facebook-m

Are you one of the few hundred lucky Facebook users who got an unusual chance last month to talk to a mysterious new contact called M?
M is actually Facebook's Cyborg Virtual Assistant powered by a mixture of algorithms and human operators.

Facebook’s Vice President of Messaging Products David Marcus today announced 'M' in August this year, however its not available all FB users but limited to testing for few hundred users in US.

M's cyborg design makes it efficient for handling much more complex requests than its Apple (Siri), Microsoft (Cortana), and Google counterparts. While Siri, Cortana, and Google's search app are capable of understanding simple commands or queries, such as "What’s the weather forecast for Delhi?" But they can't answer a more complex question such as "Where can I get good Punjabi Food in Bangalore?" They aren't capable of doing a back-and-forth conversation or booking a taxi or a flight.

m-exampleM, on the other hand, can do all the above stated tasks because the software hands over the things it can't do to human operators called "trainers". Of course in some cases, the trainer has to do all the work, but M is equally capable of taking queries it recognizes but can’t filter them into easy-to-process summaries that make a trainer's handling the task more efficient.

Morover, unlike other AI-based service like Apple Siri and Google voice search, M can actually complete tasks on your behalf. It can purchase items, get gifts delivered to your loved ones, book restaurants, travel arrangements, appointments and way more.

In August this year Wired reported that a few hundred San Francisco Bay Area (in California) Facebook users opened their Messenger apps to discover 'M'. Facebook would prompt them to test it with examples of what M can do: Make restaurant reservations. Find a birthday gift for your spouse. Suggest—and then book—weekend getaways.

Currently, this model isn't efficient enough for M to be more than an experiment. This is mainly because it requires too many human workers. But, according to Alex Lebrun, who leads the team working on the assistant, It has the capability to become a real product because the actual work of the human trainers is to gradually teach the software how to do a greater share of the work itself.
Advertisements

Post a Comment

Comment

Previous Post Next Post