Crowd Wisdom Adjusted to Measure Information Popularity

June 2, 2017

The article on ScienceDaily titled In Crowd Wisdom, the ‘Surprisingly Popular’ Answer Can Trump Ignorance of the Masses conveys the latest twist on crowd wisdom, or efforts to answer questions by asking many people rather than specialists. Unsurprisingly, crowd wisdom often is not very wise at all, but rather favors the most popular information. The article uses the example of asking various populations whether Philadelphia is the capital of Pennsylvania. Those who answered yes also believed that others would agree, making it a popular answer. The article goes on to explain,

Meanwhile, a certain number of respondents knew that the correct answer is “no.” But these people also anticipated that many other people would incorrectly think the capital is Philadelphia, so they also expected a very high percentage of “yes” answers. Thus, almost everyone expected other people to answer “yes,” but the actual percentage of people who did was significantly lower. “No” was the surprisingly popular answer because it exceeded expectations of what the answer would be.

By measuring the perceived popularity of a given answer, researchers saw errors reduced by over 20% compared to straightforward majority votes, and by almost 25% compared to confidence-weighted votes. As in the case of the Philadelphia question above, those who predicted that they were in the minority deserve the most attention because they had enough information to expect that many people would incorrectly vote yes. If you take away nothing else from this, let it be that Harrisburg, not Philly, is the capital of Pennsylvania.

Chelsea Kerwin, June 2, 2017

Booz Boo Boo: Blue Chip? Maybe Not

June 1, 2017

I read “Booz Allen, NGA Probe Intel Leak.” Let’s assume that the information in the write up is “sort of” accurate. I suggest this because the article invokes the name of “Edward Snowden” and the name of “Hal Martin.” Both of these individuals allegedly behaved with a bit of professional “looseness.”

But the write up does more than remind me that the once highly regarded blue chip management consulting firm has become an example of how not to manage its own employees and contractors.

Too bad. I worked at Booz, Allen & Hamilton when the firm’s reputation was reasonably well regarded. Today I am not so sure I would place the Booz Allen outfit identified in the FCW article in my “I want to work their” Top 10.

The main point of the write up seems to me to be:

Edward Snowden, Hal Martin and now another Booz Allen Hamilton employee could be involved in the leak of sensitive intelligence data — though in the latest case, it appears it could be accidental.

The information, according the FCW, was sensitive. The error was a result of a misconfiguration error.

Nevertheless, a company charged with working within the constraints set forth by the client should have management procedures in place to prevent alleged security issues.

Booz, Allen & Hamilton once kept a low profile. Now the firm finds itself making headlines.

FCW is not the grocery store tabloid-type of “real news” outfit, of course. However, I ask myself, “Management or mismanagement?”

And from an outfit which once provided management consulting services to the world’s leading organizations.

Interesting.

Stephen E Arnold, June 1, 2017

Elastic Stack Offers Machine Learning Functionality on the Side

June 1, 2017

Elastic, the company behind the Elasticsearch stack, has announced the release of a commercial add-on available via X-Pack. The product will detect unusual changes or anomalies in Elasticsearch’s real-time data results.

Elastic is not overpromising the features of the add-on – a fact that is praised by Infoworld in Elasticsearch Stack Wises Up with Machine Learning:

One possible issue is that non-open-source machine learning applications can look more impressive than they actually are. Elastic is avoiding that (for now) by confining the promise of the new features to specific, well-defined goals. It’s also likely to be even more powerful when a full non-beta version is available at the scale provided by cloud partners like Google.

Elastic, based on Lucene, has emerged as the go-to choice for enterprise search. A free and open source version of the software is available at https://www.elastic.co/downloads/elasticsearch. By keeping its goals realistic, is Elastic poised to not only be in the race for the long haul, but win the search gold medal?

Mary Pattengill, June 1, 2017

Break into Netflixs Stockroom with This Chrome Extension

June 1, 2017

The article titled Search Hidden Netflix Categories and Save Your Favorites With This Extension on LifeHacker calls attention to Netflix’s treasure trove of hidden category codes. Using Netflix often feels like a very limited exercise, especially if you don’t use the DVD service. But part of that is because Netflix is only showing you titles based on what it thinks you will like. The algorithm has its perks, but it can also become a spiral of narrowing cultural interests. The article illumines,

Netflix has a ton of hidden categories codes you can use to find movies and shows you’re into. The aptly-named Chrome extension Netflix Categories helps you find and save the ones you like. The extension adds a button to your Chrome menu bar. Click it and you’ll see a drop down list of categories that you might not find on the Netflix site proper. You can search the categories by name to find something more specific.

What sort of categories are available? Everything under the sun, from “Movies for ages 0 to 2” to “Film Noir” to “Military Documentaries” to “Belgian Movies” to “Korean TV Shows.” These categories offer a great way to branch out and be exposed to content that might unlock new interests. Or they can help to pinpoint an area of interest and see everything that Netflix has to offer on the subject. At any rate, it is a helpful tool to navigate Netflix’s full inventory.

Chelsea Kerwin, June 1, 2017

« Previous Page

  • Archives

  • Recent Posts

  • Meta