Special thanks to the BlueSky DCVLR crew: @yuhuiz.bsky.social @thaottn.bsky.social @vishaalurao.bsky.social @saining.bsky.social @sarameghanbeery.bsky.social
18.06.2025 14:27 β π 2 π 1 π¬ 0 π 0@benjaminfeuer.bsky.social
PhD researcher at NYU, working on LLMs, VLMs, and tabular foundation models from a data-centric perspective. Father of two, NYC diehard.
Special thanks to the BlueSky DCVLR crew: @yuhuiz.bsky.social @thaottn.bsky.social @vishaalurao.bsky.social @saining.bsky.social @sarameghanbeery.bsky.social
18.06.2025 14:27 β π 2 π 1 π¬ 0 π 0Check out:
Our Website: dcvlr-neurips.github.io
Our Starter Kit (Curate, Train, Eval): github.com/oumi-ai/oumi...
π§΅ 6 / n
* A submission = a curated reasoning dataset on @huggingface with 1k or 10k samples and a scalable, reproducible curation strategy you document in a write-up
* You donβt need to train a model
* You can submit with nothing more than a free Colab or Kaggle account for basic testing
π§΅ 5 / n
πͺanyone can compete for free πͺ: Thanks to our sponsor @LambdaAPI we offer three free submissions for up to 500 teams. This is unprecedented in data-centric research, which tends to be very expensive because you have to train lots of models!
π§΅ 4 / n
π€ open-models π€: every model we present results for will have open weights, and one of those models will be Molmo-O from @allen_ai (a recent best paper honorable mention from @cvpr at #CVPR2025), trained on open data.
π§΅ 3 / n
DCVLR is data-centric: we train an ~7B VLM on your dataset. The best performer (on benchmarks like MathVista, VMCBench and LiveXiv) will be eligible to win $1500 and a talk at #NeurIPS2025!
We also have a few twists compared to prior data-centric competitions β
π§΅ 2 / n
So excited to announce the DCVLR (Data Curation for Vision-Language Reasoning) competition at #NeurIPS2025, led by @oumi-pbc.bsky.social and Lambda AI!
πopen-data π
π€ open-models π€
π» open-source π»
πͺanyone can compete for free πͺ
dcvlr-neurips.github.io
π§΅ 1 / n
Co-organizing with wonderful collaborators from MIT, NYU, Stanford and UW: @thaottn.bsky.social , @sewoong79.bsky.social , @sarameghanbeery.bsky.social , @yuhuiz.bsky.social !
01.05.2025 17:04 β π 2 π 1 π¬ 0 π 0We are excited to be sponsored by @datologyai.com
, who will be providing prizes for best paper awards π
πWe welcome any submission that discusses domain-specific data curation pipelines and/or generalizable curation principles, getting us closer to building data-centric methods that are robust, efficient, and adaptable across domains.
Refer to our website for the call for papers!
π’ Announcing our data-centric workshop at ICML 2025 on unifying data curation frameworks across domains!
π
Deadline: May 24, AoE
π Website: dataworldicml2025.github.io
We have an amazing lineup of speakers + panelists from various institutions and application areas!
That's not what they did, they used gpt-4o for program synthesis, it's fundamentally different than asking the LLM to provide the correct response in the prompt
22.12.2024 11:06 β π 0 π 0 π¬ 1 π 0Thanks for sharing! FWIW, I sensed mostly optimism and excitement at NeurIPS -- the people I spoke to were eager to talk about their research and learn about mine. Let's meet up in the new year and compare notes @kyunghyuncho.bsky.social
22.12.2024 11:02 β π 1 π 0 π¬ 0 π 0That does seem like a sound rule! Although, interestingly, they did not apply it to me. π
14.12.2024 15:40 β π 0 π 0 π¬ 0 π 0Hello Vancouver!
11.12.2024 16:19 β π 1 π 0 π¬ 0 π 0or AI for science!
baskargroup.github.io/BioTrove/
Or tabular deep learning ...
www.linkedin.com/posts/benjam...
Or LLMs ...
sites.google.com/berkeley.edu...
NeurIPS folks, excited to connect next week at the conference!
HMU to talk about VLMs ...
neurips.cc/virtual/2024...
Unpopular opinion: the #ICLR2025 reviews were better quality than in the last few years.
I think its mainly because they had people review fewer papers.
Opinions?
This book helped me learn how to understand ideological and inconsistent intellectual stances (a bit)
www.amazon.com/Righteous-Mi...
I feel like my Macbook Pro battery is starting to go; it used to last all day, now it's dead by the afternoon. The thing is only 2.5 years old. π€¨
27.11.2024 11:45 β π 0 π 0 π¬ 0 π 0Excited to be making my first post on BlueSky! Let's talk AI research.
@eugenevinitsky.bsky.social, can I get a who's who on here? :-)