🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

My Rumelhart cognitive science prize talk was all about children and grandchildren both biological and academic. So here I am with the medal and both varieties of offspring, extremely proud of all three!

New paper: KiVA: Kid-inspired Visual Analogies for Testing Large Multimodal Models We present KiVA, a simple benchmark for visual analogies that even 3-year-olds can pass, but LMMs cannot. Paper: arxiv.org/abs/2407.17773 Benchmark: github.com/ey242/KiVA 1/7

"Causal Relational Problem Solving in Toddlers" 📢 New paper from: Mariel Goddu (@marielgoddu.bsky.social) Eunice Yiu @alisongopnik sciencedirect.com/science/articl…

KiVA is visiting European Conference on Computer Vision #ECCV2026! Check out our poster tomorrow: Sep 30th, 15:15 | Amber 5 (EVAL-FOMO workshop) Eunice Yiu w/ Maan Qraitem, Charlie Wong, Anisa Majhi, Yutong Bai, Alison Gopnik, Kate Saenko, and yours truly. #ECCV2024 🇮🇹

📷 This was another great Simons event, I'm reporting brand new work with Eunice Yu, Shiry Ginosar, Mariel Goddu, et al. on simple analogies that kids can solve and Large Language and Vision Models can't. youtube.com/watch?v=NFQIQ6…

Children can solve simple visual puzzles that stump large multimodal models! Check out our benchmark "Kid-inspired Visual Analogies," which has recently been accepted to ICLR, here: openreview.net/forum?id=vNATZ…

Think LMMs can reason like a 3-year-old? Think again! Our Kid-Inspired Visual Analogies benchmark reveals where young children still win: ey242.github.io/kiva.github.io/ Catch our #ICLR2025 poster today to see where models still fall short! Thurs. April 24 3-5:30 pm Halls 3 + 2B #312

Curious where 3–5 year olds still beat big AI models at analogical reasoning? Come find me at #ICLR2025 — Hall 3, Poster #312 — for KiVA: Kid-Inspired Visual Analogies.

Children & adults don't just explore—they seek empowerment: controllable and variable interventions. This intrinsic drive supports causal learning & generalization. 📄 Forthcoming in Phil. Trans. A (2026) osf.io/ept4n/download Alison Gopnik Kelsey Allen Shiry Ginosar

I’m excited to share that KiVA (Kid-inspired Visual Analogies), our image‐understanding benchmark grounded in developmental psychology, is launching as an open challenge at the Third Perception Test Workshop at ICCV 2025. Compete now for a chance to win prizes!