It provides a mathematical framework for representing and manipulating uncertainty. Unlike deterministic systems which operate under the assumption of exact and complete information, probabilistic systems acknowledge that the real world is fraught with uncertainties. By employing probability, AI systems can make informed decisions even in the face of ambiguity.
Need for probabilistic reasoning in AI
- Machine Learning
- Robotics
- Natural Language Processing
- Decision making systems
Key concepts in Probabilistic Reasoning
1. Bayesian Networks
- It is a relationship between variables and their conditional probabilities
- Advantages: Are very effective to express cause-effect and reasoning about missing information.
2. Markov Models
- Predicts the future state of a system from its current state and its past history.
- Example: A simple Markov model of weather would predict that the next day after a sunny day would be a sunny day with high probability, as compared to a rainy day the next day.
- Advantages: Markov models are easy to implement and are widely used for prediction systems.
3. Hidden Markov Models
- Used to predict future states of a system which also has some hidden states and hidden variables.
- Examples: A stock market prediction engine where the variables that govern prices (human sentiment) are not fully transparent.
4. Probabilistic Graphical Models
- This is a broader framework encompassing both Bayesian networks and Markov models.
- An approach for representation and reasoning in a framework of uncertain information, given in graphical structure
- Advantages: PGMs are powerful, flexible and expressive language for doing probabilistic reasoning, which is well suited for complex relationships that may capture different types of uncertainty.
Techniques in Probabilistic reasoning
1. Inference
- The process of computing the probability distribution of certain variables given known values of other variables.
- The inference methods include variable elimination and the junction tree algorithm, while approximate inference methods include Markov Chain Monte Carlo (MCMC) and belief propagation.
2. Learning
- Involves updating the parameters and structure of probabilistic models based on observed data. Techniques include maximum likelihood estimation, Bayesian estimation, and expectation maximization (EM)
3. Decision making
- Utilizing probabilistic models to make decisions that maximize expected utility.
- Techniques involve computing expected rewards and selecting actions accordingly, often implemented using frameworks like POMDPs(Partially Observable Markov Decision Process).
How does probabilistic reasoning empower AI systems
- Quantifying uncertainty
- Reasoning with evidence
- Based on past experience
- Effective decision making
Applications of Probabilistic Reasoning in AI
- Robotics
- Healthcare
- Natural Language Processing
- Finance
Advantages of Probabilistic Reasoning
- Flexibility
- Robustness
- Interpretable