Results for "probabilistic loss"

A model that assigns probabilities to sequences of tokens; often trained by next-token prediction.

Scales logits before sampling; higher increases randomness/diversity, lower increases determinism.

Samples from the smallest set of tokens whose probabilities sum to p, adapting set size by context.

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

Exponential of average negative log-likelihood; lower means better predictive fit, not necessarily better utility.

Bayesian parameter estimation using the mode of the posterior distribution.

AI subfield dealing with understanding and generating human language, including syntax, semantics, and pragmatics.

Categorizing AI applications by impact and regulatory risk.

Estimating parameters by maximizing likelihood of observed data.

Models that define an energy landscape rather than explicit probabilities.

Probabilistic model for sequential data with latent states.

Simultaneous Localization and Mapping for robotics.

Eliminating variables by integrating over them.

Sampling multiple outputs and selecting consensus.

Probabilities do not reflect true correctness.

Software pipeline converting raw sensor data into structured representations.

External sensing of surroundings (vision, audio, lidar).

Sampling-based motion planner.

Estimating robot position within a map.

Modeling environment evolution in latent space.

Understanding objects exist when unseen.

Human-like understanding of physical behavior.

Inferring human goals from behavior.

Controlling robots via language.

Predicting case success probabilities.

AI proposing scientific hypotheses.

Risk threatening humanity’s survival.

The field of building systems that perform tasks associated with human intelligence—perception, reasoning, language, planning, and decision-making—via algori...

Inferring the agent’s internal state from noisy sensor data.

Learning a function from input-output pairs (labeled data), optimizing performance on predicting outputs for unseen inputs.