If the bot answers a question logically or solves a given task, it should be considered smart. A lot of personal assistants like Apple Siri, Amazon Alexa and Google Home can be regarded as intelligent. But can we say these assistant bots are Turing smart? Well, when we cannot distinguish between a bot and a human, then the bot can be called Turing smart.
To check whether a bot is human level intelligent or not, a lot of ways are there, and the most prominent one is the Turing test. The Turing test, developed by Alan Turing in 1950, is a test of a machine’s ability to exhibit intelligent behavior equivalent to, or indistinguishable from, that of a human. Turing proposed that a human evaluator would judge natural language conversations between a human and a machine designed to generate human-like responses. The evaluator would be aware that one of the two partners in conversation is a machine, and all participants would be separated from one another. Interactions would be limited to a text-only channel such as a computer keyboard and screen so the result would not depend on the machine’s ability to render words as speech. If the evaluator cannot reliably tell the machine from the human, the machine is said to have passed the test. The test does not check the ability to give correct answers to questions, only how closely answers resemble those a human would give. 
To know more about the test and the bots that have won it in the past, check my previous blog.
How close are chatbots to pass Turing test?
The question is same since the term AI was coined at a conference at Dartmouth College in 1956 : Can a machine be ever…
The question is how to make the bot smart enough that it compares to Human Intelligence. There are mainly two ways of doing so:
- Rule-Based Approach: In a rule-based approach, a bot answers questions based on some rules on which it is trained on. The rules defined can be very simple to very complex. The creation of these bots are relatively straightforward using some rule-based approach, but the bot is not efficient in answering questions, whose pattern does not match with the rules on which the bot is trained. One of such languages is AIML (Artificial Intelligence Markup Language): a language based on XML that lets developers write rules for the bot to follow. Also, writing rules for different scenarios is very time taking and it is impossible to write rules for every possible scenario. The bots can handle simple queries but fail to manage complex queries. Hence, the bot can never pass the Turing test if based on some rule-based models.
2. Self learn-able bots: These are the bots that use some Machine Learning-based approaches that make them more efficient than rule-based bots. These bots can be of two types:
2.1 Retrieval based models: These bots are trained on a set of questions and their possible outcomes. For every question, the bot can find the most relevant answers from the sets of all possible answers and then outputs the answer. Although, the bot cannot generate new answers if trained on a lot of question and answers dataset, and if the data set is pre-processed smartly, the bot can handle queries fairly good. The complexity can range from simple rules for a query to complex rules using some machine learning algorithm to find the most appropriate answer. Also, there is no issue with the language and grammar as the answers are pre-determined and it cannot go wrong in syntax manner.
2.2 Generative models: Generative models are better than rule-based models in a way better that they can generate the answers and not always replies with one of the answers from a set of answers. This makes them more intelligent as they take word by word from the query and generates the answers. It also makes them more prone to errors as they need to take the spelling and grammar into account. To make them better at handling these errors, these models need to be trained more precisely. Once trained, they outperform the rule-based models as they can answer complex and unseen queries. Language translation models can be used in creating such a model.
Check my next blog for creating a sequence to sequence learning model that generates its answers.
Generative Model Chatbots
As discussed in my previous post about the types of bots and it seemed that the generative bots are the smartest…
We’re a team of bot creatives and AI scientists with one common goal:
blowing business objectives out of the water with bots and cognitive solutions