Unfortunately this only comes with practice and it only applies to a specific domain. A physicist cannot easily weed out computer science literature and vice versa.

In fact, there are identical problems that are solved by different communities and you would not know because they use completely different lingo. Math optimization/dynamic programming/reinforcement learning is one of these.

Many of the accomplished scientists just read papers from other domains and adapt them to their own domain making huge progress.

So yes I see tremendous value to what you describe. A Google translate for academic work that can translate between domain specific lingos and common language.