we had studies unequivocally (empirically & conceptually) demonstrating the idea of “debiasing” training data/algorithms is akin to putting lipstick on a pig as far back as 2019. yet, the amount of papers that publish claiming to come up with “debiasing” techniques is astounding. our field is broken
May 22, 2026 at 4:08 PM
we had studies unequivocally (empirically & conceptually) demonstrating the idea of “debiasing” training data/algorithms is akin to putting lipstick on a pig as far back as 2019. yet, the amount of papers that publish claiming to come up with “debiasing” techniques is astounding. our field is broken
Out of all the mountains of work on "debiasing" word2vec, BERT, and LLMs, what anecdotes or evidence do we have that debiasing techniques have *actually been used in practice*, in industry or research?
(not referring to "cultural alignment" techniques)
(bonus if used and *to good effect*)
(not referring to "cultural alignment" techniques)
(bonus if used and *to good effect*)
May 5, 2025 at 9:17 PM
Out of all the mountains of work on "debiasing" word2vec, BERT, and LLMs, what anecdotes or evidence do we have that debiasing techniques have *actually been used in practice*, in industry or research?
(not referring to "cultural alignment" techniques)
(bonus if used and *to good effect*)
(not referring to "cultural alignment" techniques)
(bonus if used and *to good effect*)
fucked up my gender-debiasing regex and accidentally created the ottowoman empire which carried out the arwomenian genocide
May 4, 2026 at 4:01 AM
fucked up my gender-debiasing regex and accidentally created the ottowoman empire which carried out the arwomenian genocide
trying to trace the source of bias is a fool’s errand. training data alone doesn’t account for model bias. the mantra “bias in, bias out” is misleading at best and dangerous at worst
10/
10/
March 16, 2026 at 8:42 PM
trying to trace the source of bias is a fool’s errand. training data alone doesn’t account for model bias. the mantra “bias in, bias out” is misleading at best and dangerous at worst
10/
10/
New article about debiasing the #BoardGameGeek ranking: blog.recommend.games/posts/debias... 🤓
Debiasing the BoardGameGeek ranking
Bias is a bit of an ugly word, isn’t it? It certainly has become one of those battle phrases in the culture war, where both sides of the argument accuse the other of forcing their biases onto …
blog.recommend.games
September 20, 2024 at 8:43 PM
New article about debiasing the #BoardGameGeek ranking: blog.recommend.games/posts/debias... 🤓
Thrilled to share our new paper! We introduce a generalized autoDML framework for smooth functionals in general M-estimation problems, significantly broadening the scope of problems where automatic debiasing can be applied!
Lars van der Laan, Aurelien Bibaut, Nathan Kallus, Alex Luedtke
Automatic Debiased Machine Learning for Smooth Functionals of Nonparametric M-Estimands
https://arxiv.org/abs/2501.11868
Automatic Debiased Machine Learning for Smooth Functionals of Nonparametric M-Estimands
https://arxiv.org/abs/2501.11868
January 22, 2025 at 1:54 PM
Thrilled to share our new paper! We introduce a generalized autoDML framework for smooth functionals in general M-estimation problems, significantly broadening the scope of problems where automatic debiasing can be applied!
On this International Day for the Elimination of Racial Discrimination, we plug our “Debiasing Committee Composition and Deliberative Processes”, a 1-pager w/strategies for diversifying perspectives & reducing biases for hiring, promotion, tenure, and funding decisions.
sfdora.org/resource/ret...
sfdora.org/resource/ret...
Rethinking Research Assessment: Debiasing Committee Composition and Deliberative Processes | DORA
This is part of DORA’s toolkit of resources to support academic institutions that are improving their policies and practices. Find the other resources in the toolkit here. Debiasing Committee Composit...
sfdora.org
March 21, 2025 at 2:31 PM
On this International Day for the Elimination of Racial Discrimination, we plug our “Debiasing Committee Composition and Deliberative Processes”, a 1-pager w/strategies for diversifying perspectives & reducing biases for hiring, promotion, tenure, and funding decisions.
sfdora.org/resource/ret...
sfdora.org/resource/ret...
Barker
Debiasing
V limited new pressing of Barker's 2018 EP, mastered by Pole and tipped FFO Mark Fell, Lorenzo Senni, Second Woman, Rian Treanor
now shipping
boomkat.com/products/deb...
Debiasing
V limited new pressing of Barker's 2018 EP, mastered by Pole and tipped FFO Mark Fell, Lorenzo Senni, Second Woman, Rian Treanor
now shipping
boomkat.com/products/deb...
September 1, 2026 at 12:32 PM
Barker
Debiasing
V limited new pressing of Barker's 2018 EP, mastered by Pole and tipped FFO Mark Fell, Lorenzo Senni, Second Woman, Rian Treanor
now shipping
boomkat.com/products/deb...
Debiasing
V limited new pressing of Barker's 2018 EP, mastered by Pole and tipped FFO Mark Fell, Lorenzo Senni, Second Woman, Rian Treanor
now shipping
boomkat.com/products/deb...
STORE // Barker's Debiasing EP returns for 2026. Four tracks of morphing synth stabs and pared-down metronomic rhythms in varying time signatures that function, against all logic, as driving, hypnotic and unlikely dancefloor tools. inverted-audio.store/products/bar...
September 21, 2026 at 7:53 AM
STORE // Barker's Debiasing EP returns for 2026. Four tracks of morphing synth stabs and pared-down metronomic rhythms in varying time signatures that function, against all logic, as driving, hypnotic and unlikely dancefloor tools. inverted-audio.store/products/bar...
Also, the message of the article is not that there are no facts or that there's no point in arguing about facts across ideological lines! The opposite. It's a call for dialogue and debiasing.
January 11, 2026 at 10:45 PM
Also, the message of the article is not that there are no facts or that there's no point in arguing about facts across ideological lines! The opposite. It's a call for dialogue and debiasing.
August 2, 2025 at 6:43 AM
It feels like most of this debiasing has been pushed to the alignment / safety training so I’m not sure there is any standard on debiasing techniques nowadays
May 6, 2025 at 9:02 PM
It feels like most of this debiasing has been pushed to the alignment / safety training so I’m not sure there is any standard on debiasing techniques nowadays
8️⃣ So will democracy survive the internet?
Maybe — but only with action: debiasing algorithms, transparency under the DSA, & tools to help people recalibrate their perception of what views are actually prevalent. The technology to fix this exists. The political will is lacking 🧵/end
10/10
Maybe — but only with action: debiasing algorithms, transparency under the DSA, & tools to help people recalibrate their perception of what views are actually prevalent. The technology to fix this exists. The political will is lacking 🧵/end
10/10
June 4, 2026 at 6:53 PM
8️⃣ So will democracy survive the internet?
Maybe — but only with action: debiasing algorithms, transparency under the DSA, & tools to help people recalibrate their perception of what views are actually prevalent. The technology to fix this exists. The political will is lacking 🧵/end
10/10
Maybe — but only with action: debiasing algorithms, transparency under the DSA, & tools to help people recalibrate their perception of what views are actually prevalent. The technology to fix this exists. The political will is lacking 🧵/end
10/10
Who benefits (or doesn't) from debiasing training? We find that intuitive-analytic thinking dispositions, and not cognitive ability, predict training effectiveness.
New paper from: boissinesther.bsky.social & gordonpennycook.bsky.social
www.sciencedirect.com/science/arti...
New paper from: boissinesther.bsky.social & gordonpennycook.bsky.social
www.sciencedirect.com/science/arti...
May 8, 2025 at 12:25 PM
Who benefits (or doesn't) from debiasing training? We find that intuitive-analytic thinking dispositions, and not cognitive ability, predict training effectiveness.
New paper from: boissinesther.bsky.social & gordonpennycook.bsky.social
www.sciencedirect.com/science/arti...
New paper from: boissinesther.bsky.social & gordonpennycook.bsky.social
www.sciencedirect.com/science/arti...
Officially announcing the FIRST-EVER MCSR CENSUS!
If you have ever watched a Minecraft speedrun, please retweet this - we want maximum exposure for debiasing!
shoutout to the wonderful orgs nessampah, sivianox and emma tortona, as well as the 34 testers and 25 translators that made this happen!
If you have ever watched a Minecraft speedrun, please retweet this - we want maximum exposure for debiasing!
shoutout to the wonderful orgs nessampah, sivianox and emma tortona, as well as the 34 testers and 25 translators that made this happen!
July 22, 2025 at 3:20 PM
Officially announcing the FIRST-EVER MCSR CENSUS!
If you have ever watched a Minecraft speedrun, please retweet this - we want maximum exposure for debiasing!
shoutout to the wonderful orgs nessampah, sivianox and emma tortona, as well as the 34 testers and 25 translators that made this happen!
If you have ever watched a Minecraft speedrun, please retweet this - we want maximum exposure for debiasing!
shoutout to the wonderful orgs nessampah, sivianox and emma tortona, as well as the 34 testers and 25 translators that made this happen!
Predicted structures from the AFDB are poor substitutes for experimental structures when training inverse folding NNs. Here, authors show that this is because they are "too perfect", and passing them through a custom-trained AFDB-to-PDB NN helps fix this
threadreaderapp.com/thread/19384...
threadreaderapp.com/thread/19384...
June 27, 2025 at 11:33 AM
Predicted structures from the AFDB are poor substitutes for experimental structures when training inverse folding NNs. Here, authors show that this is because they are "too perfect", and passing them through a custom-trained AFDB-to-PDB NN helps fix this
threadreaderapp.com/thread/19384...
threadreaderapp.com/thread/19384...
❗New open access paper out with @boissinesther.bsky.social, Alexandra Delmas & @wimdeneys.bsky.social in Psychologica Belgica!
🔗 psychologicabelgica.com/articles/10....
⬇️ In this paper, we explored whether we could replicate previous debiasing findings among native French speakers. (1/4)
🔗 psychologicabelgica.com/articles/10....
⬇️ In this paper, we explored whether we could replicate previous debiasing findings among native French speakers. (1/4)
Adieu Bias: Debiasing Intuitions Among French Speakers
Recent debiasing studies have shown that a short, plain-English explanation of the correct solution strategy can improve reasoning performance. However, these studies have predominantly focused on Eng...
psychologicabelgica.com
April 22, 2024 at 7:16 AM
❗New open access paper out with @boissinesther.bsky.social, Alexandra Delmas & @wimdeneys.bsky.social in Psychologica Belgica!
🔗 psychologicabelgica.com/articles/10....
⬇️ In this paper, we explored whether we could replicate previous debiasing findings among native French speakers. (1/4)
🔗 psychologicabelgica.com/articles/10....
⬇️ In this paper, we explored whether we could replicate previous debiasing findings among native French speakers. (1/4)
📚 Biases and debiasing in policy decision-making doi.org/10.1111/eci....
Biases and debiasing in policy decision‐making
Policymakers often succumb to fallacies, leading to suboptimal decision-making and maladaptive practices. We map the key biases involved at three different, but communicating, domains: the scientific....
doi.org
May 21, 2025 at 9:14 AM
📚 Biases and debiasing in policy decision-making doi.org/10.1111/eci....
1/9
We are excited to share our new working paper:
arxiv.org/abs/2502.12323
If you use ML predictions (like remote-sensed data) as outcomes, the resulting regression coefficients can be biased by measurement error. With @megan-ayers.bsky.social @mdgordo.bsky.social @eliana-stone.bsky.social
We are excited to share our new working paper:
arxiv.org/abs/2502.12323
If you use ML predictions (like remote-sensed data) as outcomes, the resulting regression coefficients can be biased by measurement error. With @megan-ayers.bsky.social @mdgordo.bsky.social @eliana-stone.bsky.social
Adversarial Debiasing for Unbiased Parameter Recovery
Advances in machine learning and the increasing availability of high-dimensional data have led to the proliferation of social science research that uses the predictions of machine learning models as p...
arxiv.org
March 17, 2025 at 7:30 PM
1/9
We are excited to share our new working paper:
arxiv.org/abs/2502.12323
If you use ML predictions (like remote-sensed data) as outcomes, the resulting regression coefficients can be biased by measurement error. With @megan-ayers.bsky.social @mdgordo.bsky.social @eliana-stone.bsky.social
We are excited to share our new working paper:
arxiv.org/abs/2502.12323
If you use ML predictions (like remote-sensed data) as outcomes, the resulting regression coefficients can be biased by measurement error. With @megan-ayers.bsky.social @mdgordo.bsky.social @eliana-stone.bsky.social
Barker - Debiasing
September 16, 2025 at 5:10 PM
Barker - Debiasing
#eeac Day 4: @gubri.bsky.social JMP "Targeting Behavioral Change Interventions: An Experiment on Debiasing Savings for College" demonstrates the importance of theory-driven identification of behavioral barriers and subsequent personalized interventions targeting specific barriers. #econsky #econjmp
December 5, 2024 at 3:02 AM
#eeac Day 4: @gubri.bsky.social JMP "Targeting Behavioral Change Interventions: An Experiment on Debiasing Savings for College" demonstrates the importance of theory-driven identification of behavioral barriers and subsequent personalized interventions targeting specific barriers. #econsky #econjmp
The next ViCE seminar is this Thursday, April 9, at 1pm ET.
Sofia Amaral will present: "Can Debiasing Law Enforcement Officers Improve Performance? Evidence from an Expressive Arts Intervention in India"
Sofia Amaral will present: "Can Debiasing Law Enforcement Officers Improve Performance? Evidence from an Expressive Arts Intervention in India"
April 6, 2026 at 8:30 PM
The next ViCE seminar is this Thursday, April 9, at 1pm ET.
Sofia Amaral will present: "Can Debiasing Law Enforcement Officers Improve Performance? Evidence from an Expressive Arts Intervention in India"
Sofia Amaral will present: "Can Debiasing Law Enforcement Officers Improve Performance? Evidence from an Expressive Arts Intervention in India"
racism vs "innovation" apparently a trade-offs that one has to choose from
September 7, 2026 at 3:19 PM
racism vs "innovation" apparently a trade-offs that one has to choose from
"So will democracy survive the internet? Perhaps..."
June 4, 2026 at 11:06 PM
"So will democracy survive the internet? Perhaps..."