Seung-Goo Kim, Ph.D.
@seunggookim.bsky.social
Human neuroscientist | #musicscience #fmri #eeg #meg #ml | 🎶 in 🧠&🤖 | Postdoc @ae.mpg.de | Fellow @ jqya.de | Acad. Editor @plosone.org | Cons. Editor @ online.ucpress.edu/mp | Recomm. @ neuro.peercommunityin.org | 🏠 seunggookim.github.io | by dur 🇰🇷🇩🇪🇺🇸🇬🇧🇯🇵
Re: 1. Why don't YOU get out more?
Re: 2. Please read the doc. Thanks! 😊
Re: 2. Please read the doc. Thanks! 😊
April 6, 2025 at 4:16 PM
Re: 1. Why don't YOU get out more?
Re: 2. Please read the doc. Thanks! 😊
Re: 2. Please read the doc. Thanks! 😊
And what if I told you that the model in the figure was trained by pure noise?🤯😱 This scary artifact is due to a faulty cross-validation design, which I call "Reverse Double-Dipping" in my new preprint. Please read it to avoid RDD!👹 doi.org/10.1101/2025...
Reverse Double-Dipping: When Data Dips You, Twice—Stimulus-Driven Information Leakage in Naturalistic Neuroimaging
This article elucidates a methodological pitfall of cross-validation for evaluating predictive models applied to naturalistic neuroimaging data--namely, 'reverse double-dipping' (RDD). In a broader co...
doi.org
April 6, 2025 at 3:28 PM
And what if I told you that the model in the figure was trained by pure noise?🤯😱 This scary artifact is due to a faulty cross-validation design, which I call "Reverse Double-Dipping" in my new preprint. Please read it to avoid RDD!👹 doi.org/10.1101/2025...
Isn’t it a classic case of class confounding and the attribution bias? 🤣
February 4, 2025 at 10:42 AM
Isn’t it a classic case of class confounding and the attribution bias? 🤣
I will co-host a workshop for Open Science, talk about the neural encoding of musical emotions, and present a poster about LLM-based conference programming via iterative hierarchical clustering (i.e., "How the Bratwurst🌭 is made"). If you're around in the Frankfurt am Main area, please join!
January 9, 2025 at 9:14 AM
I will co-host a workshop for Open Science, talk about the neural encoding of musical emotions, and present a poster about LLM-based conference programming via iterative hierarchical clustering (i.e., "How the Bratwurst🌭 is made"). If you're around in the Frankfurt am Main area, please join!
Reposted by Seung-Goo Kim, Ph.D.
I hate being a stereotypical epidemiologist, but… They only asked if people commuted by bike at baseline and never again. They did not control for obesity, income, or other forms of exercise. The most we can conclude from this is that exercise is good, or that healthy people commute by bike.
December 18, 2024 at 12:42 PM
I hate being a stereotypical epidemiologist, but… They only asked if people commuted by bike at baseline and never again. They did not control for obesity, income, or other forms of exercise. The most we can conclude from this is that exercise is good, or that healthy people commute by bike.
Reposted by Seung-Goo Kim, Ph.D.
This is how you do it, America.
Celebrations in South Korea as the President is impeached for attempting a coup d'état. Democracy has been protected.
(📹/imyour_space)
(📹/imyour_space)
December 14, 2024 at 2:23 PM
This is how you do it, America.
It’s not whether you can believe or not an entire discipline. The motivation to cheat exists by the academic system. This events represent self-regulation of the discipline by peers, if any.
December 8, 2024 at 11:12 AM
It’s not whether you can believe or not an entire discipline. The motivation to cheat exists by the academic system. This events represent self-regulation of the discipline by peers, if any.