Results for "full pass through data"

A wide basin often correlated with better generalization.

Compromising AI systems via libraries, models, or datasets.

Low-latency prediction per request.

Models accessible only via service APIs.

Measures similarity and projection between vectors.

Model behaves well during training but not deployment.

Model relies on irrelevant signals.

AI used without governance approval.

Interpreting human gestures.

Acting to minimize surprise or free energy.

Automated assistance identifying disease indicators.

US approval process for medical AI devices.

AI discovering new compounds/materials.

Predicting borrower default risk.

A learning paradigm where an agent interacts with an environment and learns to choose actions to maximize cumulative reward.

An RNN variant using gates to mitigate vanishing gradients and capture longer context.

Generates sequences one token at a time, conditioning on past tokens.

Stepwise reasoning patterns that can improve multi-step tasks; often handled implicitly or summarized for safety/privacy.

Breaking documents into pieces for retrieval; chunk size/overlap strongly affect RAG quality.

Constraining outputs to retrieved or provided sources, often with citation, to improve factual reliability.

Ensuring model behavior matches human goals, norms, and constraints, including reducing harmful or deceptive outputs.

Rules and controls around generation (filters, validators, structured outputs) to reduce unsafe or invalid behavior.

Techniques to understand model decisions (global or local), important in high-stakes and regulated settings.

Studying internal mechanisms or input influence on outputs (e.g., saliency maps, SHAP, attention analysis).

Raw model outputs before converting to probabilities; manipulated during decoding and calibration.

Reconstructing a model or its capabilities via API queries or leaked artifacts.

A discipline ensuring AI systems are fair, safe, transparent, privacy-preserving, and accountable throughout lifecycle.

Tendency to trust automated suggestions even when incorrect; mitigated by UI design, training, and checks.

Early architecture using learned gates for skip connections.

Methods for breaking goals into steps; can be classical (A*, STRIPS) or LLM-driven with tool calls.