A Data-Transparent Probabilistic Model of Temporal Propositional Abstraction
This work addresses fundamental limitations in probabilistic modeling for AI and machine learning, though it appears incremental as it builds on existing Markov chain concepts.
The authors tackled the challenges of data scarcity, large hypothesis space, and poor transparency in probabilistic models by proposing a data-driven temporal propositional reasoning model that reverses the conventional relationship between data and domain knowledge, showing equivalence to a full-memory Markov chain and addressing issues like the zero-frequency problem.
Standard probabilistic models face fundamental challenges such as data scarcity, a large hypothesis space, and poor data transparency. To address these challenges, we propose a novel probabilistic model of data-driven temporal propositional reasoning. Unlike conventional probabilistic models where data is a product of domain knowledge encoded in the probabilistic model, we explore the reverse direction where domain knowledge is a product of data encoded in the probabilistic model. This more data-driven perspective suggests no distinction between maximum likelihood parameter learning and temporal propositional reasoning. We show that our probabilistic model is equivalent to a highest-order, i.e., full-memory, Markov chain, and our model requires no distinction between hidden and observable variables. We discuss that limits provide a natural and mathematically rigorous way to handle data scarcity, including the zero-frequency problem. We also discuss that a probability distribution over data generated by our probabilistic model helps data transparency by revealing influential data used in predictions. The reproducibility of this theoretical work is fully demonstrated by the included proofs.