Algorithm-Driven Innovation

Algorithm-Driven InnovationCould an algorithm generate innovation? Christopher Steiner poses a similar question in a Wall Street Journal article titled “Automatons Get Creative.” Steiner posits that automation is slowly making its way into creative fields that previously were viewed as immune to machine-generated content. After all, the very essence of a creative individual is the fact that such a person can identify new pathways, whereas a machine must be programmed in advance to follow a specific set of guidelines. Imagination is typically the opposite of automation, but as Steiner points out, perhaps “[c]omputers can be creative after all.”

Steiner focuses on the power of the algorithm, which takes a series of inputs and follows a decision tree to generate certain outputs. With the increased use of Big Data analytics and incredibly fast computers, algorithms can process more and more data and, presumably generate more interesting output. Steiner cites several examples in his analysis of the state of the art in computer algorithms performing creative-like tasks, ranging from tools that predict the popularity of a certain piece of music, to quantitative tools that anticipate which movie scripts will deliver the greatest box office success, to programs that grade student essays, to programs that automatically generate news reports about football games using quantitative data captured during the game.

For the student of innovation, the increasing prevalence of powerful algorithms operating in the creative fields raises the question of whether such tools could be used in the innovation process. At first glance, algorithms could be of value to the innovation process in the ideation process and in innovation scoring or assessment. The ideation phase, when innovators work to define a new set of approaches to solve a particular problem, could benefit from the assistance of a more powerful tool to gather and assess the different aspects of an innovation theme. For instance, the innovator could create an algorithm that scanned certain websites to perform basic word association activities to look for connections that the innovator or workshop participants might have missed. Similar to the visual word clouds or tag clouds, an innovation ideation algorithm could identify patterns and connections between concepts that a human might not see. Moreover, the algorithm could scan through infinitely larger data sets than could a human, casting a broader net.

Although there is debate in the innovation arena concerning the value of a shotgun approach versus a more focused and precise ideation strategy, there is definitely value in being able to process a large quantity of data in the case where an innovator has done work to define a specific innovation theme or area of investigation. The innovation scoring or assessment process could also benefit from the incorporation of more rigorous, and less subjective, methods. Just like the movie script algorithm, an innovation scoring algorithm could assess an innovation concept against a large data set of known innovation successes and post a likelihood of success, providing an additional data point for the practitioner in determining whether to proceed with a given approach.

Steiner concludes with an interesting anecdote about the power of algorithms, citing the example of the political science professor Bruce Bueno de Mesquita and his work with the CIA in political and military predictions. Dr. Bueno de Mesquita’s game-theory inspired algorithms have been used to assess over 1,700 different political and military scenarios and his approach has predicted the outcomes successfully at double the rate of the CIA’s internal analysts. The advantage of the computer model, the professor notes, is that the CIA analysts are focused more on “personal back stories and gossip that will have no effect on the future,” whereas the machine just focuses on the data. This is an occurrence that many of us have witnessed in innovation workshops, where either the facilitator (I am guilty of this) or participants will focus on one idea or an aspect of an idea to the detriment of the overall process. Perhaps when we facilitate innovation sessions we should step back and think more like an algorithm in terms of processing the inputs we are generating and receiving from the room.

Source:
Christopher Steiner, “Automatons Get Creative,” Wall Street Journal (August 18, 2012).

image credit: gorkieryluk.com

Join the global innovation community

Don’t miss an article (4,600+) – Subscribe to our RSS feed and join our Innovation Excellence group!


Scott BowdenScott Bowden works on Innovation Programs for IBM Global Services.

This entry was posted in Innovation, People & Skills, Processes & Tools and tagged . Bookmark the permalink.

2 Responses to Algorithm-Driven Innovation

  1. Some stuff cannot be automated.
    Innovation is probably one of them.
    Generating trust among people is another one.

  2. Pingback: Innovation Excellence | Innovation Quotes of the Week – September 2, 2012

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Keep Up to Date

  • FeedBurner
  • LinkedIn
  • Twitter
  • Facebook
  • Slideshare
  • Email
  • YouTube
  • IPhone
  • Amazon Kindle
  • Stumble Upon

Innovation Authors - Braden Kelley, Julie Anixter and Rowan Gibson

Your hosts, Braden Kelley, Julie Anixter and Rowan Gibson, are innovation writers, speakers and strategic advisors to many of the world’s leading companies.

“Our mission is to help you achieve innovation excellence inside your own organization by making innovation resources, answers, and best practices accessible for the greater good.”