EXPLORE / 06
Terms index
Specialized AI and software-development terms with concise explanations
Entity records
4520 records · 89/113
- 3521
Symbol Grounding Problem
How to connect abstract symbols with real-world perceptual experiences.
Website pending - 3522
Compositionality
The meaning of complex expressions is determined by their parts and how they are combined.
Website pending - 3523
Action Recognition
Technology that automatically identifies and classifies human actions from video data
Website pending - 3524
Temporal Segment Model
An architecture that divides videos into segments and encodes them independently to capture long-range dependencies
Website pending - 3525
Reproducible Research
Research method providing full code and data to ensure others can verify results
Website pending - 3526
Language Emergence
The phenomenon where agents spontaneously form communication symbols or rules.
Website pending - 3527
Interactive Visualization
Technology allowing users to dynamically explore data and model internals via interface operations
Website pending - 3528
Sentiment Analysis
The process of identifying and extracting subjective information such as opinions, sentiments, and emotions from text.
Website pending - 3529
Stochastic Neural Network
A network structure introducing randomness to enhance exploration or robustness
Website pending - 3530
OpenAI Gym
A toolkit for developing and comparing reinforcement learning algorithms.
Website pending - 3531
Soft Q-Learning
Maximum entropy RL variant with entropy regularization
Website pending - 3532
DQN
A reinforcement learning algorithm combining deep neural networks with Q-learning.
Website pending - 3533
Upper Confidence Bound
A decision strategy based on the upper bound of confidence intervals to balance exploration and exploitation
Website pending - 3534
Q-Ensemble
Training multiple independent Q-value networks to estimate the distribution and uncertainty of the value function
Website pending - 3535
UCB
An exploration strategy based on optimism bias that prioritizes actions with the highest potential payoff
Website pending - 3536
Robotic Simulation
The technique of simulating robot dynamics in virtual environments for testing and training
Website pending - 3537
MuJoCo
A high-performance physics engine for computational physics, machine learning, and biomechanics
Website pending - 3538
Adversarial Examples
Inputs with small perturbations that cause models to produce incorrect outputs.
Website pending - 3539
Curriculum Learning
A training strategy that arranges samples from easy to hard, mimicking human learning.
Website pending - 3540
Hindsight
Re-defining goals based on actual outcomes to generate valid samples
Website pending - 3541
Experience Replay
Storing past interactions and sampling them randomly to break correlations
Website pending - 3542
Proximal Policy Optimization
A reinforcement learning algorithm that stabilizes training by limiting policy update steps
Website pending - 3543
Exploration-Exploitation
The trade-off between trying new actions to gather information and choosing known best actions.
Website pending - 3544
Reward Function
Numerical scoring mechanism defining agent behavior quality in reinforcement learning
Website pending - 3545
Supervised Learning
A machine learning paradigm using labeled data to train models for prediction.
Website pending - 3546
Sample Efficiency
Higher efficiency means fewer data points or interactions are needed to achieve specific performance.
Website pending - 3547
TRPO
An algorithm ensuring monotonic improvement by limiting the step size of policy updates.
Website pending - 3548
A3C
A reinforcement learning algorithm that updates policies using multiple parallel worker threads.
Website pending - 3549
Supervised Deep Learning
Model training method relying on labeled static datasets.
Website pending - 3550
Theory of Mind
The ability to infer others' beliefs, intentions, and mental states, often used to describe AI modeling of interaction partners.
Website pending - 3551
Opponent Modeling
Process of predicting future behaviors or policy changes of other agents
Website pending - 3552
Opponent-Learning Awareness
Method where an agent explicitly models the future learning behavior of opponents during policy updates
Website pending - 3553
Deep Linear Network
Neural network composed of multiple linear layers without activation functions
Website pending - 3554
Autoregressive Model
A generative architecture that predicts subsequent sequence elements based on prior outputs.
Website pending - 3555
Visual Reinforcement Learning
Reinforcement learning methods that learn decision-making policies directly from image pixel inputs.
Website pending - 3556
Dynamics Randomization
A method enhancing policy robustness by randomly varying physical parameters during training
Website pending - 3557
Asymmetric Architecture
Network design where different components use distinct structures or information levels
Website pending - 3558
Actor-Critic
A reinforcement learning framework combining policy gradients with value estimation
Website pending - 3559
High-level Action
Abstract behavioral units composed of low-level policies, used to simplify decision spaces
Website pending - 3560
Closed-loop Control
Control system that uses feedback to adjust outputs in real time to maintain stability or achieve goals
Website pending