Tag: google

Prediction Markets

Internet gambling is good for consumers. Too bad America wants to ban it

The spread of the internet has made the online gambler king. The emergence of large online gambling companies has slashed gaming operators’ margins and driven up payout ratios for gamblers. And the punters have embraced it in their millions, especially in America, where illegal gambling has long flourished. Last year 12m Americans placed about $6b in online bets, 50% the world’s total. You might have hoped politicians would greet such a demonstration of popularity with moderation—welcoming online gambling’s benefits and curbing its inevitable excesses. Instead they have put all their chips on red. Last weekend Congress passed a bill that will stop banks making payments to online gambling sites, adding to an already formidable legislative arsenal that outlaws most online gaming.

congress just killed a golden goose.
2006-12-03: uh crap i would have LOVED to attend this one. prediction markets is one of my permathreads.
2007-01-30: performance matters for betting markets too. to get even faster you’d have to factor out the humans clicking the buttons, but i guess that would require betting to grow up from being entertainment.
2007-04-24:

The key drivers in both filtering and consensus making in such systems is often positive feedback. For Digg, good stories get “dug” which makes them more likely to be seen and dug further. Poor stories do not get dug, rank lower, are less likely to be dug and slowly disappear. In ant trails, pheromone deposition stimulates further deposition if the food source really is good. Perhaps software developers should be given a cache of tokens. They can give the manager of a project 2 tokens each day if they think the project worth continuing. The manager can hire a programmer for the day [for 1 token] if he has enough tokens. Thus, crappy projects soon get filtered out by consensus. [The 2 vs 1 token prevents too much fluidity in the project and damps some random fluctuations].

2007-12-06:

In a new book, “Imperfect Knowledge Economics”, Mr Frydman sets out an alternative approach to prediction, in which the forecaster recognizes that his model will inevitably be less than perfect. Their work has received glowing praise from Nobel-prize-winning economists such as Kenneth Arrow and Edmund Phelps, who wrote the introduction to the book—though it is unlikely to have gone down so well with Robert Lucas, who won the Nobel for his work on rational expectations.

2008-01-07:

Despite the markets’ strong forecasting abilities, there is a slight optimistic bias driven mainly by new employees. On average, outcomes that were good for Google were overpriced by 20%. This bias was strongest on days after appreciations in Google stock and, ironically, for outcomes under our own control! We also find biases against extreme outcomes and short selling. Given a range of 5 outcomes, the middle ones were typically overpriced and unprofitable by comparison with the outliers.

hmm, i must have slept through that one.
2021-04-21:

are prediction markets doomed to repeat errors as grave as giving Trump a 15% chance of overturning the election in early December, and a 12% chance of overturning it even after the Supreme Court including 3 judges whom he appointed telling him to screw off? My answer is, surprisingly, an emphatic yes, and I see a few reasons for optimism.

2022-02-09:

In 2010, Philip Tetlock (one of the signatories on the pro-prediction market letter) did some pretty basic forecasting work, not even prediction market level, and proved that he could significantly outperform top analysts at the CIA with access to classified information. The government refused to hire him or use any of his methods, and continued shutting down new prediction markets as they arose. Polymarket is probably the biggest prediction market currently available. US law considers unlicensed prediction markets to be somewhere between illegal gambling and illegal futures trading, ie definitely illegal. The US is becoming the North Korea of forecasting. Every other civilized country allows prediction markets. In a perfect world, they could ignore our constant own goals and move on without us. But because America has a disproportionate share of money, users, coders, and entrepreneurs, a US-less prediction market ecosystem won’t be living up to its potential. That means decreased ability to gather and process information and worse decision-making worldwide.

Pinging Service API


The Google Blog Search Pinging Service API allows users who frequently update their blog to programattically inform Google Blog Search about changes to their blogs. Blogging provider admins can also use this API to notify Google of changes to blogs on their platform(s). To set up automated pinging of Google Blog Search, create either an XML-RPC Client or a REST Client which sends requests as noted below. It doesn’t matter which method you choose for notification; both are handled in the same way.

this would be useful for other web content too (news items, say). unfortunately this will be filled with splogs

Translation

hmm

GT now gets 55% accuracy on English to Arabic. Human agreement on human translations is 60%. After this point they have no standard by which to measure their progress

2016-09-27: Getting amazingly close to human level performance. it’s interesting that for all languages, the gap between human and perfect translation is much much larger than between human and machine.

Neural Machine Translation: Much better translation quality
Full technical report (23 exciting pages of bedtime reading)

Research blog post

I’m very excited to announce that our new neural machine translation system closes the quality gap between the existing Google Translate production system and human quality translations by 58% to 87% for a variety of different language pairs (see table below, from the technical report we published today). This work has been a close collaboration between the Google Brain team and the Google Translate team.

Thanks to lots of hard engineering work and the computational efficiency of our Tensor Processing Units (see report), we are also rolling these benefits out to users of Google Translate, starting today with Mandarin to English as the first language pair live in production that uses this new system. We’ll be rolling out many more language pairs over the coming weeks.

This highlights the success of neural models at more accurately capturing the complexities of real human language, and is a powerful demonstration of the research our group has been doing on language understanding.

2016-11-15: Nice behind the scenes article on the recent translation breakthrough.

With this update, Google Translate is improving more in a single leap than we’ve seen in the last 10 years combined.

3 overlapping stories converge in Google Translate’s successful metamorphosis to A.I. — a technical story, an institutional story and a story about the evolution of ideas. The technical story is about 1 team on 1 product at 1 company, and the process by which they refined, tested and introduced a brand-new version of an old product in only about a quarter of the time anyone, themselves included, might reasonably have expected. The institutional story is about the employees of a small but influential artificial-intelligence group within that company, and the process by which their intuitive faith in some old, unproven and broadly unpalatable notions about computing upended every other company within a large radius. The story of ideas is about the cognitive scientists, psychologists and wayward engineers who long toiled in obscurity, and the process by which their ostensibly irrational convictions ultimately inspired a paradigm shift in our understanding not only of technology but also, in theory, of consciousness itself.

2023-07-08: Akkadian translation, with modest BLEU scores.

In its transliteration to English test, the AI model scored 37.47. In its cuneiform to English test, it scored 36.52. Both scores were above their target baseline and in the range of a high-quality translation. The model was able to reproduce the nuances of each test sentence’s genre. The AI model works best when it is translating short- to medium-length sentences. It also does better with more formulaic genres, like royal decrees and administrative records, than literary genres such as myths, hymns, and prophecies. With more training on a larger dataset, they aim to improve its accuracy. “100s of 100s of clay tablets inscribed in the cuneiform script document the political, social, economic, and scientific history of ancient Mesopotamia. Yet, most of these documents remain untranslated and inaccessible due to their sheer number and limited quantity of experts able to read them”