Reinforcement learning for chatbots

A system comprising:

• a processor;
• and
• a computer-readable storage medium communicatively coupled to the processor and storing program instructions which, when executed by the processor, are configured to cause the processor to:
• select a plurality of conversations, wherein each conversation includes an agent and a user;
• identify, in each of the plurality of conversations, a set of turns and on or more topics;
• associate the one or more topics to each turn of the set of turns;
• generate, based on the set of turns, a conversation flow for each conversation, wherein the conversation flow identifies a sequence of the one or more topics;
• apply an outcome score to each conversation;
• create a reinforced learning (RL) model, wherein the RL model includes a Markov chain and wherein the RL model is based on the conversation flow of each conversation and the outcome score of each conversation;
• deploy the RL model, wherein the deploying includes sending the RL model to a chatbot and the RL model is configured to maximize a future outcome score for a future conversation;
• initiate, in response to updating the RL model and by the chatbot, a new conversation;
• develop, based on the RL model and based on a set of topics in the new conversation, a new conversation sequence;
• and
• complete, by the chatbot, one or more tasks, wherein the one or more tasks are based on the set of topics in the new conversation.

A computer program product, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processing unit to cause the processing unit to:

• select a plurality of conversations, wherein each conversation includes an agent and a user;
• identify, in each of the plurality of conversations, a set of turns and on or more topics;
• associate the one or more topics to each turn of the set of turns;
• generate, based on the set of turns, a conversation flow for each conversation, wherein the conversation flow identifies a sequence of the one or more topics;
• apply an outcome score to each conversation;
• create a reinforced learning (RL) model, wherein the RL model includes a Markov chain and wherein the RL model is based on the conversation flow of each conversation and the outcome score of each conversation;
• deploy the RL model, wherein the deploying includes sending the RL model to a chatbot and the RL model is configured to maximize a future outcome score for a future conversation;
• initiate, in response to updating the RL model and by the chatbot, a new conversation;
• develop, based on the RL model and based on a set of topics in the new conversation, a new conversation sequence;
• and
• complete, by the chatbot, one or more tasks, wherein the one or more tasks are based on the set of topics in the new conversation.