An educational crypto, finance, and AI vocabulary puzzle about protecting wallets, understanding bonds, evaluating models, and coordinating computing resources.
Wallet Safety
Wallet safety combines secure key handling, clear transaction review, and recovery planning to reduce avoidable loss.
Private KeyHardware WalletTransaction SimulationMultisignature
- Private Key: A private key is a secret cryptographic credential that can authorize transactions for an associated blockchain address.
- Hardware Wallet: A hardware wallet is a dedicated device that helps keep private keys isolated while approving transactions.
- Transaction Simulation: Transaction simulation estimates what a proposed blockchain transaction may do before a user signs and submits it.
- Multisignature: Multisignature is a setup that requires approval from more than one authorized key before a transaction can proceed.
Bond Market Basics
Bond markets price promised cash flows by considering interest rates, repayment timing, and the issuer's ability to pay.
Bond YieldDurationCoupon RateCredit Spread
- Bond Yield: Bond yield is the return implied by a bond's market price and its expected interest and principal payments.
- Duration: Duration estimates how sensitive a bond's price is to changes in interest rates, based on the timing of its cash flows.
- Coupon Rate: A coupon rate is the stated annual interest rate a bond issuer pays relative to the bond's face value.
- Credit Spread: A credit spread is the additional yield investors demand for a bond compared with a similar lower-risk benchmark.
Model Evaluation
Model evaluation uses representative tests, defined metrics, and human review to measure accuracy, safety, and consistency.
BenchmarkEvaluation SetHallucinationHuman Evaluation
- Benchmark: A benchmark is a standardized task or dataset used to compare the performance of AI models.
- Evaluation Set: An evaluation set is a collection of examples held back from training and used to measure model behavior.
- Hallucination: A hallucination is a model output that appears plausible but is unsupported, inaccurate, or invented.
- Human Evaluation: Human evaluation uses people to assess qualities such as helpfulness, correctness, safety, or clarity that may be hard to capture with one automatic metric.
Compute Networks
Compute networks coordinate processors, storage, and scheduling so workloads can run with appropriate capacity and reliability.
GPU ClusterWorkload SchedulerInference ServerResource Utilization
- GPU Cluster: A GPU cluster is a group of graphics processors connected to run compute-intensive workloads such as AI training or inference.
- Workload Scheduler: A workload scheduler assigns computing jobs to available machines according to capacity, priority, and resource requirements.
- Inference Server: An inference server runs a trained model and returns outputs for incoming application requests.
- Resource Utilization: Resource utilization measures how much of a computing resource such as processor time, memory, or storage is actively being used.