2,960 crowdsourcing and crowdfunding sites
Editor's Note: The following post by Clinton Bonner was originally published on the TopCoder blog and is reposted here with their permission.
Algorithms: you can’t see them, touch them, or if you cared to, taste them, but you know they are everywhere. Algorithms are the O2 of the technology blood stream. Without them, networks die, information ceases to exchange, and the pace of innovation slows to a crawl. Algorithms are what make your favorite apps so personalized and what deliver exceedingly larger sets of Big Data to clouds to speed up computational decisions that have legitimate impacts on humanity. Even companies who traditionally wouldn’t consider themselves a “data company” are increasingly spending time and resources to prep data, mine data, understand data and pair otherwise non-related data sets in the hopes of mashing-up a solution that pays dividends for their enterprise. All of this data work is driven by sophisticated algorithms. So the question that matters to your enterprise is: How do you approach algorithmic innovation? Here are three compelling reasons you should consider an Open Innovation approach to tackle your biggest algorithmic challenges.
Open Innovation, especially EOI (Enterprise Open Innovation), takes that raw power that is promised by Crowdsourcing and applies process. This process equates to gaining scale and this means you can have repeatable, and even predictable success in algorithmic innovation and advancement via Open Innovation. Why is this the case? Well, let’s get back to the core of Open Innovation. Two powerful factors converge inside an Open Innovation community such as TopCoder; they are competition and collaboration.
TopCoder Algorithm challenges often take place in the shape of what we call our Marathon Matches. These are long form challenges dealing with highly complex and sophisticated work, typically spanning two weeks in competition time. Our Marathon Matches attract hundreds of registrants and competitors and when you have that many solid algorithmists focused on a problem and bringing their unique experiences and perspective to their submission(s), the results are staggering.
Remember in a competitive Open Innovation community, you the end-user of the output, are after one thing, a truly valuable outcome. Repeatable, rigorous, standards-based algorithm challenges hosted on a mature Open Innovation platform, powered by a global community of solvers is the most predictable way to attain extreme, repeatable successes.
Fans of the culinary viewing delights Iron Chef, Top Chef, or even Chopped – all based on competition we should add – will know the term “Two-Ways” straight away. It’s the same main element, let’s say Duck, served two unique ways within a singular plating. Here, we are talking about a double dose of what Maverick & Goose were continually chasing in the iconic ’80s flick Top Gun. We are of course referring to the need, the need for speed.
The first “speed” benefit is the obvious one. At TopCoder, as mentioned above, we host Marathon Matches that create stellar algorithmic solutions and typically they take place in or under a two week time frame. This is revolutionary in the arena of sophisticated algorithms. As an example, for a Harvard Medical School competition focused on DNA sequencing alignment, the TopCoder winning solutions were delivered in two weeks time. The existing gold standard that was in-use was developed by a Full-Time resource, working on the problem for a full year. The TopCoder solution was delivered 96.2% faster than a traditional avenue could produce. But how did the algorithm perform? That is the second “speed” we are about to discus.
Whether we point to the growth and fragmentation of mobile & tablet or the shift caused by all things “social”, we talk about a hyper-paced technology environment quite often. When we discuss “pace,” there is no greater factor than that of neo-data creation. In fact, we just canvassed this very topic in a recent blog – Why this infographic on data is astounding, terrifying, and revolutionary – and the simple truth is that data creation is greatly outpacing the ability to handle the data effectively, and the pace is accelerating!
In the Harvard Medical School example above; where the existing gold standard solution was able to calculate an accurate edit distance between a Query DNA string and the original DNA string in 400 seconds, TopCoder’s solution was able to do the same – at an improved level of accuracy – in only 12 seconds.
James Kobielus, Senior Program Director, Product Marketing, Big Data Analytics at IBM, recently penned a fantastic article canvassing the need to continuously and incrementally innovate in the era of Big Data. Granted, the above examples we just showcased are not incremental innovations, they are, big and staggering improvements over existing gold standard solutions. But James’ article has many stellar points, this being one of them:
James goes on to say that the key to these advancements is experimentation in data science, and again, his point is spot on. But how does your organization hope to continuously innovate in Big Data? To experiment like mad scientists in a traditional business environment, you need incredible resources, not to mention the pure man-power who can handle this highly specialized work. Fact is most have neither, very few have both. So what do you do?
You access platforms, you tap into global communities, and you run your data experiments – just as James is suggesting – but you do it through Open Innovation, powered by competitions. We discussed above the why, (Extreme value outcomes, speed two-ways), now it’s about how you execute.
To create value from advanced algorithms, you need faster, smarter solutions, delivered quicker on a scalable platform that allows you to experiment. Enterprise Open Innovation (EOI) is proving to be a remarkable way to do just that.