Of late, Chatbots have enhanced a lot, all thanks to leveraging the power of machine learning along with the neutral networks. However, even today, most of them working on specific situations are not seen working on the confabulate on a different range of topics. This could change in the chatbot introduced by Google called Meena. In a recent paper published in the technical journal called Towards a Human-Like Open Domain Chatbot, the tech giant claimed that it has designed a chatbot that is superior when compared to others. It has been designed after exploring 341GB of social-media communications along with checking not less than 2.6 billion parameters claimed the official of Tech giant.
It further added stating that Meena Google’s Chatbot can carry out sensible communication and specific along with current state of art chatbots claimed the researchers. With the traditional metric required for calculating the realism of Chatbots that are seen under the way of Turing Test, however, over the years, the Chinese Room Argument has shown several disadvantages of the test. Apparently, Meena was later tested on factors like Sensibleness and Specificity Average (SSA), which happens to be a new metric developed by the Tech Giant that catches the basic, but at the same time, the vital attributes for the human conversation at the background of the ensuing conversation.
With all the parameters coming on the board, Meena was reported scoring high with 79 percent of the new metric. This could score and place the new chatbot of Google found in between the venerable Mitsuko along with scoring 56 percent and the human conversing partner that seemed to have scored 86 percent on the same test. However, Google is not at all in the mood of releasing the demo of Meena at the moment in the public domain but the company assures about the factors like safety and potential biases to the test.
Once all the checks pass all the go, the tech giant would then release a demo video in the coming months only with the help of advanced research methods carried out over the neural conversation models and chatbots at one go. The company also has claimed that they want to keep things focussed on sensibilities along with the specifications in this project. At the same time, the key features including factuality and personality would be kept as a target for the following projects. Well, this seems interesting, there is much more to come on this, stay tuned.