Page 40 - Touhpad Ai

u Involves incomplete data, external factors, or human behaviour.
u Predictions are based on probabilities (e.g., 70% chance).
u Used in real-world situations where perfect accuracy is not possible.
Real-Life Examples

Some real-life examples of probabilistic problems are as follows:
u Google maps travel time: Provides an estimated arrival time, but it can change due to traffic or accidents.
u Stock market prediction: AI analyses trends but results vary based on news, economy, or global events.
u Movie and music recommendations: Platforms like Netflix and Spotify suggest options based on user history, but
there’s no guarantee that the suggested content will match every mood or preference.
u Weather forecasting: Meteorologists use satellite data and models to predict rain or storms, but the exact timing and
intensity can still be uncertain.
u Health predictions: Doctors or AI tools may predict the probability of a disease based on symptoms and reports, but
other unseen health factors may alter the final diagnosis.
u Online shopping suggestions: Amazon or Flipkart recommends products based on your browsing history, but it can't
be sure you'll actually buy them.
u Credit card fraud detection: Banks flag transactions as “likely fraudulent” using AI models, but some flagged ones
might be genuine, and some frauds might go undetected.
u Job recommendation systems: Platforms like LinkedIn suggest jobs based on your profile, but not every suggestion
will match your interest or skills.
u Voice assistants: They interpret speech using probabilistic models; background noise or accents can lead to errors.
u Energy demand forecasting: Power companies predict electricity needs, but unexpected events like heatwaves can
change the demand suddenly.

Role of Probabilistic Models in AI
Artificial Intelligence (AI) frequently uses probabilistic approaches because the real world is rarely black-and-white.
Some applications include:
u Machine learning algorithms: Models such as decision trees, Bayesian networks, and neural networks learn from
data patterns to make probability-based predictions.
u Natural Language Processing (NLP): Speech recognition tools (like Siri or Google Assistant) work with probabilistic
models since accents, noise, and tone make exact interpretation challenging.
u Computer vision: Image recognition systems detect objects and faces with a percentage of confidence (e.g., “This is
92% likely to be a cat”).
u Spam filtering: Emails are classified as spam or safe based on probability scores.
u Medical diagnosis systems: AI predicts possible illnesses (e.g., “There is an 80% likelihood of flu”), helping doctors
make better decisions.

Balancing Deterministic and Probabilistic Thinking in AI
AI systems often need to combine deterministic rules with probabilistic reasoning:
u Deterministic tasks: When exact steps and outcomes are required (e.g., calculating 2 + 2 or executing a fixed
algorithm).
u Probabilistic tasks: When the environment is uncertain, and predictions are based on best possible estimates (e.g.,
predicting weather, recognizing speech).
AI experts learn when to apply deterministic logic and when to rely on probabilistic reasoning, one develops the mindset
of an AI thinker that enable to handle both certainty and uncertainty intelligently.

38 Touchpad Artificial Intelligence - XI