A recent report from Facebook tips that researchers from the Facebook Artificial Intelligence Research Lab(FAIR) created chatbots to converse with each other. The chatbots made up their own way of communicating using machine learning, which showed a marked deviation from human languages.
A distinct pattern that enables the chatbots to communicate freely using machine learning has been identified. These chatbots designed by the humans act independently and used their own unique language to emphasize their conversational strategies.The single, impulsive growth of an own language was the most thrilling development for the researchers, undoubtedly. The chatbots are also the evidences of how machines are redefining human’s understanding of multiple realms and domains, that were earlier believed to be solely dominated by humans and their languages.
Facebook’s chatbots are the clear indicators that predictions involved in machine learning research need to be more accurate. Till date, machine learning has generally entailed a fair amount of guesswork and speculation. Abundant data is fed to the neural network, followed by a detailed examination of the output. This gives an insight into the thinking pattern of the machine, though there is room for further exploration.
Facebook’s researchers wrote in their paper,
“There remains much potential for future work, particularly in exploring other reasoning strategies, and in improving the diversity of utterances without diverging from human language.”
Most of these studies have revealed that despite being designed by humans, these machines have their own non-human ways of communication. This clearly indicates that there are still ample areas of machine behavior that need to be investigated further.