Victoria Chua

[Boze Herrington, Library Owl 😴🧙‍♀️ @SketchesbyBoze]

I’m not kidding: those who delegate all their writing, thinking and creative expression to machines are going to wake up one day and discover that they can no longer write or think. You need to make your own art. You need to keep your brain working. You need to stay human.

[The PhD Place @ThePhDPlace]

Do a PhD in a field you are passionate about and you’ll never work a day in your life, because there are no jobs.

[Benjamin (offline for a bit) @bschne]

ok this is the best one I've seen yet

[Victoria Chua @viccyhan]

One of the many interesting projects I got to work on in @BLIPLabNTU with @suzyjstyles - we investigated how babies & parents coordinated speech with each other. It's like a game of ping-pong (or tennis)! Check out our poster at #ICIS2024 today 🏓

[Dr Suzy J Styles @suzyjstyles]

@BLIPLabNTU is back at @Infantstudies for #ICIS2024 with data from our Talk Together Study: 146 parents narrated a storybook to their children. Here, every row is one parental narration (8m/o at bottom, 40m/o at top). Something changes with child age, but what is it?
(P1-E-94)

[gena ?! @kytalli]

beginning with “automate the boring stuff away w/ python” is like a good hook into learning how to code to make discrete tasks disappear, but then the real reward of code is beginning to think about solving problems thru systems that can vapourise a class of task altogether

[Eiko Fried @EikoFried]

Eating disorders are among the mental health problems with the highest mortality rate.
It is at best completely nonsensical to frame them as a “survival mechanism” given the data.

[Skyler Wang @skylrwang]

So proud to announce that No Language Left Behind (NLLB) has officially been published by @Nature—this was years of hard work from a really dedicated group of researchers. I'm also so honored to be a part of the editorial team for this one. LET'S GO TEAM! https://t.co/A9R8j2Gc2g

[Tanu Mitra @tanmit]

If you're a recently graduated or about-to-graduate PhD looking for a postdoc in the space of auditing AI systems, epistemic risks of generative AI, and thinking about human-centric ways of addressing these risks, please apply and reach out: https://t.co/Td034CK7Vl

[Ms. G 💖 @pragueyerrr]

@Prigoose @visakanv @AskYatharth the human condition is that we will continue to create languages that can only be heard and spoken between two people and we will continue to grieve the extinction of that language when the only 2 people who speak it don’t anymore. it is tragic and it is beautiful

[James Bedford @james_s_bedford]

https://t.co/hhmxGqrkOo

Here is the AI guidance slide I developed for UNSW. This will be rolled out with an ethical framework on the use of AI and shown in classrooms for students to understand responsible AI use. If you would like to use or adapt any info please acknowledge.

[ICIS @Infantstudies]

.@CecileIcecile @CogTalesTweet @CristiaAlex's (2023) meta-analysis: infants prefer speech over other vocal, natural, or familiar sounds from birth, not only from 3mos. Caveat: pub bias & low representation of certain stimuli types and ages #infancypapers https://t.co/nrdpFLfrCR

[Xuechunzi Bai @baixuechunzi]

🚨 New Preprint 🚨
OpenAI says that the new GPT-4o model is carefully designed and evaluated to be safe and unbiased. But implicit biases often run deep and are harder to detect and remove than surface-level biases. In our preprint, we ask, are GPT-4o and other LLMs really unbiased?

Details in the🧵 below!

[Miranda ✨️ @MirandaevaM]

The most annoying part of academic writing is knowing the thing, but looking for some other paper to say it for you. You can't just say bread is soft, you have to say it according to Baker (2001) that bread is soft.

At the same time, you can't quote Baker 2001 because it's outdated. Although bread did not stop being soft, you have to find another Baker who said it in 2023.

Weeeeeeeeeeee aiy kuningi! Ngeke singahlanyi makunjena

[Jaclyn A. Siegel, PhD @jacasiegel]

I believe it should be a federal crime to publish a scale validation paper without including the items, instructions, response options, and scoring within the manuscript. All reviewers and editors who let scale validation papers be published without these are complicit.

[Moshe Poliak @MoshePoliak]

Once, we ran a study on Prolific and a participant wrote on Reddit that the study “Felt like I was losing the will to live.” I went on the Prolific Subreddit (24k members!) and asked what matters. Here is what they told me. A thread on happier participants and better studies 1/9

[Stuart Ritchie @StuartJRitchie]

-Sample size: 49
-p-values: 0.045, 0.052, 0.029, 0.032 (i.e. prima facie p-hacking)
-Scientific value: very little
-Number of views on Twitter regardless: 3,000,000

Doesn’t this Huberman guy have any standards for the stuff he promotes to his huge audience?

(via @alexabelson9)

[Emily @the_aiju]

you may think this is silly but humans get stuck all the time as the dog in more abstract versions of this exact situation and a lot of ‘healing’ moves are just the higher-dimensional equivalent of opening the invisible door

[Marvin Lavechin @LavechinMarvin]

Are you using models to study human speech processing? 🗣️🤖🧠👶

Consider submitting your work to our special session "Computational models of human language acquisition, perception, and production" at @ISCAInterspeech 2024

organized by @ojrasanen,@thomashueber, and myself!