You are currently browsing the tag archive for the ‘testing’ tag.
COVID-19: AI to the rescue?
By Shlomo Maital
Today’s daily Haaretz * carries a brief report of how three brilliant Israeli scientists have tackled a pressing problem – the need to know where the COVID-19 hotspots are, in order to focus spatial separation without shutting down the economy of the entire country.
The three are Prof. Eran Segal, an expert in computational and systems biology, Weizmann Institute, Rehovot; Prof. Benny Geiger, also from Weizmann; and Prof. Yuval Dor, Hebrew University.
Segal notes that experience from studying previous epidemics, as well as knowledge about how COVID-19 spreads, show that the virus spreads through clusters of infection and that early identification of such clusters can help stop the virus from spreading, ot at least slow it considerably.
We have seen such clusters, or hot spots, in New Rochelle, NY, in Washington State (Seattle), and initially, in Wuhan, China.
Segal notes that one possibility is to use massive testing, as they did in South Korea. More than 10,000 persons are tested daily there for COVID-19.
Israel can’t do such extensive testing, at this stage, he notes. Hence, the solution the team found was to ask members of the public to fill out online daily questionnaires, which take less than two minutes to complete, that include details about various symptoms and place of residence, including street and zip code.
This information will be analyze, Segal notes, using machine learning algoithms that give researchers and the Health Ministry a variety of information. If enough data are collected, the tool will help give up-to-date assessment of the spread of the illness.
This ‘early warning’ system can help spot these clusters, long before other methods do. The AI algorithms could also determine the effectiveness of public health measures, such as self quarantine, to limit COVID-19’s spread.
The information, noted Segal, is collected using Google DOCS. No privacy is violated.
Segal says we need as many people as possible to fill out the questionnaire, in the initial pilot stage.
I wonder whether Israel can offer this approach to the US, where testing remains quite limited.
* Haaretz. “Israeli Researchers Hope AI Can Tame COVID-19, and They Want Our Help.” Asaf Ronel. March 17 2020.
How Everyone Can Be Better Than Average:
Why “No Child Left Behind” Leaves Kids Behind
By Shlomo Maital
In Garrison Keillor’s wonderful radio program Prairie Home Companion, that aired live from 1974 to 2016 – an incredible 42 years! — Keillor did regular segments on “Lake Wobegone” where “all the children are above average”.
He always ended the segment with these words: “Well, that’s the news from Lake Wobegon, where all the women are strong, all the men are good-looking, and all the children are above average.”
Now, all the children cannot be above average, if you understand what an average is.
But in fact – it turns out, in one sense, they CAN!! Let me explain.
In his excellent New York Times Op-Ed (Tuesday June 18, international edition), Alfie Kohn asks, Why Can’t Everyone Get A’s? he makes the distressing point that America’s educational system has for two decades been built on the wrong belief that “excellence is a zero-sum game”.
Why?
When George W. Bush was elected President in 2000 (actually, he lost, but Florida’s Republican Supreme Court screwed Democrat candidate Al Gore), the first thing he did was initiate No Child Left Behind legislation. That law mandated widespread standardized testing in US schools. The idea, based on free-market economics, was – you promote excellence only by measuring it.
But – how do you measure it?
My wife Sharone, an experienced school psychologist, explained the two alternate ways of assessment: a) norm-reference tests, and b) criterion-reference tests. Please take a moment to understand the difference:
Norm-referenced tests report whether test takers performed better or worse than a hypothetical average student, which is determined by comparing scores against the performance results of a statistically selected group of test takers, typically of the same age or grade level, who have already taken the exam.
A criterion-referenced test is a style of test which uses test scores to generate a statement about the behavior that can be expected of a person with that score. Most tests and quizzes that are written by school teachers can be considered criterion-referenced tests.
Let’s simplify. Norm reference tests are tests ‘on a curve’. There are always those who excel, and always those who flunk. It’s the nature of a curve. Zero sum.
No Child Left Behind was based on norm reference tests. And as a result a great many kids were and are being left behind.
There is a better way. Define a criterion for excellence, or anything else you want to measure. For instance: Answering 80% or more math questions correctly.
Test kids. See how many meet the criterion. The goal: Let every kid be ‘above average’, like in Lake Wobegone, where ‘average’ means ‘meeting the criterion’.
With norm reference tests, 20% of kids, for instance, will get A’s. No matter how hard the rest study, or learn, only 20% can get an A. It’s zero sum.
With criterion reference tests, EVERYONE can potentially get an A.
When schools report a high number of A’s, experts say, “grade inflation”. Why? Isn’t the goal of education to be inclusive, to help EVERYONE get an A, to make sure that truly, no child is left behind?
But norm reference tests BY DEFINITION leave 80%, say, behind.
Everyone CAN get A’s. Everyone can be above ‘average’, as in Lake Wobegone. America has sold a dangerous, false educational ideology to the world, including my country Israel.
It’s time to rethink how we assess our kids.
TIMnovate 