The Semantic Web Blog mentioned here, there is speculation that the Siri intelligent personal digital assistant technology may come to light in Apple’s fall iOS 5 release. Which is all well and good for users on Apple’s mobile platforms, but myBantu founder and vice president of products and strategy Bharath Yadla sees an opportunity to bring personalized recommendations for entertainment to a wider swath of the populace. The iPhone and Android version of its application, to join its web and Facebook applications. officially launch next week.

“We see absolutely a great opportunity with other platforms from a mobile standpoint,” he says. “And when you leverage the application on Facebook, that is a predominant presence.”

What users are leveraging in this intelligent assistant is its ActiveRelevance platform for providing relevant and personalized recommendations based on their profiles and queries. The platform leverages both artificial intelligence and social relevance, assessing some nine dimensions, including personal interests, proximity of choices, popularity, and peer influence, in order to deliver a handful or two of results. The social relevance comes by way of crowd-sourcing recommendations from sites such as OpenTable and Rotton Tomatoes, friends on social networking platforms, and the output of search as well as semantic engines. Users can enter their requirements in natural language, such as top romantic restaurant nearby, for the ActiveRelevance engine to parse intent and come to some conclusions.

“It understands the context of what I am looking at and gives recommendations,” including considering the location a user lists in his profile or using GPS data for the mobile apps. “From our semantic set it dynamically understands categories, distance, user ratings and so on.” To accomplish its goals, the ActiveRelevance technology leverages some existing ontologies, such as Freebase, and has created some of its own. It also makes use of entity recognition services such as OpenCalais.

For a request to find a good comedy movie playing nearby, for instance, it might deliver recommendations for Bridesmaids, Midnight in Paris, and so on from multiple sources, Yadla says. “It dynamically picks up the semantics of the recommendation itself.” It also can parse Facebook friends’ comments into “aye” or “nay” opinions about that movie – or dining, entertainment or shopping.

As he sees it, the differentiator for the myBantu service starts with the context that lets it pick up relevant content from multiple parties. One challenge it has yet to conquer for the mobile apps, however, is voice enablement, but that’s expected in an upcoming release a few months down the road.