Facebook Is Working To Transform Chatbots Into A Real Chit-Chat Models

Facebook, the social networking giant was in news for shutting down its AI equipped chatbot project due to unwanted results. But, it seems that the incident didn't stop its researchers from exploring new ways to improve the chatbots performance. In the shared research paper, Facebook has revealed that its AI research team have been working on the true nature of the chatbots. The breakthrough in the self-operating chatting apps could be a shocker for all in the coming time.

In the published paper, Facebook has totally declined the chatbots of being an actual chatting tool. According to the tech giant, the software only talks to the user on the basis of predefined algorithms. Whereas the chatting is more than just talking it requires to understand the meaning of replies, remembering previous talks and responding on the basis of chat precisely instead of putting a well-structured sentence.

While explaining the failure of the chatbots, Facebook explained how these apps are not what we think about. According to the company, the bots fail in a number of standards of being a chatbot. Starting with the incapability of showing the ‘Consistent Personality’ by sticking to the same pattern throughout the conversation. Along with this, these apps also don't remember any piece of information from the previous chats. And finally, the researchers said, when the chatbots don't understand something they went to ‘I don't Know’ mode with a diversionary response.

“Chit-chat models are known to have several problems: they lack specificity, do not display a consistent personality and are often not very captivating. In this work we present the task of making chit-chat more engaging by conditioning on profile information”, wrote the AI research team in the paper. “We collect data and train models to (i) condition on their given profile information; and (ii) information about the person they are talking to, resulting in improved dialogues, as measured by next utterance prediction. Since (ii) is initially unknown our model is trained to engage its partner with personal topics, and we show the resulting dialogue can be used to predict profile information about the interlocutors”.

With the aim to break the shell of chatbots and unleash their true nature the researchers have worked on ‘Persona-chat dataset’. The Persona-chat is the ‘crowdsourced’ data set consisting of 1155 personas along with minimum 5 profile sentences. In the data set each pair of speakers delivers their statement in the confined profile. The AI team of Facebook used this data set to overcome the failure of ‘consistent personality’ in the chatbots. Below is example of a persona-chat data from the Facebook’s research paper over the chatbots.

[PERSON 1:] Hi

[PERSON 2:] Hello ! How are you today ?

[PERSON 1:] I am good thank you , how are you.

[PERSON 2:] Great, thanks ! My children and I were just about to watch Game of Thrones.

[PERSON 1:] Nice ! How old are your children?

[PERSON 2:] I have four that range in age from 10 to 21. You?

[PERSON 1:] I do not have children at the moment.

[PERSON 2:] That just means you get to keep all the popcorn for yourself.

[PERSON 1:] And Cheetos at the moment!

[PERSON 2:] Good choice. Do you watch Game of Thrones?

[PERSON 1:] No, I do not have much time for TV.

[PERSON 2:] I usually spend my time painting: but, I love the show

This could be a revolutionary change and will redefine the chatbots usability forever. However, until Facebook actually implements the idea into a chatbot for the users we all are stuck with the chatting tools responding with famous movie dialogues.