Page 20 of 43« First...1018192021223040...Last »

Speaking in tongues: how the crowd is transforming translation

April 12th, 2011 by

microtask_translations_tonguesEver since the first letter was chiseled out of a stone tablet, people have needed to translate text into other languages.

And while keyboards are somewhat faster than chisels, the hassle involved in translating text hasn’t changed much since Moses tried to tweet the Ten Commandments: you have to find a translator, negotiate the price, send the text, and then wait a few days for it to come back.

Because this process is both expensive and time consuming, we’re used to avoiding it. Usually we only translate documents when we have to, such as an application for a foreign visa or an instruction manual for a product to be sold overseas.

Machine translation, such as Google Translate, has been around for a few years, but quality is still an issue. It is a great tool when you personally need to understand some foreign text, but it falls far short of the standard required for publishing.

Enter the crowd
In the last year translations have entered the era of crowdsourcing, transforming the industry profoundly. There are now numerous crowdsourced translation providers, each with thousands of translators on call. They provide native speakers, professional translators, or a mix of the two.

This has changed the industry in several key ways. The first is time. Through these networks you can always get a translator instantly. You can’t pick who does the work (this is the idea of crowdsourcing after all…), but you can choose if you want the work to be done by a native speaker or a certified translator. (Most of the time all you really care about is that a qualified translator translates the text.) Without the normal administrative hassle and waiting, turnaround time is only as long as it takes the individual to translate the text, which is about 200 words per hour for a professional translator.

The second change is cost. With translators working from home via the internet, overheads plunge. This makes it cheaper of course, but also means you can translate short pieces of text, or even single words. Short sentences introduce a problem with potential ambiguity when the translator is not familiar with the product, but all the translator networks allow you to submit additional instructions that help the translator understand what it is that you wish to say.

API: A fan of man
But perhaps the most significant change is that these networks can now be controlled through an application programming interface (API). This means that the ordering of translations can be fully automated, while the actual work is performed by real human beings. This opens up a whole new world of innovation.

Take PremiumFanPage (PFP) as a completely random example (it is pure coincidence that I am CEO and founder). Using an API and an army of crowdsourced translators around the world, it can automatically read text from social media streams, translate it, and make it available online.

This means, for example, that you can write a tweet in English at any time of the day, knowing that the tweet will also appear in Japanese and French a few minutes later. Traditionally this would involve a lot of tedious work per tweet, but PFP can automate the entire process. The efficiency, low overhead costs and quick turnaround time mean that even large quantities of small tasks are now economic.

Still unsure? Talk to a professional
As with anything, some people are skeptical of crowdsourced translations. Some see crowdsourcing as a hobby, rather than a professional service. This probably stems from Facebook’s early success with the concept, where anyone could help translate the site’s user interface.

But the Facebook-model is only one form of crowdsourced translation. As mentioned above, there are now a number of networks of professionals who are paid to do small individual tasks. Because you pay for the work you can expect professional quality in return.

Another criticism is that pay is too low to attract qualified translators. In some cases this is valid, but we need to keep in mind that the industry is very young. There is pressure to increase the levels of pay, and at the same time new and innovative business applications may help justify higher price brackets for those who require the highest quality work.

Despite these concerns, the industry is developing rapidly. Getting tweets automatically translated to a number of languages within minutes of posting them is very exciting, but I believe we are just at the beginning of a whole new era. In years to come, the language will no longer be a barrier of entry to new markets. We are one huge step closer to a world where everyone can understand each other, and participate in a global marketplace.


Too cool to play: the gamification backlash

April 11th, 2011 by

microtask_gamification_backlashIn high school life was simple (brutal, but simple): popularity equaled coolness. As my math teacher would say, the relationship had both correlation and dependence (for some reason A marks in statistics class failed to improve my social status).

Post-graduation, the world sometimes seems to work in reverse. Take Apple. In 2000, Mac was the hip, underrated indie-kid of computing. Now everyone’s got an iPhone and suddenly Apple is an evil corporate giant out to steal your money, digital rights and free will (if you believe the Guardian).

People are instinctively suspicious of stuff that gets too big, too fast. Two years ago the term “gamification” barely existed, now it’s everywhere. There are gamification books, startups and even a dedicated gamification summit. With so much hype, backlash is inevitable.

Missing the points?
Here at Microtask we’re “out and proud” gaming fans. We play games, we write about games, we make games. However, being a mature and friendly organization, we realize that sometimes it’s good to check in with the opposition (even if it’s just so you know where to aim). Happily for us, critics of gamification rarely lay into full-blown games like Digitalkoot. Instead they tend to attack companies like Foursquare, who apply game-mechanics to non-game sites and apps.

Counter-intuitively, game designers are often the biggest gamification skeptics. Designers argue that game mechanics are subtle and complex. Good games, they say, are like art: engaging players deeply and emotionally (not sure how Grand Theft Auto fits in here). In contrast gamification, as game developer Margaret Robertson puts it, “tricks people into believing that there’s a simple way to imbue their thing (gym, job, genital health outreach program, etc) with the psychological, emotional and social power of a great game.” In other words, it takes more than points and a leader-board to turn a banking website into the next Farmville.

Master class
It’s easy to see why game designers look down on gamification. Modern console games are hugely complex, multi-layered virtual realities. Gamified apps are often just primitive reward-systems with flashing scores and badges. But what if designers stopped sniping from the sidelines and instead helped to “level-up” gamification?

In a recent Google Tech Talk designer and self-confessed “grumpy German scholar” Sebastian Deterding outlines how he believes companies could “gamify better”. His presentation really stands out from the general gamification hype/hate. Deterding identifies three key ingredients to successful games: meaning, mastery and autonomy. Meaningful games are ones which connect deeply with users. It could be via compelling narrative (only you can save mankind) or via a game’s social value (only you can save the Finnish library archive). Mastery is about giving players a sense of control and well-defined game progression. Autonomy is basically the art of making games enough fun so that people actually choose to play them.

For all the backlash, gamification is still the popular new kid in the virtual playground. As critics have (endlessly) pointed out, one-size-fits-all gamification is a sure-fire way to frustrate users and lose money. But equally, personalized well-designed gamification is capable of achieving exactly the opposite. Gamification is here to stay, the question is whether game designers will swallow their pride and help us gamify better.


Lost in the Virtual Economy? Here’s a map

April 7th, 2011 by

microtask_virtual_economiesEvery day we hear more about goldfarming, crowdsourcing, distributed work and other areas of the so-called “virtual economy”.

As with any industry in its infancy, exactly what is meant by these terms is unclear, and rapidly evolving. (If you find it slightly confusing, don’t worry: it probably just shows you have been paying attention.)

This is why we are so excited about the new study Knowledge Map of the Virtual Economy by Vili Lehdonvirta (researcher at the University of Tokyo and occasional guest writer on this blog) and his colleague Mirko Ernkvist.

Along with giving us something to blog about, the aim of the study is to provide an updated picture of the scale and development of the virtual economy. It focuses on its economic impact, business models and value chains. The two major areas of the virtual economy identified in the report are third party gaming services and, more interesting for us, the role of microwork in virtual worlds.

The bottleneck: translating problems into microtasks
The report explains that, although microwork itself requires no technological expertise, converting computational and business problems into microtasks requires a great deal. This is well known to our tech-experts: Digitalkoot, the first Microtask-powered service, combines advanced OCR techniques, human recognition skills and game mechanics to seamlessly distribute work to end users. What we are really proud of is that the crowd does the work by playing games. (Let’s call it “gamesourcing”, just to confuse everyone even more.)

Thankfully, the report also considers what some of these new terms actually mean. Usually the word “crowdsourcing” is used interchangeably with “microwork”. According to Lehdonvirta and Ernkvist, they are in fact distinct: Crowdsourcing entails outsourcing tasks traditionally performed by employees or contractors to a large group of people through the internet. Microwork on the other hand tends to involve breaking the work down to suitably sized microtasks, integrating quality assurance into the process, and recombining the completed microtasks into a final deliverable. This is a view that we certainly share, and a good reminder of how powerful words are: expect more Microtask’s tweets with the “microwork” hashtag from now on.

Back to the future: appetite for disruption
The report goes on to argue that “Microwork is today emerging as a separate concept from crowdsourcing, rather than a subset of it [...] The innovation in microwork is the transformation of information work into microsized units, similarly to how Taylorism and scientific management transformed manufacturing work in the late 19th century.” (Again, it is a point we agree with: for further discussion of the historical roots of distributed work see our earlier post).

Although it may be tempting to question whether this is a step forward or a step back, it is important to consider the new opportunities. Tasks that are not economically feasible using traditional work practices will soon (if not already) be possible using microwork. Imagine a long-time professional photographer wanting to sell his pictures online. He would have to tag thousands of shots, consuming weeks if not months of his precious time. Soon he will be able to submit his pictures to a microwork service and let the crowd work its magic, shrinking weeks and months into mere hours.

The report goes on to discuss some of the key problems the industry must overcome. These include payment channels (especially when using workforce from other countries) and legislation relating to worker protection (are microworkers employees or independent contractors? How to ensure there is no child labor involved?). Despite these issues, the report considers the potential of the mechanism self-evident.

A final message to take from the report is a call to action: if the upgrading strategies mentioned in the study (coordination of workers, and building capacities to supply and grow microwork in developing countries) are tackled quickly, it suggests that the whole market could be worth several billion dollars within the next five years, as the technology matures. Not bad for tasks that you can complete in just a couple of seconds (whatever they end up being called).


Searching Questions: how good is Google?

April 6th, 2011 by

microtask_google_goodDo you remember the first time you used Google? The white screen, the goofy logo, the incomprehensible “I’m feeling lucky” button.

What did you search for? (if you were a teenage boy, don’t answer that). In the bad old days of dial-up, Google was a (relatively) fast, flashing and seemingly benevolent oasis of free information.

These days Google is big. Really big. Founders Larry Page and Sergey Brin are both multi-billionaires. 64% of all web searches start beneath those famous letters. To millions of users “Google” and “the web” are virtually interchangeable terms.

But is Google’s virtual monopoly good for the internet? In his book The Googalization of Everything, media expert Siva Vaidhyanathan asks if it’s healthy for one largely unregulated corporation to dominate online.

A big friendly giant?
Google is much more than just a search engine. You can do a Google search for a Google Map on a Google browser using a Google phone while being driven by a Google car (spot the odd one out). Prof. Vaidhyanathan’s problem (well, one of his problems: it’s a 300 page book) is that as Google grows, the company is becoming less about “organizing the sum total of human knowledge” and more about shopping.

The professor argues that Google products and searches increasingly favor “the recent over the classic, the local over the global, the personal over the universal.” A search engine focused on “now”, “here” and “me” is great if you want to find the best iPad 2 deals or hook up with your friendly local crowdsourcing startup. But less useful if you need to search complex scientific data sets or track down news headlines from twenty years ago.

Keep your Google friends close
As if to confirm Prof. Vaidhyanathan’s worst fears, Google has just launched a new social networking style +1 button. This innocent-looking icon allows users to tag web pages and recommend them to their “Google friends”. Eventually, Google may use +1 data to personalize user searches. So if all your friends Like, sorry, +1 the same site it’ll get ranked higher in your search results.

Unlike Prof. Vaidhyanathan I can see the appeal of personalized, localized crowd-based search. Friends are less likely to spam you with fake results. Friends can be trusted to provide you with reliable information (well usually, unless you’re asking things like: “do I look fat in this?” or “do I really need another beer?”)

Genius +1
The question is: will friend-centered search teach people anything they don’t already know? Prof. Vaidhyanathan doesn’t think so. He predicts that “the knowledge web” will eventually get taken over by smaller specialist engines like Wolfram Alpha.

It’s a depressing vision. But what if Google could combine the trustworthiness of friend groups with the depth of crowdsourced expert knowledge?

Crowdsourced Q&A services Quora and Stack Overflow already tap into a community of experts who provide reliable answers to user questions. Answers are voted up or down by the community. On Stack Overflow, experts can also build their reputations with awards, points and badges.

Imagine if Google +1 incorporated elements of this kind of crowdsourcing. Specialists could +1 articles and sites in their subject area and make them visible to the whole web. An expert cardiologist could rank his favorite heart transplant journals. Siva Vaidhyanathan could flag his all time top corporate-responsibility papers. (Of course, Google would have to work out a way of registering and verifying the experts to prevent spammers.)

The web is a collection of human knowledge – global and local, personal and universal. A search engine that was as reliable as a close friend and as knowledgeable as a Harvard professor would be awesome. But while we wait I’m sticking with Google.


Science and the “Nobel” art of gaming

March 31st, 2011 by

microtask_biotic_gamesAs regular readers will know, here at Microtask we love a bit of science fun. Back in November we blogged about Foldit, a freely-available online protein-folding game. Foldit players contribute directly to scientific discovery: the more proteins they fold, the closer scientists get to curing diseases like Alzheimer’s and AIDS.

Refusing to be out-innovated by mere protein professors, geneticists at Carnegie Mellon and Stanford University have created EteRNA. EteRNA is another folding game: this time the goal is to help create the first large-scale library of synthetic RNA designs (which sounds like you’re studying for a PhD as opposed to playing a puzzle game).

It doesn’t take a genius to see that gamification is getting a totally new crowd interested in scientific experiments. So what’s the next step? Instead of creating games that model biological processes, what if scientists started “playing” with actual molecules?

Microscopic evolution, real life fun
That’s exactly what Stanford researcher Ingmar Riedel-Kruse and his lab group have done. The Biotic Games project enables players to interact directly with microorganisms. The game’s “hardware” is a simple console which is hooked up to a lab slide. When players push buttons on the console the microorganisms on the slide react. These reactions are displayed onscreen in real-time via a microscopic camera.

So far, the team has come up with several varieties of protozoa-based fun, such as:
PAC Mecium allows players to move paramecia around by controlling electrical fields. Biotic Pinball lets you influence the direction of microorganisms by injecting chemicals onto the lab slide. Polymer Race has players betting on the fastest reactions among millions of copies of an organism’s DNA.

Like many great science projects, Biotic Games was created out of intellectual curiosity (the “let’s see what happens if we try this” impulse) rather than for any specific application. Riedel-Kruse hopes that:

“By playing games involving biology on a scale too small to see with the naked eye, people will realize how amazing these processes are… We are talking about microbiology with these games, very primitive life forms.”

At this point, the actual game-play of Biotic Games is still very basic (plus players need an inconveniently expensive biotechnology lab to run the system). However, with a bit of development (and maybe a decent game designer), there’s definitely potential for crowdsourced “biotic tasks”.

Console campaigners
Medicine and science are not the only fields that could benefit from gamified crowdsourcing. Back in 2008 Steve Puma showcased Macrocosm, the “climate change simulation game that lets you save the earth”. Puma’s concept was a game that allowed users and companies to work together to solve complex climate change and social problems.

Developers behind the recently released game Fate of the World would definitely agree that climate change is ideal gamification material. A single-player strategy game, Fate of The World encourages players to explore serious political and social issues. Unlike many “message-focused” platforms, Fate of the World is a game you actually want to play: complex, immersive and visually appealing.

Both Biotic Games and Fate of the World have huge crowdsourcing potential, especially if they borrow a few ideas from each other. Fate of the World could develop ways to interact with real world processes – allowing the crowd to suggest concrete solutions to global problems. Biotic Games could adopt more high-spec, mainstream game-play in order to attract a bigger crowd to medical research tasks.

It’s still very early days for scientific crowdsourced gamification. But the beauty of science is that a sudden, critical discovery is always possible. Who knows, one day there really might be a Nobel prize winning crowd.


Page 20 of 43« First...1018192021223040...Last »