However, once trained, the generative models outperform the retrieval-based models in terms of handling previously unseen queries and create an impression of talking with a human (a toddler may be) for the user. These models are difficult to train, as they need to learn the proper sentence structure by themselves. Due to this, the responses generated are prone to grammatical errors. They generate a response, word by word based on the query. The Generative models are quite intelligent. It does not generate any new sentences, hence we don’t need to worry about grammar. The Retrieval-based models pick a response from a collection of responses based on the query. The Intelligent models can be further classified into: Let us call these models that automatically learn from data, Intelligent models. What if we can build a bot that learns from existing conversations (between humans). Also, it is time consuming and takes a lot of effort to write the rules manually. The pattern matching is kind of weak and hence, AIML based bots suffer when they encounter a sentence that doesn’t contain any known patterns. But it is incredibly difficult to create a bot that answers complex queries. Rule based models make it easy for anyone to create a bot. Let us call this model of bots, Rule based model. Basically, you write a PATTERN and a TEMPLATE, such that when the bot encounters that pattern in a sentence from user, it replies with one of the templates. It is written in AIML (Artificial Intelligence Markup Language) an XML based “language” that lets developers write rules for the bot to follow. Then I encountered another bot, Mitsuku which seemed quite intelligent. But it was fun to see people willingly interact with a program that I’ve created. The program uses the linux utility fortune, a pseudorandom message generator. I created Myshkin in nodejs that answers any query with a quote.
IN THE DEEP MOVIE 2016 WTF SOFTWARE
We (Free Software Community) created a group for interacting with the bots we built. Immediately people started creating abstractions in nodejs, ruby and python, for building bots. Last year, Telegram released its bot API, providing an easy way for developers, to create bots by interacting with a bot, the Bot Father. Part II of Sequence to Sequence Learning is available - Practical seq2seq
0 kommentar(er)
