31 Followers
6 Following
16 Posts
Researcher at https://five.ai
Previous Uni of Oxford TVG student with Prof. Phil Torr
Intern META
Cofounder https://girlswhoml.com
GirlsWhoMLhttps://girlswhoml.com/
Sometimes I feel like I'm unearthing a treasure trove of information from the past, then I just realise I'm reading a CVPR paper from 2008

How robust are unsupervised representation learning methods (e.g. SSL) to distirbution shift compared to supervised learning?

𝐒𝐡𝐨𝐫𝐭 𝐚𝐧𝐬𝐰𝐞𝐫: Quite!
𝐋𝐨𝐧𝐠 𝐚𝐧𝐬𝐰𝐞𝐫: Our #ICLR2023 paper http://arxiv.org/pdf/2206.08871.pdf

Joint work with Imant Daunhawer & Amartya Sanyal @amartya

@eric_brachmann very true, I guess there are some examples when it's appropriate, but I think these are edge cases
Just finalising reviews for CVPR, what are peoples thoughts about seeing MNIST in the main paper? Red flag imo
@jarartur that's more like it? What book?
@the_heruman Why do people set these deadlines? I have a AAAI video due on the 3rd Jan, I just don't get the motivation
It's almost that time of year when everyone starts mentioning how many papers they're going to read over Christmas, how many are you aiming for? I'm going to be reading a big fat 0

RT @[email protected]

Is having multiple modalities a blessing or a curse? What is a good representation? Let's find out together!
We are proud to announce the 1st hybrid workshop on Multimodal Representation Learning at ICLR2023 🚀

More info: https://mrl-workshop.github.io/iclr-2023/
Organisers: Miguel Vasco, Adrian Javaloy, Imant Daunhawer, Petra Poklukar, Isabel Valera, Danica Kragic, Yuge Shi

Home

First Workshop on Multimodal Representation Learning (ICLR 2023)

What's your favourite seed? 👀
@kristinmbranson sounds like you've converged, i would think unpredictability in how you code is undesirable!