Eric Green @ericpgreen

associate prof @DukeU @DukeGHI, cofounder and chief scientist @asknivi, digital health, clinical psych, global mental health, he/his ericpgreen.com Durham, NC Joined February 2009

Tweets

2K
Followers

980
Following

2K
Likes

7K

Eric Green @ericpgreen

2 years ago

@DukeATHLETICS @DukeBASE @jerseymikes Tucker Green 39

0 0 2 24 0

View Details

Eric Green @ericpgreen

2 years ago

This is cheaper than registering a boat name. So there. Done.

0 0 0 88 0

View Details

Eric Green @ericpgreen

2 years ago

are we understanding what i am going for here

Eric Green @ericpgreen

3 years ago

I need to name a small 2005 sailboat. Please help.

4 1 4 2K 0

1 0 5 406 0

View Details

@yudapearl @avicenna @SolomonKurz I'm wondering how one might put the two together to enable non-technical team members to more easily ask causal questions via LLMs. ChatGPT will soon be able to see images, so maybe it could see a DAG. Given a dataset, a DAG, and a question, could an LLM answer?

0 1 1 42 1

View Details

Eric Green @ericpgreen

3 years ago

I'm want to annotate datasets to facilitate a wider group of team members asking questions of curated datasets via LLMs. I'm not finding examples of annotating w/ causal models like DAGs. The data are dumb, right @yudapearl? How to annotate data w/ causal assumptions for LLMs?

1 3 3 1K 2

View Details

Eric Green @ericpgreen

3 years ago

@avicenna @SolomonKurz @yudapearl That’s a nice point. I think we could amend my ask to passing dag to LLM more generally (with the question or possibly as part of the data annotation for some uses)

0 0 0 48 0

View Details

Eric Green @ericpgreen

3 years ago

@VincentAB 👀

0 0 2 196 0

View Details

Durham Bulls @DurhamBulls

3 years ago

Who wants to join us in the PNC Triangle Club on Friday night? We've got a 4-pack of tickets to give away thanks to @PNCBank! Repost for your chance to win! (Winner will be chosen at noon tomorrow, August 29)

11 306 123 29K 0

View Details

Eric Green @ericpgreen

3 years ago

@LSenderowicz I get that @LSenderowicz. When you're able to step back from it, I think you'll conclude that you acted admirably and that this is normal. Nearly every analysis has mistakes. Some get discovered, few get corrected.

1 0 1 68 0

View Details

Durham Bulls @DurhamBulls

3 years ago

We've got another 4-pack of PNC Triangle Club tickets for tonight’s game to give away thanks to @PNCBank! Retweet for a chance to win! (Winner chosen at 2pm today)

8 226 70 22K 0

View Details

Eric Green @ericpgreen

3 years ago

Eve says office h’oars

0 0 1 169 0

View Details

Eric Green @ericpgreen

3 years ago

I need to name a small 2005 sailboat. Please help.

4 1 4 2K 0

View Details

Eric Green @ericpgreen

3 years ago

@FedExHelp It says I need to be verified to DM you

1 0 0 9 0

View Details

Eric Green @ericpgreen

3 years ago

@FedExHelp your phone service won’t accept my tracking number and I’m close to losing it. Please send help.

1 0 1 15 0

View Details

Eric Green @ericpgreen

3 years ago

@sd2nyc2nc 🥹

1 0 0 15 0

View Details

Eric Green @ericpgreen

3 years ago

Lots of helpful resources and ideas shared in response to this question. Is this revised figure helpful? Trying to separate reliability from agreement in teaching about reliability.

Eric Green @ericpgreen

3 years ago

Can anyone point to resources you've found helpful for evaluating test-retest reliability? I am preparing a teaching example, and most psych references to the topic teach using pearson r correlation. #psych #psychometrics

6 10 39 15K 30

1 0 3 394 2

View Details

Eric Green @ericpgreen

3 years ago

@DemidenkoMI @SolomonKurz Yes it struck me as counterintuitive when I was trying to explain it that r could equal 1 yet it’s possible no one has the same score at time 2 they had at time 1.