←
Laure Ciernik32 posts
Laure Ciernik avatar

Laure Ciernik

@lciernik

PhD student at ML Group @TUBerlin @bifoldberlin @HFA_academy @ELLISforEurope | MSc thesis at Boeva Lab, @ETH_en

◷ Joined September 2023▧ 8 Photos and videos
Luca Eyring avatar
⟳ Laure Ciernik retweeted
Luca Eyring@LucaEyring·Jul 7

I‘m at ICML this week and we’re presenting Diamond Maps at Hall A #2713 today at 2pm @electronickale, @zeynepakata, @peholderrieth! Also, feel free to DM if you wanna chat about diffusion/flows/reward alignment/etc.

Accurate reward guidance in diffusion models has been a long-standing challenge. We propose Diamond Maps 💎, through stochastic flow maps we make reward guidance both accurate and scalable.

Laure Ciernik avatar
Laure Ciernik@lciernik·Jul 7

✨Tomorrow's the day! Excited to present this work with @LucaEyring at ICML in HALL A #2711 at 10:30 am✨ If you're interested in extracting knowledge from vision transformers for downstream tasks, come by! Looking forward to great discussions and new research ideas.

Why you should probe more than just the final layer of your Vision Transformer to maximize performance. 🧵👇

Laure Ciernik avatar
Laure Ciernik@lciernik·Jan 16

4/ Interpretation: Attention heatmaps reveal that specialized tasks (medical, satellite) rely heavily on intermediate layers, while natural images favor later layers.

Post attachment
Laure Ciernik avatar
Laure Ciernik@lciernik·Jan 16

3/ The impact: ✅ Consistent gains across 20 datasets. ✅ +5.54 pp avg. improvement over standard linear probes. ✅ Works across model scales (Small to Large) and training objectives (CLIP, DINOv2, Supervised)

Laure Ciernik avatar
Laure Ciernik@lciernik·Jan 16

2/ How it works: A cross-attention mechanism dynamically weights and fuses [CLS] and Average-Pooled [AP] tokens from ALL layers. It automatically identifies the most relevant abstraction levels for your task.

Post attachment
Laure Ciernik avatar
Laure Ciernik@lciernik·Jan 16

1/ Task-relevant info is distributed across the entire hierarchy, not just the final layer. We propose Attentive Multi-Layer Fusion to unlock this potential.

Post attachment
Laure Ciernik avatar
Laure Ciernik@lciernik·Jan 16

Why you should probe more than just the final layer of your Vision Transformer to maximize performance. 🧵👇

Laure Ciernik avatar
Laure Ciernik@lciernik·6 Jun 2025

The core message remains: Training objective drives the consistency of pairwise representational similarities, and similarity-performance correlations depend on the dataset structure. 🧵3/3

Laure Ciernik avatar
Laure Ciernik@lciernik·6 Jun 2025

During the rebuttal, we've added: improved local structure measurements, a CKA stability analysis, and additional dataset observations. 🧵2/3

Laure Ciernik avatar
Laure Ciernik@lciernik·6 Jun 2025

🎉 Update: This work got accepted to #icml2025!! Huge thanks to my amazing co-authors @LorenzLinhardt, Marco Morik, @jdppel, @skornblith, and @lukas_mut for their great work and to all collaborators! 🙏 📄 Paper: arxiv.org/abs/2411.05561 💻 Code: github.com/lciernik/similari… 🧵1/3

Post attachment

If two models are more similar to each other than a third on ImageNet, will this hold for medical/ satellite images? Our preprint analyzes how vision model similarities generalize across datasets, the factors that influence them, and their link to downstream task behavior. 🧵1/7

Fabian Theis avatar
⟳ Laure Ciernik retweeted
Fabian Theis@fabian_theis·12 Dec 2024

1/🚀 Excited to share RegVelo, our new computational model combining RNA velocity with gene regulatory network (GRN) dynamics to model cellular changes and predict in silico perturbations. Here's how it works and why it matters! 🧵👇
biorxiv.org/content/10.1101/…

Post attachment
Valentina Boeva avatar
⟳ Laure Ciernik retweeted
Valentina Boeva@val_boeva·26 Nov 2024

🚀 New preprint from our lab, @ekrym2 and @fabian_theis : UniversalEPI, an attention-based method to predict enhancer-promoter interactions from DNA sequence and ATAC-seq🌟 🔬 Key Highlights: - Predicts chromatin interactions across unseen cell types with no retraining. - Outperforms state-of-the-art models like C.Origami and ChromaFold with Spearman’s Rho > 0.92 in unseen cell types. - Tracks chromatin dynamics in processes like macrophage reprogramming and cancer transcriptional states. - Scalable for both bulk and single-cell chromatin accessibility data. 🧬UseUniversalEPI for in-silico 3D chromatin modeling in your favorite model! Examples of applications: studies on genetic diseases, cell differentiation, and the regulatory impacts of non-coding variants. 📄 Read the full preprint: doi.org/10.1101/2024.11.22.6… by @AayushGrover8 @ekrym2 @ilibarra @fabian_theis et al.

Post attachment