was produced.
Data OwnerColumn DescriptionSource SystemRefresh Cadence
- Data Owner: A data owner is the person or team accountable for decisions about a dataset and its management.
- Column Description: A column description explains what a field represents, including relevant units or interpretation rules.
- Source System: A source system is the application or platform from which a dataset originates.
- Refresh Cadence: Refresh cadence describes how often a dataset is expected to receive updated data.