Research Publications Newsletters Above the Fold Above the Fold - 2 June 2016

Jim Michalko retired from OCLC in February 2016. He'll be greatly missed, but Above the Fold will continue on. Next week we'll bring you Jim's concluding  thoughts. From there several of our colleagues will continue to select new articles offering views and insights that can help our work evolve and have impact.


How to Teach Yourself About Algorithms Future Tense blog • 9 February 2016

The author asks "Have you ever thrown around the word algorithm without knowing what it means?" While you constantly hear people complain or marvel at the [Facebook/Netflix/Google/etc.] algorithm, they usually don't understand what it is. Golbeck thinks that "as algorithms gain increasing importance in our lives, it's critical for everyone to have a very basic understanding of what they are." Her article tries to provide that basic understanding.

I admire this kind of article that takes what might be a daunting topic and explains it easily enough for the moderately-motivated reader to come away with a basic understanding. Check out her very cogent explanation of the binary search algorithm.

This Future Tense article is part of a series called Futurography which introduces readers to the technologies that will define tomorrow. Each month from January through June 2016, they choose a new technology and break it down. Futurography is the product of a joint effort among Arizona State University, New America, and Slate. You can guess at the ASU and Slate motivations. But I'd never heard of New America—part of their mission is to identify and nurture public intellectuals. Wow. (Michalko)



Algorithms and the (Incomplete) Stories They Tell Us

UX Magazine • 25 January 2016

You better explain yourself to the UPS driver. The author of this post contends that "the stories algorithms provide must be designed to create true value for people, or we miss the great potential of IoT and the data it provides." He explains three ways in which algorithms under-deliver on the value they promise—gaps in the data, over-optimization, and opaque intelligence.

The three examples Goodhand uses to illustrate these algorithm failures are compelling. I was particularly taken by the United Parcel Service routing algorithm example. Its intelligence was so opaque that drivers mistrusted the instructions and would override the directions. For more about the UPS Orion project read this: At UPS, the Algorithm Is the DriverWSJ. (Michalko)



Companies Are Reimagining Business Processes with Algorithms • 8 February 2016

What outcomes get improved by machine learning? In this post the authors describe the new re-engineering of business processes that is occurring because of the power of machine learning and algorithms that get better as they are used. The first round of process re-engineering occurred in the 90's as businesses brought to bear information technology. This round is different say the authors, "With machine-reengineering, process changes are constant and driven not just by history but also by the predictive capabilities of machine-learning algorithms." They try to determine if the type of machine learning employed contributes to particular types of business success such as customer satisfaction or revenue.

I agreed with the authors when they observed that the first round of process engineering was too aggressive and tried to change too many processes simultaneously. That created problems and rough patches before the benefits kicked in. The interactive graphs get across their observations nicely.(Michalko)



The Crusade Against Multiple Regression Analysis • 21 January 2016

Articles based on using this statistical method need warning labels. "A huge range of science projects are done with multiple regression analysis. The results are often somewhere between meaningless and quite damaging," according to Richard Nisbett. In this wide-ranging conversation he explains why the correlational (observational) evidence so often deviates from the experimental evidence.

Later in the interview he takes on the inability to replicate a lot of social psychology "cueing" experiments. That many are irreplicable doesn't bother him. He says "We may sometimes describe a particular experiment as an example of some point, and that particular experiment might not replicate, but the theory that the experiment exemplifies has been established in any number of different experimental contexts." The entire conversation is peppered with descriptions of intriguing exemplar experiments. My favorite—put dots on the coffee urn in a pattern that looks like human eyes and people will contribute more money to the coffee expense jar. Really? (Michalko)



Skin (artificial, in the game, left behind)

Ioannis Yannas looks back wistfully on the surprising discovery of artificial skin (VIDEO).
Slate • 20 January 2016

All The Money In The World, In A Single Chart
Co.Exist • 11 January 2016

Here's How You're Going To Die
Co.Exist • 8 January 2016

The first because it's a moving story told with great emotion. It is a breakthrough that was a direct result of a failed experiment. "Go wherever ignorance is maximum."

The second because it's a fun visualization and gives you context for markets and numbers which are likely to be unfamiliar. Derivatives‽

The third because it's a nice display that appeals to a universal (morbid) interest. (Michalko)



Above the Fold Quiz

According to an item in this week's News and Views section, how many distinct publications are in the RLUK collective collection?

Get the answer.