Original title: Dynamic probabilistic decision networks
V. I. Yukalov, E. P. Yukalova · September 2026
Making good choices is hard because feelings and biases often get in the way. Researchers designed a new kind of decision network to study this issue. In this network, connected agents make choices. These agents can represent humans, brain neurons, or artificial intelligence units. Each agent picks an option based on probability, weighing both practical use and emotions. The system is dynamicChanges over time as conditions evolve. and varied. Some agents rely on long-term memory, while others use short-term memory. As agents swap information over time, their choice probabilities shift. The researchers found that this information exchange reduces decision errors. They demonstrated this by resolving a famous puzzle called the Allais paradox. Finally, they found that machine-learning techniques can regulate agent behavior, steering them toward specific choices.
You and a group of friends are deciding whether to go out or stay in for the evening. Each person has a different mood and preference, and as you discuss, opinions shift based on what others say. In this paper, the friends are like agents in a network, and their shifting moods represent emotional biases. Talking over time corresponds to information exchange, which helps the group make a better collective decision with fewer regrets.
This could help build AI systems that make better decisions by mimicking human-like emotions and learning from each other. It also offers a way to study how biases affect choices in groups.
The abstract doesn't state specific limitations. It presents this as a suggested model with an illustrative example, but it doesn't discuss real-world implementation or empirical testing beyond the example.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Dynamic probabilistic decision networks — V. I. Yukalov, E. P. Yukalova
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.