← 研踪 PaperTrailarXiv:2606.23715 · v2
Random coloured digraphs defined by a Markov logic network
由马尔可夫逻辑网络定义的随机着色有向图
首次提交 2026-06-16 19:49
版本更新 2026-11-20 00:36
雷达观测 2026-11-20 12:52
摘要
A Markov Logic Network (MLN) is a probabilistic relational model used in Statistical Relational Artificial Intelligence for defining a probability distribution on the set of possible worlds with domain for an arbitrary finite domain . An MLN consists of first-order formulas, which we call soft constraints, with associated weights which are nonnegative real numbers. The behaviour of the probability distribution, including the effect of the weights, on different domain sizes is poorly understood. We approach the problem in a systematic way by studying sequences where is a first-order sentence and is the probability that holds in where . We consider a first-order language that can express a property and relation and show that all of the following cases can be realized for an MLN such that each of its soft constraints is a boolean combination of , and in which both and appears: For every first-order sentence , converges to 0 or 1 as . For every first-order sentence , converges as but not necessarily to 0 or 1. There is a first-order sentence such that does not converge as . In fact, all three cases can be realized even after fixing the soft constraints and, for any , by varying their weights by less than . We also show that, given any MLN as described above, we can efficiently determine which case it belongs to by inspecting the MLN. Then we show that if the weights are scaled by then we end up in case~(1) no matter how the weights are chosen.
作者备注:This is a major revision, especially with respect to the exposition, of the first arxiv version
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-20 00:36 · 最新观测版本
Random coloured digraphs defined by a Markov logic network
A Markov Logic Network (MLN) is a probabilistic relational model used in Statistical Relational Artificial Intelligence for defining a probability distribution on the set of possible worlds with domain for an arbitrary finite domain . An MLN consists of first-order formulas, which we call soft constraints, with associated weights which are nonnegative real numbers. The behaviour of the probability distribution, including the effect of the weights, on different domain sizes is poorly understood. We approach the problem in a systematic way by studying sequences where is a first-order sentence and is the probability that holds in where . We consider a first-order language that can express a property and relation and show that all of the following cases can be realized for an MLN such that each of its soft constraints is a boolean combination of , and in which both and appears: For every first-order sentence , converges to 0 or 1 as . For every first-order sentence , converges as but not necessarily to 0 or 1. There is a first-order sentence such that does not converge as . In fact, all three cases can be realized even after fixing the soft constraints and, for any , by varying their weights by less than . We also show that, given any MLN as described above, we can efficiently determine which case it belongs to by inspecting the MLN. Then we show that if the weights are scaled by then we end up in case~(1) no matter how the weights are chosen.
查看此版本解读 →查看此版本原文 ↗