Results for "real-time"

Classical statistical time-series model.

Repeating temporal patterns.

Identifying abrupt changes in data generation.

Training with a small labeled dataset plus a larger unlabeled dataset, leveraging assumptions like smoothness/cluster structure.

Automated detection/prevention of disallowed outputs (toxicity, self-harm, illegal instruction, etc.).

Constraining model outputs into a schema used to call external APIs/tools safely and deterministically.

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

Attention mechanisms that reduce quadratic complexity.

Models trained to decide when to call tools.

Models evaluating and improving their own outputs.

Model execution path in production.

Running new model alongside production without user impact.

Running predictions on large datasets periodically.

Centralized repository for curated features.

Using production outcomes to improve models.

Interleaving reasoning and tool use.

Dynamic resource allocation.

Running models locally.

AI systems that perceive and act in the physical world through sensors and actuators.

Internal sensing of joint positions, velocities, and forces.

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

Continuous loop adjusting actions based on state feedback.

Using output to adjust future inputs.

Planning via artificial force fields.

Ensuring robots do not harm humans.

Closed loop linking sensing and acting.

AI systems assisting clinicians with diagnosis or treatment decisions.

Ultra-low-latency algorithmic trading.

Signals indicating dangerous behavior.

Balancing learning new behaviors vs exploiting known rewards.