Previous algorithms for generating the end of a story tend to favour generic sentences, such as “They had a great time,” or “He was sad” – these final lines can be seen as bland. However, Alan Black, a professor in CMU’s Language Technologies Institute, said they aren’t necessarily worse than a non sequitur such as “The UFO came and took them all away.”

In a paper presented at the Second Workshop of Storytelling in Florence on Thursday (1 August), Black and students Prakhar Gupta, Vinayshekhar Bannihatti Kumar and Mukul Bhutani showcased a model for generating endings that will be both relevant to the story and diverse enough to be interesting.

According to Black, one trick to balancing these goals is to require the model to incorporate some key words into the ending that are related to those used early in the story. At the same time, the model is rewarded for using some rare words in the ending, in hopes of choosing an ending that is not totally predictable.

During the presentation, the team used this bot-generated story as an example: “Megan was new to the pageant world. In fact, this was her very first one. She was really enjoying herself but was also quite nervous. The results were in and she and the other contestants walked out.”

Existing algorithms generated these possible endings: “She was disappointed the she couldn’t have to learn how to win,” and “The next day, she was happy to have a new friend.” The CMU algorithm, on the other hand, produced the ending: “Megan won the pageant competition.”

Black acknowledged that none of the selections represent deathless prose, but the endings generated by the CMU model scored higher than the older models, both when scored automatically and by three human reviewers.

Although researchers have worked on conversational agents for some years, automated storytelling presents new technical challenges, he explained.

“In a conversation, the human’s questions and responses can help keep the computer’s responses on track,” Black said. “When the bot is telling a story, however, that means it has to remain coherent for far longer than it does in a conversation.”

Black also suggested that automated storytelling might be used for generating sub-stories in video games, or for generating stories that summarise presentations at a conference.

Another application might be to generate instructions for repairing something or using complicated equipment that can be customised to a user’s skill or knowledge level, or to the exact tools or equipment available to the user.

In May, a study commissioned by the charity BookTrust has suggested that an increasing number of parents are happy to replace books with digital technology when they read their children bedtime stories.

Previous algorithms for generating the end of a story tend to favour generic sentences, such as “They had a great time,” or “He was sad” – these final lines can be seen as bland. However, Alan Black, a professor in CMU’s Language Technologies Institute, said they aren’t necessarily worse than a non sequitur such as “The UFO came and took them all away.”

In a paper presented at the Second Workshop of Storytelling in Florence on Thursday (1 August), Black and students Prakhar Gupta, Vinayshekhar Bannihatti Kumar and Mukul Bhutani showcased a model for generating endings that will be both relevant to the story and diverse enough to be interesting.

According to Black, one trick to balancing these goals is to require the model to incorporate some key words into the ending that are related to those used early in the story. At the same time, the model is rewarded for using some rare words in the ending, in hopes of choosing an ending that is not totally predictable.

During the presentation, the team used this bot-generated story as an example: “Megan was new to the pageant world. In fact, this was her very first one. She was really enjoying herself but was also quite nervous. The results were in and she and the other contestants walked out.”

Existing algorithms generated these possible endings: “She was disappointed the she couldn’t have to learn how to win,” and “The next day, she was happy to have a new friend.” The CMU algorithm, on the other hand, produced the ending: “Megan won the pageant competition.”

Black acknowledged that none of the selections represent deathless prose, but the endings generated by the CMU model scored higher than the older models, both when scored automatically and by three human reviewers.

Although researchers have worked on conversational agents for some years, automated storytelling presents new technical challenges, he explained.

“In a conversation, the human’s questions and responses can help keep the computer’s responses on track,” Black said. “When the bot is telling a story, however, that means it has to remain coherent for far longer than it does in a conversation.”

Black also suggested that automated storytelling might be used for generating sub-stories in video games, or for generating stories that summarise presentations at a conference.

Another application might be to generate instructions for repairing something or using complicated equipment that can be customised to a user’s skill or knowledge level, or to the exact tools or equipment available to the user.

In May, a study commissioned by the charity BookTrust has suggested that an increasing number of parents are happy to replace books with digital technology when they read their children bedtime stories.