Zohar Strinka

@ka_schu I once found Calvin having diligently picked out all of the marshmallows from our box of Lucky charms.

Kids were suspiciously quiet. I found them in the pantry taking shots of sprinkles straight from the sprinkle shaker

At TB39 "Using Optimization to Maximize Equity." A key takeaway is that for measures where bigger numbers are bad for equity (e.g., pollution, distance traveled to amenities), you can use the Kolm-Pollak EDE score. Which happens to have an exact linear formulation #INFORMS2022

On the topic of hepatitis C elimination and understanding that an expensive cure is much cheaper than perpetual treatment.

At the Sanjay and Panna Mehrotra research excellence award session. They presented another "what if" tool for helping decision makers use OR to support their understanding and choices #INFORMS2022

At the second set of Wagner talks MB51. Nikos Trichakis shows the tool they developed to help UNOS decision makers test different prioritization weights for who gets a donated organ. I love what if tools for making prescriptive analytics accessible #INFORMS2022

I asked if she has seen anything to help reduce the burden of returns, especially in the online apparel industry. She called for adding return fees as a simple solution that seems to work.

Final slide with ideas for opportunities to reduce the damage of the garment industry.

At the plenary from Maxine Bedat on the life and death of your jeans. Her suggested alternate title regarding the environmental impact of the conventional garment industry: "really great benefits, let's consider the other side." #INFORMS2022

At a panel discussion about the future of COIN-OR. I've used PuLP on several projects. What have you used this set of projects for? coin-or.org #INFORMS2022

Final comment: in an uncertain world, a 1% difference in model accuracy may or may not play out into the future. So whether or not your sub-optimal theoretical solution is sub-optimal in practice may or may not play out.

Tool from their team for playing with all the sparse trees that are almost optimal: poloclub.github.io/timbertrek/

At the Sunday plenary at #INFORMS2022. Cynthia Rudin suggests that in tabular datasets with a lot of noisy features (which mean prediction is hard), then there are a lot of good models (i.e., similar to the most accurate model possible), including some very simple ones.

My Indy-local family recommended checking out the Indy monthly magazine indianapolismonthly.com/best-restauran… to find good restaurants while we're in town #INFORMS2022

@ka_schu Calvin hasn't started that yet, but I tended to answer Riker and Saria with discussions of "The Manufacturer." We also liked to say "Employee" which is a little bit hilarious to hear a 4 year old try to say.

My 3yo is in a "where does that come from" phase so there's a lot of discussions about global supply chains in our household right now