Decision-making

Important aspects in decision-making

  • Setting goals and strategy

  • Data collection and analysis

  • Feasibility and risk assessment

  • Creativity and innovation

  • Participation and cooperation

AI-powered decision-making models

AI-powered decision-making models use algorithms and methodologies to make decisions based on available data.

  • Purpose: analyze, optimize, or recommend actions based on decision criteria.

  • Examples: decision trees, statistical models, ML models

AI decision-making models and trees

Analytical decision-making models

  • Decision trees: a model of a hierarchical structure that relies on multiple choices and conditions.

Recommendation decision-making

  • Recommendation trees: like decision trees, these trees generate recommendations based on available data and rules.

Optimization models

  • Optimization trees: these trees look for the best solutions according to the criteria set.

Decision-making process with AI

  • Data collection

  • Data processing: cleaning data

  • Model selection or development: choosing appropriate model based on requirements

  • Model training: utilizing past data to train the model.

  • Decision-making: complexity, loops, iteration, direct transition

Parameter Tuning vs. Forward and Backward

Propagation

  • Parameter Tuning: adjusting model parameters for performance improvement

  • Forward and backward propagation: fundamental processes in neural network training

Forward Propagation

  • Prediction process

    • Passing input data through network layers

    • Transformation and prediction calculation

  • Errors calculation: comparing predictions to actual outcomes

Backpropagation

  • Learning mechanism: iterative adjustment of model parameters

  • Error reduction: minimizing overall error through parameter updates

Iterative process of forward and backpropagation

  • Training neural networks: connection between forward and backward propagation

  • Learning cycle: Prediction, error calculation, parameter update

  • Importance: foundation for developing adaptable AI systems

Human or AI?

The use of AI

  • Quick search and analysis

  • Decision support by data

  • Routine tasks, data management

  • Risk and feasibility assessment

  • Problem-solving modeling

The greater role of a human

  • Creativity and innovation

  • Setting goals and strategy

  • Participation and cooperation

  • Consideration of alternatives

  • Alignment with ethics

Human or AI success area?

Humans are doing better

  • Intuitive decisions

  • Difficult decisions

  • Tactical decisions

  • Strategic decisions

  • Group decisions

AI is doing better

  • Routine decisions

  • Repetitive decisions

  • Programmable

  • Urgent decisions