Dr Casey Becker

Postdoc @pittcogneuro, dept neurosurgery - intracranial EEG, naturalistic face perception, AI stuff. #Firstgen yr12 🦘🏳️‍🌈🍉

[Matt McDermott @mattmfm]

Must watch: NYT just released a damning forensic analysis of the ICE shooting in Minneapolis.

It flatly refutes Trump administration claims — confirming the motorist was driving away, not toward the officer, and the officer was not hit by the vehicle.

https://t.co/n8opkxbtmU

[David Milner @DaveMilbo]

Free Palestine, Melbourne, today

[Dr Casey Becker @Casey_Becker]

@Derek_Vis_Nerd Neat!

[Dr Casey Becker @Casey_Becker]

I’m scared abt the grandparents that’ll get scammed, but I’m not scared for the younger generation. They’ll adjust their expectations and favour real people / faces.

[Beck 🎈 @beckylitv]

It’s terrible but I still think it’s wild that this can all be done with AI. Imagine in 6 months…

[Nancy Kanwisher @[email protected] @Nancy_Kanwisher]

via @NYTimes https://t.co/bSvQj0zFU3

[Dr Casey Becker @Casey_Becker]

@nataliepeluso I feel this in my bones. Who knew research had the same kick as performing?!

[Dr Casey Becker @Casey_Becker]

Dynamic diffusion deepfakes are finally here.

Microsoft releases VASA-1, which generates a realistic video from audio in real time, using a single portrait photo. Considering it generates its own "biological" movement, I'd say it's pretty impressive. Uncanny, but impressive.

[Dr Casey Becker @Casey_Becker]

@GntknB I’m sorry for your loss 🤍

[Dr Casey Becker @Casey_Becker]

@Dr_N_Butcher Yep, that seems to be the issue. Morphs make all expression features move at the same time and rate (usually linear), but any disruption to natural movement looks off.

[Dr Casey Becker @Casey_Becker]

First experiment from my PhD!

Dynamic faces are great, but videos are a nightmare for experimental control, so many of us use dynamic morphs of photos.

Except we showed that if you can’t use videos with natural face motion, you might be better off using static faces 😬

https://t.co/FJFDTSCM1p

[Dr Casey Becker @Casey_Becker]

@ufangyang I think they skew results quite a bit tbh. The naturalistic motion they remove is an important feature of emotional expressions. There are some good video databases that are controlled enough for most research :) Unless you're doing certain EEG research, then deepfakes can help!

[Dr Casey Becker @Casey_Becker]

@mzloteanu Agreed, videos should be used a lot more. Some experiments don’t have room for a time dimension, and videos can be tricky for EEG, so statics can be useful.

[nihanalp @_AlViN_lab]

I am looking for a #postdoc, who has experience in conducting EEG exp and interested in how the brain compress information. #jobopportunity please contact me for further details. The position is at least for one year at @sabanciu with possibility of extension.

[Dr Casey Becker @Casey_Becker]

@TheRealJaffman Thank you for these kind words

[Dr Casey Becker @Casey_Becker]

My dad died shortly before I graduated, so turning up was hard. He'd tell anyone who'd listen about me, and read my papers even though I told him not to. I wish I’d known he needed to talk about the things on his mind, all the things that fill the journals that he left behind.

[Dr Casey Becker @Casey_Becker]

I had a great time being a part of the AI symposium at @ACNS_Official this year, thank you @TGrootswagers & Patrick Cooper for inviting me, and @astrid_zeman for joining me. I also did a lil poster, that was fun too.

[Dr Casey Becker @Casey_Becker]

@akbobak I too am chief word-cutter, commander of conciseness. I feel for you. It’s time to kill your darlings.

[David Milner @DaveMilbo]

Gutted. If white Australia felt uncomfortable about the Voice part of the Uluru Statement, just wait until we get to the Truth Telling bit

[Kiera 🦋 @KieraGorden]

I can’t believe anyone would say no to this.

#VoteYes #Referendum2023 #ConstitutionalRecognition #VoiceToParliament