Tom Slee (SAP); Algorithmic Accountability: The Big Problems; Their Blog; 2017-10.
tl;dr → You have problems, SAP has expertise in this practice area. Call now.
Yvonne Baur, Brenda Reid, Steve Hunt, Fawn Fitter (SAP); How AI Can End Bias; In Their Other Blog, entitled The D!gitalist; 2017-01-16.
Teaser: Harmful human bias—both intentional and unconscious—can be avoided with the help of artificial intelligence, but only if we teach it to play fair and constantly question the results.
New Theory Cracks Open the Black Box of Deep Learning; Natalie Wolchover; In Quanta Magazine, also syndicated out to copied onto Wired.com; 2017-10-09; pdf.
Teaser: A new idea called the “information bottleneck” is helping to explain the puzzling success of today’s artificial-intelligence algorithms — and might also explain how human brains learn.
tl;dr → the “information bottleneck,” an explainer; as the metaphor.
and → <quote><snip/> that a network rids noisy input data of extraneous details as if by squeezing the information through a bottleneck, retaining only the features most relevant to general concepts.</quote>
Phases of Deep Learning
“fitting” or “memorization”
Is shorter (than the longer phase).The network learns labels for training data.
“compression” or “forgetting”
Is longer (than the shorter phase).
The network observes new data, to generalize against it. The network
optimizes (“becomes good at”) generalization, as measured differential with the (new) test data.
Alex Alemi, Staff, Google.
…quoted for color, background & verisimilitude; a booster.
Kyle Cranmer, physics, New York University.
…quoted for color, background & verisimilitude; a skeptic.
…quoted for color, background & verisimilitude; is non-committal, “It’s extremely interesting.”
Faculty, University of Toronto
Brenden Lake, assistant professor, psychology & data science statistics, New York University.
In which a data scientist is a statistician who performs statistics on a Macintosh computer in San Francisco; and Prof. Lake’s employer is the university system of the State of New York.
Burke, Craxton, Kolstad, Onda; Some Paper (a “deep and subtle discussion of discount rates”); 2016.
<quote>First, we discuss the social cost of carbon (SSC) and how it could be improved, including the consideration of catastrophes, nonmarket damages, impacts in developing countries, growth versus level effects, adaptation, and the use of discount rates. We then turn our attention to the integrated assessment models (IAMs) used in the computation of the SCC, arguing that, in addition to the need for incorporating the latest scientific understanding, we need to examine leading models’ consideration of uncertainty, the aggregation of heterogeneous agents, and technology options. Finally, we look at ways to improve climate policy design, in particular through the use of ex post analyses, insights from behavioral economics, the consideration of technology policy, and considerations specific to the developing world. With significant time and resources, we believe that progress can be made and many of these gaps filled.</quote>
tl;dr → they are using robots so that reporters can focus on “high value work.” i.e. most of what you’re reading, right now, today, even at the big brightly-lit news shops, is farmed content auto-formfill-assembled drek.
The Associated Press
Jeremy Gilbert, director of strategic initiatives, The Washington Post.
Francesco Marconi, strategy manager, and AI co-lead (that’s an actual title?) Associated Press.
Seth Lewis, professor, of journalism, University of Oregon.