george tziralis log

I attach the presentation file, full paper and abstract of the work I presented earlier today at APMS 2008, here in beautiful Espoo, Finland. Your comments are welcome.

Market Drive Innovation Management, from the inside in

View SlideShare presentation or Upload your own. (tags: tziralis predictionmarkets)

abstract: Modern enterprises and organizations seem like market-agnostic, semi-permeable bubbles, in a market-powered universe. Although markets external ly define companies’ operations, such mechanisms have yet no place in supporting the decisions of those entities. We attempt to address this reality by exploring the use and impact of an internal market mechanism to support decision-making in innovation management. We step up by focusing on a generic innovation’s management problem, the one of idea selection. An abstract model of an internal market is shaped, with contracts traded on selecting the idea that has the greatest potential for deployment. The questions that we ultimately address regard the properties of market’s convergence towards equilibrium. In our case, this translates to a meta-mechanism able to aggregate the totality of available information throughout knowledge workers and utilize it to highlight among the available ideas the one with the greatest potential for development. A market-approach to innovation management is rather novel; we believe that this text provides some unique contributions in this developing field, while introduces various options for further research.

full paper (.pdf)

Comments

Have you ever realized how many times you use Google search every single day? No? Get prepared to be amazed if you navigate to Google web history and click on ‘trends’. In my account, for example, Google has already tracked 23,025 searches, an average of about 45 queries per day. Sounds like a lot, right? And you may say it actually is, for a rudimentary service that has almost remained unchanged since its introduction 10 years ago and ‘just provides relevant results to search queries’. Well, let me argue that it’s not; at least for users being slightly smarter than dumb. Let me elaborate the reason.

Google, except from being relevant, is also fast. And speed is a matter of utter importance, if not the most significant one. Yes, it’s faster to search the web than your hard disk, enter arithmetic calculations or find word definitions into your browsers’ search box rather than opening a calculator or a dictionary, search for ‘wikipedia silvereye’ and feel lucky, instead of navigating to wikipedia’s homepage and finding your way from there; it keeps going. It should be by now apparent, Google turns out to be a launcher -over time and the more you use it-, next to its default search functionality, the recent Chrome browser being the most ubiquitous proof.

Turning back to my personal experiences -I bet they sound familiar- what I realize is that more and more often I enter longer queries and I don’t even click on any of the search results; Google web history comes again into support and reveals that this is the case for roughly half of my very recent queries. So, why is that, has Google become less relevant? No, not at all. It appears that I’m just looking for something else, which turns out to be different than any of the specific results. And that is their total number. The total number of results returned by a query in Google might look worthless, and it actually is when considered in an absolute sense. But, move that to a relative setting -aka comparing the results of two or more search queries- and yet you have a fantastic tool for a wide variety of purposes, powerful while elegant on its simplicity. Feeling unsafe about two potential spellings of a word? Run a query for both of them and select the one with the most results. Wanna pick a product and you don’t know which of the reviews to trust? Search for “productname good” and “productname bad” and find out how many people have written about the first versus the second, that will probably turn out to be a more sincere advice, at least worth of the time you spent for the ‘survey’.

Well, to be honest and while speaking of time spent, the concept is fantastic, but its implementation is not-so-fast. Needing to iterate the search process a couple of times or more and then remembering the specific number of results that Google returned each time is a process that could profoundly be improved. And the conception was common and clear in a recent Open Coffee Athens meeting, when my co-founder Efthimios Mpothos, our good friend Dimitris Athanasiadis and I were talking about how each one of us utilizes Google in daily practice: Why don’t we optimize the process by ourselves?

So, today we are happy enough to introduce you a tiny tool, result of limited effort and time, but potentially able to make you reconsider the value a couple of google queries can offer you. And while we cannot claim of reinventing the wheel (clearly, there are many wheels out there), we do think that our app is slick and simple, next to cute and time and attention efficient. It’s name? Well, each group of searches serves as an attempt to outsource a task, ranging from a quick spell checker task, to an elementary poll, to a basic sentiment analysis tool, or to whatever else you can imagine it being serving of. And, particularly, this is an attempt to outsource this task to Google, essentially to outsource it to the web-at-large. So, let it be named WebSource.it, and that’s the name of it.

Yes, by now you may navigate to http://websource.it, enter up to 5 terms separated by a Tab stroke and hit enter: What you’ll get is just the number of relevant Google search results, as provided by the Google API. And you may rest on the shoulders of the crowds to select the term with the most results, highlighted with green. Or, as an another uber-fast option for you “big-guys-who-don’t-click”, you may enter websource.it/search/term1/term2 to get your results directly, without leaving the address bar.

We hope it will be useful for you, at least as it was proved to be for us so far and we do await for your own very innovative uses.

Comments

We just added another option -facebook connect that is- to let practically anyone get into and take a look at our pride and joy, askmarkets. So, by now you can navigate to askmarkets.com, press the facebook button, enter your facebook credentials and -boom- you’re in, go trade, spread the word and let us know what you think askmarkets.

Comments

I was honored enough to be quoted among the former ATHOC President Gianna Angelopoulos-Daskalaki and the former Mayor of Athens and current Minister of Foreign Affairs Dora Bakoyannis, in a today’s article of the much respected Chicago Tribune, regarding the post-Games Athens (update: the article actually appeared at the newspaper’s front page).

To give you some more info on that, I have met the journalist Philip Hersh in Guatemala, at an IOC conference a year ago. Interestingly, we both attended the Sport Accord 2008 exhibition, held early June in Athens and I was at Chicago a month ago; however the interview was conducted via phone and I was asked to share my humble opinion about the status of the Olympic legacy of Athens, as a senior researcher of the Athens OGGI team.

After all, Mr. Hersh quotes me saying “I do think there are some opportunities that haven’t been used, and I don’t know if and when they will be” and [regarding the main Athens Olympic Sports complex] “I would like to see it as an Olympic Park, where people would use it, we in Athens don’t have a large park or one like [Chicago’s] Millennium Park” (well, clearly I pretty much liked the audio picnic option there, but this is another topic).

While I do not put into question the validity of quotes attributed to me, I feel that there was maybe a slightly biased and somehow disjointed citation of them, especially in the given context of the article’s corpus. I mean, in the phrase “Well, ‘A’ is true, but consider also ‘B’ and ‘C’, which clearly change the broader picture ‘D’.”, you may quote someone saying ‘A’, but this is only the partial truth, and not the full story.

I’m not willing to open the huge debate on 2004 Olympics and their legacy right here and now, nor I consider myself as the most appropriate one to handle or take part in it. However, if you want to get a much broader picture, you may go read this paper and this text (more and specific data are on their way to release) and shape your very own opinion. As for me, I clearly learned my lesson, concerning dealing with (old) media: If you don’t want to be quoted of saying ‘A’, then just don’t say it, in any context and collocation of your words. :)

Comments

Everything under the sun is on “public domain”. Thus, it might end up on the web, too. Get used to it, or stay at home.

a re-tweet, inspired by boris and mike.

please accept/correct/upgrade, or discard (I’m not the law guy)

Comments

cross posted at “the official askmarkets blog”

Askmarkets is an internet-based service for prediction markets, which consists of virtual marketplaces for information trading. Now, while this sentence might seem like an accurate description to our biased eyes, we bet that you can find a much better pitch line and submit it in the comments below. Still, allow us for now, to shed some light on the concept behind askmarkets.

Our inspiration for askmarkets stems from several years of research on the topic of prediction markets, enabling us - hopefully - to separate the wheat from the chaff. The concept, at its very core, goes like this. Markets are institutions that are everywhere. From the stock market in New York to the fish market in Tokyo to the flower market in Amsterdam, markets exist and are used for a variety of purposes. But, while stock markets, for example, seem to be so different than betting, or fish and flower markets, in their core all those institutions are essentially the same: Markets bring people together, they sum up their information and transmit it through prices.

Askmarkets intends to bring this functionality to the masses. We provide a tool with which anyone can create a virtual market or trade in markets created by others. In other words, we provide a tool which can aggregate the opinions and knowledge of the many and transform these into a meaningful result. Markets arise as the ideal tool to crowdsource cognitive tasks and arrive at consensus results which are typically proven to be more accurate or correct than the opinions of the few experts, as suggested by both theory, experiments and practice.

Practically speaking, the whole concept of markets as an information aggregation tool has a relatively short history of existence. Better captured by the term ‘prediction markets’, it is only during the last years that the rise of the Web has given an actual shape to the idea and its potential. Especially among companies, the promise of markets as a decision support tool has been put into practise and been used as an effective decision-making tool only during the second half of the running decade, while by now all of the following companies are reported to be using markets internally: Siemens, HP, Intel, General Electric, Google, Eli Lilly, Microsoft, Archelor, Nokia, Intercontinental, Best Buy, Electronic Arts, Swisscom, Cisco, Motorola, Qualcomm, InfoWorld, MGM, Chiron, Frito Lay, TNT, Yahoo, Corning, Masterfoods, Pfizer, Abbott, Chrysler, General Mills, O’Reilly and TNT (disclaimer: no, these are not our customers, yet). Typical uses of markets for enterprise decision support range - so far and we’re still counting - from planning and forecasting, project management and marketing to new product development and market research. As our product is suitable for and targets all these applications, we could say that we naturally serve these or any other enterprise needs for dynamic information aggregation. Stay tuned for specific details coming out real soon.

Another major field of use regards public applications of aggregating people’s choices, on elections, movies, products or anything else. Especially regarding election results, relevant markets have proven to be more reliable than polls and surveys, and - lately - are gaining extended coverage from media sources like Time, Wired, CNN, Economist, New York Times, Financial Times, Newsweek, Fortune, ABC, Science, Nature, New Scientist, Business Week, Wall Street Journal and others. Moreover, many of these outlets already use markets to tap into the collective intelligence of their audience, giving them the ability to directly shape public consensus to ultimately provide a better alternative than polls. We also address such emerging needs - including yours, too -, so bear with us to find out more or just contact us in case that you want to be among our enterprise product testers. And thank you for your fantastic feedback and support so far, you really stand as our most valuable of assets!

Comments

cross posted at ‘A Course by Blog’

This semester, as you may know, I was privileged enough to teach (or, to be academically correct, to provide teaching assistance, but I truly thank my professor for entrusting me and giving me full freedom to handle everything by myself) the course ‘Information Extraction Algorithms’ -pure Data Mining in practice- at the postgraduate program ‘Applied Mathematical Sciences‘ of theNational Technical University of Athens. And the whole course actually turned out to be a one-off experience, for both me and the 10 students who appeared to take it, by deciding to host the whole course process in a blog. 

‘A Course by Blog’, a wordpress blog created on-the-fly during the first introductory lecture, finally concluded with 142 posts from 11 authors, 182 comments and 8,965 pageviews in total till today. We put everything related to the course online and publicly available, from the course’s syllabus and (optional) short CVs of its participants, to the lecture notes and videos (also, a livestream of the lectures was provided), to the weekly assignments, the final exam and grades, while in parallel a lot of more general posts -still relevant to the core machine learning topic- appeared (well, these are all in greek, apologies).

So, you may now follow the course’s proceedings in detail, or, if you are really determined to, watch on demand all lectures and study the relevant material, actually get the whole course’s juice at your convenience. Well, while this may be a pretty useful corollary effect, the main reason that triggered the whole decision was different. The focus was on enabling and fostering the conversation among the students, actually attempting to switch from a closed and locked status to an open and participative, yet competitive one. So, I’d like to add some rough thoughts on the experiences I gained through the whole implementation.

I might now say that the experiment was after all successful, actually a rather successful one. All of the students proved to be good recipients of the -novel to them- blogging platform and I’m positive that their exposure to each others’ assignments and questions resulted in a significantly improved learning experience. Plus, the recorded live streams of the lectures (16h of broadcasting time, available here) attracted 116 viewers, of 14m viewing time on average, while also resulted in some invaluable last-day-before-the-exam responses like “Hi Prof, I’m watching again all of your lectures and need to ask you a couple more things on…”. 

At the end of the day, I need to confide that going open clearly demands your time, in a way secures the minimal commitment for both the teacher and the students, as nobody wants to spoil his public reputation. And I’m convinced that all of us spent a lot more time on the course, especially when you compare it to the typical offline version. But, most of this effort was made voluntarily and happily, following uncontrolled patters, while resulting in added value for all of the participants. Isn’t that what you ideally expect from an academic course?

Comments

Comments

I attended earlier today the 3rd Workshop on Prediction Markets, organized with the support of Yahoo Research at the Kellogg School of Management, Northwestern University in Chicago. The event managed to gather top-notch researchers and practitioners of the field from Japan till, well, Greece, and I really enjoyed the focus and depth of all presentations.

What’s more, I had the chance to met up in person and chat with, among others “PM colleagues”, the Chicago-based Inkling guys -we owe them a lot at askmarkets-, Emile Servan-Schreiber of Newsfutures -he proved to be voluble and absolutely helpful on sharing experiences from many years of providing markets services-, as well as his own majesty and my very own prediction markets guru David Pennock; yes, he is such a geek but modest.

The presentation I gave for our late research attempts with Panos Ipeirotis entitled as ‘Detecting Important Events Using Prediction Markets, Text Mining and Volatility Modeling’ is attached hereby, while, as a work in progress, your comments are more than welcome.

| View | Upload your own

Comments

It’s true. It’s big. It’s awesome. After 22 Open Coffee events and hundred of participants around Greece this season, we’re very pleased to announce that we’ll cohost Open Coffee Athens XIII with Techcrunch. When? Tuesday, July 1st.

Close your mouth :), we are serious. The blog that brought the term start-up to the mainstream and is read by millions of people worldwide is on european tour and Mike Butcher, editor of Techcrunch UK, will join us to meet-up with you all and bring attention to the best and brightest of our attempts.

The setting goes like this: Take our very own entrepreneurs’ community. Add 10 Greek web start-ups, getting a 5′ chance to show and tell their story to the world. And all these with tons of karma and passion, in an uber cool-location (to be announced soon).

If you were dreaming about your start-up mentioned at Techcrunch, wake up and send us an email at opencoffee(dot)gr(at)gmail(dot)com. Or, if you would like your name among the sponsors, again email us. And, if you’d like to feel the love and passion that we Greeks share about start-ups and the web at large, just join us, we bet you will.

So, stay tuned for updates, save the date, spread the word and be there (or be square)!

cross-posted by OpenCoffee.gr

Comments