The paper introduces a new family of probability distributions built by combining two simpler distributions in a particular way. Specifically, it takes a "Bernoulli" distribution, which places probability mass at just two points, and a uniform distribution, which spreads probability evenly across an interval, and combines them using an operation called Boolean convolution. This operation comes from free probability theory, a branch of mathematics that generalizes classical probability to settings where the usual notion of independence is replaced by a different one called Boolean independence. The result is a one-parameter family of distributions, meaning there is a single tuning parameter that controls the shape of the distribution, and the paper works out its properties in careful detail.
The technical heart of the paper involves exploiting a special tool called the Boolean energy transform, which has a useful linearization property. This means that combining two distributions via Boolean convolution corresponds to simply adding their energy transforms, making calculations much more tractable. Using this, the authors derive explicit formulas for the distribution's key descriptors, including its moments (which summarize the shape of the distribution, like its spread and skewness), its Boolean cumulants (a different but related set of descriptors), and the region of the real line where the distribution actually lives. They also find a differential equation that the energy transform satisfies, and they work out how the distributions behave when scaled or viewed from far away.
One of the more conceptually interesting findings concerns whether this family forms a so-called convolution semigroup, meaning whether repeatedly applying the Boolean convolution operation within the family stays inside the family in a time-consistent way. The answer turns out to be no, which distinguishes this family from more classical examples like the Gaussian family under ordinary convolution. The paper also establishes precise estimates for how the moments grow as one looks at higher and higher order, and it pins down exactly when the distribution is supported on a single connected interval versus multiple pieces. Overall, the work contributes concrete, computable examples to the theory of Boolean probability, which tends to have fewer explicit families than its classical or free counterparts.