Finance & AI Glossary.

Excess return on an investment relative to the return of a benchmark index, representing active portfolio management value.

The simultaneous purchase and sale of an asset in different decentralized venues to lock in a risk-free price discrepancy.

Non-traditional datasets used by institutional quants to generate edge, including satellite imagery, credit card transactions, and web scraping.

An automated execution loop where autonomous AI agents plan, invoke external tools, validate outputs, and execute complex multi-step tasks.

An autoregressive integrated moving average model used for fitting and forecasting time series data based on past values.

A mathematical mechanism that enables neural networks to dynamically weigh context relevance across different parts of an input sequence.

A traffic management control mechanism used by servers to delay or reject incoming requests exceeding designated rate limits.

A measure of the volatility or systematic risk of a security or portfolio in comparison to the broader market.

An asset allocation model that combines market equilibrium returns with investor-specific views to generate stable portfolio weights.

A continuous-time stochastic process used extensively in quantitative finance to model asset prices and random market fluctuations.

The difference between the highest price a buyer is willing to pay and the lowest price a seller is willing to accept.

A financial strategy where an investor borrows money in a low-interest-rate currency to fund investments in higher-yielding assets.

The institutional clearing and final transfer of assets or cash between counterparties operating in different global jurisdictions.

The probability or likelihood that the other party in an investment agreement or contract might default on its obligations.

The maximum token capacity an AI model can process simultaneously in memory during a single inference request.

Automated regulatory frameworks designed to prevent money laundering, terrorist financing, and illicit capital flows across financial platforms.

A statistical property of a collection of time series variables indicating that a linear combination of them is stationary.

A mathematical function used to couple multivariate probability distributions to model complex asset dependencies and tail risk.

A foundational model establishing the linear relationship between systematic risk and expected return for assets.

An options strategy aimed at directional neutrality by offsetting long or short positions with the underlying asset according to its delta.

A private financial exchange or forum for trading securities where institutional investors execute large blocks without pre-trade transparency.

The process of standardizing numerical fields and schemas across disparate financial feeds to prevent distortion in quantitative models.

A risk measure that evaluates the average losses incurred beyond the Value at Risk threshold in tail-risk scenarios.

A dense numerical vector representation of words, sentences, or structured tables that captures semantic meaning in machine learning models.

An investment approach that targets specific drivers of return across asset classes, such as value, momentum, size, and quality.

An asset pricing model that expands on the CAPM by adding size risk and value risk factors to explain stock returns.

Financial statement metrics, macroeconomic indicators, and corporate disclosures utilized for long-term valuation models.

Programmatic endpoints providing real-time or historical market feeds, order execution layers, and portfolio analytics.

Providing a small number of high-quality examples directly inside the prompt context to guide an LLM toward the desired output format.

The use of borrowed capital or derivative instruments to amplify potential investment returns and risk exposure.

A standardized legal agreement to buy or sell a specific commodity or financial instrument at a predetermined future date and price.

Over-the-counter derivative contracts through which two parties exchange financial cash flows or risk profiles over time.

An uncollateralized smart contract loan where borrowing and repayment occur instantaneously within the exact same transaction block.

Mathematical variables used to quantify the sensitivity of an option’s price to changes in underlying parameters like price, volatility, and time.

Generalized Autoregressive Conditional Heteroskedasticity model used extensively to estimate and forecast financial market volatility.

A statistical measure of the dispersion of returns for a given security index over a specified historical period.

Algorithmic trading characterized by extremely high speeds, high turnover rates, and high order-to-trade ratios.

The market-forecasted volatility of an asset derived from the current pricing of its options contracts.

A ratio of portfolio returns above a benchmark compared to the volatility of those excess returns (tracking error).

The temporary loss of asset value experienced by liquidity providers when token price ratios diverge relative to deposit time.

A lightweight, stateless remote procedure call protocol encoded in JSON, widely utilized in crypto nodes and quant platforms.

An algorithm that estimates unknown states from a series of noisy measurements over time, frequently applied in statistical arbitrage pairs trading.

An electronic list of buy and sell orders organized by price level, maintained by a trading exchange.

The ease with which an asset can be converted into cash without drastically affecting its market price.

The process of adapting a pre-trained large language model on specialized domain corpora (like financial filings or code) to optimize performance.

The processing delay between sending an input prompt to an AI model and receiving the completed response stream.

Order book data displaying aggregated market depth, showing multiple bid and ask price levels alongside resting order sizes.

A crowdsourced pool of digital cryptocurrencies or tokens locked in a smart contract to facilitate automated decentralized trading.

A parameter-efficient fine-tuning technique that reduces memory overhead by freezing base weights and training low-rank matrices.

Rigorous automated testing frameworks measuring agent accuracy, reasoning latency, and hallucination rates against standardized financial datasets.

A computational algorithm that relies on repeated random sampling to model the probability of various outcomes in complex financial systems.

A firm or individual that quotes both a buy and a sell price in a financial instrument, ensuring continuous market liquidity.

A financial theory suggesting that asset prices and historical returns eventually return to their long-term average state.

A strategy where traders buy assets that have shown upward price trends or sell assets with persistent downward trajectories.

An open standard introduced by Anthropic that enables developers to build secure, bidirectional connections between AI models and data tools.

The maximum observed peak-to-trough loss experienced by a portfolio before a new peak is attained.

Modern Portfolio Theory framework designed to maximize expected return for a given level of total portfolio risk.

A broker’s demand that an investor deposit additional cash or collateral to bring account equity up to the minimum maintenance level.

The process of reducing precision weights (e.g., from FP16 to INT4) to accelerate inference speeds and minimize VRAM consumption.

The continuous stream of transactions and pending orders in a financial market, providing deep visibility into immediate supply and demand dynamics.

Aggregated financial data encapsulating Open, High, Low, Close prices and Volume over discrete time intervals.

A risk-adjusted performance measure that captures all moments of the return distribution rather than just mean and variance.

A metric measuring the structural discrepancy between total bid volume and total ask volume inside the limit order book.

A derivative that grants the buyer the right—but not the obligation—to buy or sell an underlying asset at a set strike price.

A dimensionality reduction technique used in quantitative finance to identify uncorrelated underlying risk factors from a large matrix of assets.

The systematic practice of structuring instructions, schemas, and context constraints to maximize accuracy and reliability in LLM outputs.

An expansionary monetary policy where a central bank purchases government securities to increase money supply and liquidity.

A restriction imposed by API providers on the number of requests a client can execute within a specified rolling time window.

An AI framework that combines semantic vector search with LLMs to query internal enterprise documents and real-time knowledge bases.

An asset allocation strategy that builds portfolios by balancing the risk contribution of each underlying asset class.

A metric used to calculate the risk-adjusted return of a financial portfolio, measuring excess return per unit of deviation.

A variation of the Sharpe ratio that differentiates harmful volatility from total overall volatility by using downside deviation.

The difference between the expected price of a trade and the actual execution price due to market latency or thin liquidity.

A computational trading strategy that applies mean-reversion models and machine learning to large baskets of securities.

An automated system used to divide and route order flows across multiple execution venues to achieve optimal pricing and minimal slippage.

Simulating extreme macroeconomic or geopolitical shock scenarios to evaluate portfolio resilience and capital adequacy.

Self-executing code deployed on distributed ledgers that automatically enforces agreement terms between trustless counterparties.

A stochastic process whose statistical properties such as mean and variance remain constant over time.

An automated execution instruction designed to limit an investor’s loss on a security position when prices move adversely.

The practice of selling borrowed securities with the expectation of repurchasing them later at a lower price.

A cryptographic subfield enabling multiple parties to jointly compute functions over private inputs without revealing those inputs.

Artificially generated market transactions and tabular records engineered to simulate genuine statistical properties without breaching privacy.

An execution algorithm that splits a large order into equal chunks distributed evenly across a specified time horizon.

The most granular level of financial market data, capturing every single price change or trade execution event in real-time.

The capability of an LLM to dynamically format and invoke external functions, APIs, or databases to fetch precise quantitative information.

The process of breaking down raw text or code into smaller sub-word units before feeding them into neural network token embeddings.

The economic design, supply schedules, utility mechanics, and incentive structures governing digital asset ecosystems.

A risk-adjusted performance measure that uses systematic risk (Beta) instead of total risk to evaluate portfolio returns.

The portfolio risk associated with rare, extreme market events occurring beyond three standard deviations from the mean.

A large single institutional order divided into smaller visible blocks to hide total execution size from the public order book.

A dynamic stop-loss order that adjusts automatically relative to market price movements, locking in unrealized profits.

A deep learning neural network architecture based on multi-head self-attention mechanisms, serving as the backbone for modern LLMs.

A statistical technique used to measure and quantify the level of financial risk within a firm or investment portfolio over a specific timeframe.

A trading benchmark that represents the average price a security has traded at throughout the day, weighted by volume.

A mathematical search technique utilizing high-dimensional embeddings to find semantically similar documents or financial data points.

A persistent, full-duplex communication protocol running over a single TCP connection for ultra-low latency real-time data feeds.

An HTTP-based callback mechanism that enables real-time event notifications between server applications and trading agents.

A graphical curve depicting interest rates across different contract maturities for debt instruments issued by identical borrowers.

The ability of an AI model to successfully execute a task it has not been explicitly trained or fine-tuned to perform beforehand.

A cryptographic protocol allowing one party to prove to another that a statement is true without revealing any underlying data.