Bianca Dumitrascu

@bimidu.bsky.social

computational biologist and statistician interested in how genes encode shape, form, and dynamics during development and regeneration; assistant professor @Columbia University https://www.morpho-lab.com/

Science grows through conversation and by widening who gets to shape it. Join us in Sydney to explore the opportunities and roadblocks for AI in Bio, and how we build foundations for knowledge that is transparent, diverse, and robust. Benchmarks, failure modes, and surprisingly tough tasks welcome!

ICBINB-BIO@icbinbbio.bsky.social · 2w ago

🧬 We’re excited to announce the NeurIPS 2026 workshop, I Can’t Believe It’s Not Better: Failure Modes of AI in Biology (ICBINB-BIO), to be held in Sydney, Australia, on Dec. 11 or 12, 2026! 📣 We welcome full and tiny papers sharing challenges in developing AI models for biological tasks. (1/n)

How are two embryos alike? As we collect spatio-temporal microscopy data, we want to quantify variability in the timing of key developmental events. Alignment of multiple recordings is a core engineering challenge here and we suggest a solution; read about it: www.biorxiv.org/content/10.6...

A quantitative coordinate system for developmental dynamics

Quantitative comparison of morphogenesis across individuals remains a fundamental challenge, as developing embryos vary in shape, orientation and developmental tempo. Moreover, real-time three-dimensi...

biorxiv.org

Excited to present our framework for interpretable disentanglement in multi condition data at ICML' 26 arxiv.org/abs/2506.17182 . While useful for different types of data, we think it's particularly useful if you have omics collected across distinct covariates like age or perturbations! Try it out!

Variational Learning of Disentangled Representations

Disentangled representations enable models to separate factors of variation that are shared across experimental conditions from those that are condition-specific. This separation is essential in domai...

arxiv.org

We're looking for two postdocs to join us! Topic 1: interpretable diffusion models for multi-scale morphology + genomics data! Topic 2: generation and control of biomaterials! If interested, send me CV + 3 recommendation letters over email!

Very excited about this first adventure in looking at regeneration! Thank you Maria and @smeetonlab.bsky.social for a fantastic first collaborative project :)

Joanna Smeeton@smeetonlab.bsky.social · 2y ago

📢🧪For your holiday travel reading, check out our new pre-print! Using a new joint resection injury model, we describe an innate regeneration program in adult #zebrafish that reestablishes all major synovial joint cell types, including lubricated articular cartilage. www.biorxiv.org/content/10.1...