mlsanityData to 14 Oct 2026

Flow to Learn: Flow Matching on Neural Network Parameters

arXiv 2503.19371 ↗First posted 25 Mar 2025cs.LG cs.AIcited by 3

On the map: Diffusion models → Flow matching, time series and science

Abstract (arXiv)

Foundational language models show a remarkable ability to learn new concepts during inference via context data. However, similar work for images lag behind. To address this challenge, we introduce FLoWN, a flow matching model that learns to generate neural network parameters for different tasks. Our approach models the flow on latent space, while conditioning the process on context data. Experiments verify that FLoWN attains various desiderata for a meta-learning model. In addition, it matches or exceeds baselines on in-distribution tasks, provides better initializations for classifier training, and is performant on out-of-distribution few-shot tasks while having a fine-tuning mechanism to improve performance.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 42 of its 52 references that are in the map. First posted on arXiv 25 Mar 2025.

Builds on

42 of 52 references.

Closest ancestors · 9

Same direction, newest first

  1. 2024Flow Matching with General Discrete Paths: A Kinetic-Optimal Perspective4 Dec
  2. 2024Discrete Flow Matching22 Jul
  3. 2024Metric Flow Matching for Smooth Interpolations on the Data Manifold23 May
  4. 2024Generative Flows on Discrete State-Spaces: Enabling Multimodal Flows with Applications to Protein Co-Design7 Feb
  5. 2023InstaFlow: One Step is Enough for High-Quality Diffusion-Based Text-to-Image Generation12 Sep
  6. 2023Multisample Flow Matching: Straightening Flows with Minibatch Couplings28 Apr
  7. 2023Improving and generalizing flow-based generative models with minibatch optimal transport1 Feb
  8. 2022Flow Matching for Generative Modeling6 Oct
  9. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow7 Sep

Other references · 33

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,897
  4. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,349
  5. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
  6. 2018Neural Ordinary Differential Equationscited by 3,779
  7. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,611
  8. 2018The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networkscited by 2,262
  9. 2016Matching Networks for One Shot Learningcited by 2,255
  10. 2016HyperNetworkscited by 1,027
  11. 2022Building Normalizing Flows with Stochastic Interpolantscited by 995
  12. 2019Meta-Learning with Implicit Gradientscited by 525
21 more
  1. 2020Principal Neighbourhood Aggregation for Graph Netscited by 478
  2. 2018How to train your MAMLcited by 309
  3. 2024Adjoint Matching: Fine-tuning Flow and Diffusion Generative Models with Memoryless Stochastic Optimal Controlcited by 233
  4. 2013Predicting Parameters in Deep Learningcited by 224
  5. 2023A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recommendationscited by 202
  6. 2020Meta-Baseline: Exploring Simple Meta-Learning for Few-Shot Learningcited by 179
  7. 2023Supervised Pretraining Can Learn In-Context Reinforcement Learningcited by 139
  8. 2022Pushing the Limits of Simple Pipelines for Few-Shot Learning: External Data and Fine-Tuning Make a Differencecited by 126
  9. 2022General-Purpose In-Context Learning by Meta-Learning Transformerscited by 105
  10. 2024Graph Neural Networks for Learning Equivariant Representations of Neural Networkscited by 59
  11. 2022Hypernetworks in Meta-Reinforcement Learningcited by 42
  12. 2022Relational Attention: Generalizing Transformers for Graph-Structured Taskscited by 41
  13. 2024Towards Scalable and Versatile Weight Space Learningcited by 41
  14. 2023MetaDiff: Meta-Learning with Conditional Diffusion for Few-Shot Learningcited by 40
  15. 2022Model Zoos: A Dataset of Diverse Populations of Neural Network Modelscited by 39
  16. 2024Diffusion-Based Neural Network Weights Generationcited by 39
  17. 2024A Survey of Lottery Ticket Hypothesiscited by 24
  18. 2024Learning on LoRAs: GL-Equivariant Processing of Low-Rank Weight Spaces for Large Finetuned Modelscited by 22
  19. 2023Context-Aware Meta-Learningcited by 21
  20. 2022HyperMAML: Few-Shot Adaptation of Deep Models with Hypernetworkscited by 17
  21. 2023ProtoDiff: Learning to Learn Prototypical Networks by Task-Guided Diffusioncited by 16

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026Position: Weight Space Should Be a First-Class Generative AI Modalitycited by 6
  2. 2026DeepWeightFlow: Re-Basined Flow Matching for Generating Neural Network Weightscited by 5
  3. 2026NNiT: Width-Agnostic Neural Network Generation with Structurally Aligned Weight Spacescited by 3

Similar, not cited

Close in wording, no citation either way

  1. 2025Flow Matching in the Low-Noise Regime: Pathologies and a Contrastive Remedy
  2. 2026Flow-OPD: On-Policy Distillation for Flow Matching Models
  3. 2026A Theory on Flow Matching with Neural Networks
  4. 2026Exploring and Exploiting Stability in Latent Flow Matching
  5. 2023Flow Matching in Latent Space
  6. 2025ContinualFlow: Learning and Unlearning with Neural Flow Matching
  7. 2025AlphaFlow: Understanding and Improving MeanFlow Models
  8. 2026SubFlow: Sub-mode Conditioned Flow Matching for Diverse One-Step Generation