During October 2025, Jommmycha developed a Beamer presentation for the intsystems/BMM repository, focusing on likelihood bracketing techniques for decision-making in belief networks. The work detailed theoretical foundations and algorithms for computing upper and lower bounds in sigmoid and noisy-OR networks, addressing the challenge of inference in intractable probabilistic models. Using LaTeX and Beamer, Jommmycha prepared academic slides that integrated experimental results, demonstrating how bound-based methods can support decisions without requiring exact inference. This contribution enhanced the team’s ability to communicate complex probabilistic modeling concepts and provided practical tools for reducing computational costs in machine learning applications.