| On the Web | https://kiciman.org/ |
| Papers | https://scholar.google.com/citations?user=QZCU3NkAAAAJ&hl=en |
| Code | https://github.com/emrekiciman |
The Smithsonian has made over 4 MILLION images open source! You can create, copy, distribute, etc. This is amazing!
I missed posting this for #cockatoo #tuesday but maybe it's still ok to post on #wednesday.
Here is Lila dancing funky to the viral dance music from netflix's Wednesday. Just for fun.
Just a 10sec clip for now. Full video available on request.
An AI test from 1971 shows you need to adjust your #ChatGPT expectations.
In 1971, Terry Winograd devised the Winograd Schemas, simple sentence understanding tasks that require the use of knowledge and commonsense reasoning. Humans understand them easily. ChatGPT... not so much.
I missed posting this for #cockatoo #tuesday but maybe it's still ok to post on #wednesday.
Here is Lila dancing funky to the viral dance music from netflix's Wednesday. Just for fun.
Just a 10sec clip for now. Full video available on request.
I’m gonna do a manual quote here to say that this is exactly the type of stuff we predicted, and especially with automation bias with people overtrusting model outputs. Police didn’t even check the height of the person they claimed did this, and neither was the person in Atlanta.
https://dair-community.social/@poppast[email protected]/109634255362720798
A black man wrongfully jailed for a week in Louisiana after face recognition error, report says. Lawyer says police didn't check man's height, weight or the mole on his face. #blackmastodon #facialrecognition https://peopleofcolorintech.com/articles/facial-recognition-tech-used-to-jail-black-man-for-louisiana-theft-hes-never-been-to-louisiana/
No need to go into backstory, but over dinner tonight we(*) shared a lot of frog puns to mixed reactions. The jokes were a tad pole-arizing.
(*) mostly me.
Plus, many expensive refutations now run in parallel, making them much faster (Thanks Andreas Stöffelbauer and Amey Verhade!) and as of DoWhy 0.9.1, DoWhy has cleaned up dependencies and supports python 3.10
See more in our release notes: https://github.com/py-why/dowhy/releases
If you want to see your favorite algorithm or feature in the next version of DoWhy, we have a contributor's guide (thanks to Michael Marien!) https://github.com/py-why/dowhy/blob/main/docs/source/contributing/contributing-code.rst
Open for discussions at our Discord and in GitHub issues
Jeffrey Gleason adds e-value sensitivity analysis https://pywhy.org/dowhy/v0.9.1/example_notebooks/sensitivity_analysis_testing.html
Anusha0409 adds sensitivity analysis for non-parametric estimators: https://pywhy.org/dowhy/v0.9.1/example_notebooks/sensitivity_analysis_nonparametric_estimators.html
Patrick Bloebaum and Kailash Budhathoki improve the GCM module with a new API for unit change attribution, support for modeling FCMs with auto-gluon, and more!
Egor Kraev adds support for EconML's multi-treatment estimators; and Amit Sharma
adds support for estimating direct effects (previously had been total effect) @causal
Dr. Amy, Psy.D.
Mor Naaman
Timnit Gebru (she/her).
Cassandrich