Q& Your with Cassie Kozyrkov, Files Scientist on Google

Q& Your with Cassie Kozyrkov, Files Scientist on Google

Cassie Kozyrkov, Records Scientist at Google, recently visited the main Metis Info Science Bootcamp to present into the class as part of our wedding speaker series.

Metis instructor together with Data Researchers at Datascope Analytics, Bo Peng, asked Cassie a few questions about her work and career in Google.

Bo: What their favorite element about becoming a data researcher at Yahoo or google?

Cassie: There is a selection of very interesting problems to work regarding, so you certainly not get bored! Archaeologist teams for Google talk to excellent things and it’s a lot of fun to be at the front part line of satisfying that attention. Google is as well the kind of ecosystem where you’ll expect high impact data work to be supplemented with some lively ones; for example , my co-worker and I have got held double-blind food trying sessions with a small exotic studies to determine the nearly all discerning palette!

Bo: In your chat, you speak about Bayesian vs Frequentist studies. Have you plucked a “side? ”

Cassie: A big part of this value to be a statistician is actually helping decision-makers fully understand the main insights in which data gives into their things. The decision maker’s philosophical stance will determine what s/he is actually comfortable finishing from records and it’s my responsibility in making this as fundamental as possible for him/her, which means that As i find ourselves with some Bayesian and some Frequentist projects. Continue reading “Q& Your with Cassie Kozyrkov, Files Scientist on Google”

DrivenData Contest: Building the perfect Naive Bees Classifier

DrivenData Contest: Building the perfect Naive Bees Classifier

This item was published and traditionally published by way of DrivenData. Many of us sponsored in addition to hosted the recent Novice Bees Classer contest, and the are the interesting results.

Wild bees are important pollinators and the pass on of colony collapse issue has simply made their goal more vital. Right now it does take a lot of time and effort for research workers to gather files on crazy bees. Using data put forward by citizen scientists, Bee Spotter is making this practice easier. Still they nevertheless require the fact that experts always check and indicate the bee in every image. After we challenged this community to construct an algorithm to pick out the genus of a bee based on the photograph, we were floored by the outcomes: the winners achieved a 0. 99 AUC (out of 1. 00) about the held outside data!

We involved with the top notch three finishers to learn of their backgrounds that you just they handled this problem. On true amenable data fashion, all three was standing on the shoulders of new york giants by benefiting the pre-trained GoogLeNet version, which has accomplished well in the main ImageNet rivalry, and adjusting it to the present task. Here’s a little bit around the winners and the unique treatments.

Meet the winners!

1st Place – E. A.

Name: Eben Olson and Abhishek Thakur

Family home base: Different Haven, CT and Munich, Germany

Eben’s History: I do the job of a research researchers at Yale University School of Medicine. Continue reading “DrivenData Contest: Building the perfect Naive Bees Classifier”