Artificial intelligence (AI) has the potential to disrupt virtually every industry, transform everyday operations, and touch billions of lives. Why then does it seem that most companies are only scratching the surface of what AI can do? Andrew Ng, founding lead of the Google Brain team, says, “AI will transform many industries. But it’s not magic. Surprisingly, despite AI’s breadth of impact, the types of it being deployed are still extremely limited.
Almost all of AI’s recent progress is through one type, in which some input data (A) is used to quickly generate some simple response (B).” The technical term for building this A→B software is “supervised learning.” Today’s supervised learning requires a great deal of data and AI talent, both of which have proven hard to come by. As Ng tells us, it’s important to understand what AI can and can’t do before you can incorporate it into your business strategy and truly put it to work.
What AI can do right now: task-based automation
Sentiment analysis in a comments section, scanning security footage for suspicious behavior, recognizing objects in an image… these are some examples of this A→B relationship. AI includes tasks such as learning, reasoning, perception, language understanding, and planning, meaning it can automate simple mental tasks — no longer relegating them solely to humans.
The role of human intelligence in AI is largely choosing A and B, delegating the task to be automated based on context (e.g., search for balloons in an image of a child’s birthday party). AI is one of the driving forces behind next-gen ecommerce, like Pinterest’s buyable Pins, Curalate’s Intelligent Product Tagging, or Instagram’s shoppable tags on photos. These brands have optimized for commerce, not just content. AI isn’t a magic bullet; it’s the capability of a machine to simulate intelligent human behaviors — exactly that simple, and that elusive.
How crowdsourcing can accomplish AI initiatives
The two most necessary pieces of the AI puzzle are those hardest to find: massive amounts of data and qualified AI talent. With crowdsourcing, you can tap into thousands of data scientists and developers trained in today’s must-have cognitive skills. Yesterday you had to write everything yourself; today we have open source. Better yet, we have platforms. Salesforce has AI, Microsoft has one… IBM has Watson. Crowdsourcing makes all of this more accessible and effective. At its core, crowdsourcing is having a large enough crowd of people with a favorite tool; they all take one swing, and if it connects, they stick around. If it doesn’t, they walk away. It’s an incredible filtering event, after which you’re left only with the experts you need.
Code challenges vs. sprints and ideation
With code challenges, competitors come back with solutions we can study. There are 3 possible outcomes:
- You get exactly what you needed — a clue, inspiration, or validation of a hypothesis.
- You can mine the result for more information or derive some value from it.
- The result shows promise, but it’s not quite there yet…
A code challenge is appropriate when you’re pretty sure something can be done with the data. But in the case of that third outcome, a sprint or ideation is ideal. It’s a smart option when there’s less certainty — when you have some data and you want to ask questions, but you’re not quite sure which ones to ask. Crowdsourcing makes it possible to get real results from starting points such as: “What do I need?” and “What tools should I look for?” By leveraging the Topcoder Community, you can end up with 5 or even 10 functional solutions — faster and more efficiently than before.
What can you accomplish with Topcoder? Reach out and get started on your AI initiatives today.