When to Align, When to Predict: A Phase Diagram for Multimodal Learning

• URL: http://arxiv.org/abs/2606.11190v2
• Date: Wed, 10 Jun 2026 19:46:16 GMT
• Title: When to Align, When to Predict: A Phase Diagram for Multimodal Learning
• Abstract summary: Cross-modal alignment and cross-modal prediction are the dominant paradigms for representation learning.<n>There is no systematic understanding of when each succeeds, when each fails, and when cross-modal training helps at all.<n>We develop a unified linear framework that addresses both questions.
• Score: 17.440086282242124
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Cross-modal alignment (CA) and cross-modal prediction (CP) are the dominant paradigms for multimodal representation learning, yet there is no systematic understanding of when each succeeds, when each fails, and when cross-modal training helps at all -- a gap that leaves practitioners, especially in scientific domains like biomedicine or astrophysics, with heterogeneous instruments and multiple levels of organization and measurement, unable to diagnose why standard methods underperform the best single modality. We develop a unified linear framework that addresses both questions. Under a spiked signal-plus-noise model with structured cross-modal nuisance correlation, we derive separation ratios for both objectives that expose complementary failure modes: alignment whitens each modality and fails when nuisance is strongly correlated across views; prediction encodes whatever is cross-predictable through a one-sided whitening, with recovery governed by source-modality quality. The resulting phase diagram partitions multimodal problems into four regimes: Both, CA only, CP only, and Neither. We present a data-driven procedure to locate real-world datasets in this diagram using a small labeled subsample, identifying the preferred objective and prediction direction before any cross-modal training. Experiments on synthetic data, stereo-vision benchmarks, image-caption pairs, and real astrophysical data validate the predictions in the nonlinear regime, including the Neither regime where cross-modal training is actively harmful. Our framework lets practitioners diagnose their multimodal problem and choose the right objective before committing to training. Code to reproduce the results is available at https://github.com/IlayMalinyak/mm_align_vs_pred.

Related papers

• Posterior Samplings are Missing Modalities Generators for Medical Image Translation [0.0]
  We propose a unified framework that formulates missing-modality generation as a linear inverse problem under a joint distribution.<n>By learning a joint prior over the complete modality set, our method can reconstruct arbitrary missing modalities at inference time.<n> Experiments on BraTS and IXI datasets show that our method achieves the best performance over baselines across most missing-modality scenarios.
  arXiv  Detail & Related papers  (2026-07-21T06:41:44Z)
• HiMatch-AD: DINOv3-driven Hierarchical Matching for Training-free Medical Anomaly Detection [6.789086317935452]
  HiMatch-AD is a DINOv3-driven hierarchical matching framework for training-free medical anomaly detection.<n>We introduce a unified uncertainty-based fusion mechanism that adaptively weights maps according to their reliability.<n>Experiments on the BMAD benchmark, including brain MRI, liver CT, and retinal OCT datasets, demonstrate that HiMatch-AD consistently outperforms both training-based and DINO-based state-of-the-art methods.
  arXiv  Detail & Related papers  (2026-06-21T15:26:48Z)
• Conditional Imputation for Within-Modality Missingness in Multi-Modal Federated Learning [8.35064965629059]
  Multimodal Federated Learning (MMFL) enables privacy-preserving collaborative training.<n>Real-world clinical applications often suffer from within-modality missingness caused by sensor intermittency or irregular sampling.<n>We propose CondI, a framework explicitly addressing this missingness using conditional diffusion models.
  arXiv  Detail & Related papers  (2026-04-25T02:35:08Z)
• The Mean is the Mirage: Entropy-Adaptive Model Merging under Heterogeneous Domain Shifts in Medical Imaging [3.597779662054083]
  Model merging under unseen test-time distribution shifts often renders naive strategies, such as mean averaging unreliable.<n>We introduce an entropy-adaptive, fully online model-merging method that yields a batch-specific merged model via only forward passes.<n>We extensively evaluate our method with state-of-the-art baselines using two backbones across nine medical and natural-domain generalization image classification datasets.
  arXiv  Detail & Related papers  (2026-02-24T21:06:19Z)
• Robust Multimodal Representation Learning in Healthcare [12.190907451083765]
  Real-world medical datasets commonly contain systematic biases from multiple sources.<n>We propose a Dual-Stream Feature Decorrelation Framework that identifies and handles the biases.<n>Our method employs a causal-biased decorrelation framework with dual-stream neural networks to disentangle causal features from spurious correlations.
  arXiv  Detail & Related papers  (2026-01-29T16:27:54Z)
• Dual-level Modality Debiasing Learning for Unsupervised Visible-Infrared Person Re-Identification [59.59359638389348]
  We propose a Dual-level Modality Debiasing Learning framework that implements debiasing at both the model and optimization levels.<n>Experiments on benchmark datasets demonstrate that DMDL could enable modality-invariant feature learning and a more generalized model.
  arXiv  Detail & Related papers  (2025-12-03T12:43:16Z)
• Calibrated Multimodal Representation Learning with Missing Modalities [100.55774771852468]
  Multimodal representation learning harmonizes distinct modalities by aligning them into a unified latent space.<n>Recent research generalizes traditional cross-modal alignment to produce enhanced multimodal synergy but requires all modalities to be present for a common instance.<n>We provide theoretical insights into this issue from an anchor shift perspective.<n>We propose CalMRL for multimodal representation learning to calibrate incomplete alignments caused by missing modalities.
  arXiv  Detail & Related papers  (2025-11-15T05:01:43Z)
• A Flow Model with Low-Rank Transformers for Incomplete Multimodal Survival Analysis [36.102030480314816]
  We propose a novel framework that combines a low-rank Transformer with a flow-based generative model for robust and flexible multimodal survival prediction.<n>Our method not only achieves state-of-the-art performance under complete modality settings, but also maintains robust and superior accuracy under the incomplete modalities scenario.
  arXiv  Detail & Related papers  (2025-10-22T02:27:05Z)
• Bridging Critical Gaps in Convergent Learning: How Representational Alignment Evolves Across Layers, Training, and Distribution Shifts [1.9458156037869137]
  convergent learning is the degree to which neural systems arrive at similar internal representations.<n>We present a large-scale audit of convergent learning spanning dozens of vision models and thousands of layer-pair comparisons.<n>Findings fill critical gaps in our understanding of representational convergence, with implications for neuroscience and AI.
  arXiv  Detail & Related papers  (2025-02-26T00:04:24Z)
• Scaling Multimodal Pre-Training via Cross-Modality Gradient Harmonization [68.49738668084693]
  Self-supervised pre-training recently demonstrates success on large-scale multimodal data. Cross-modality alignment (CMA) is only a weak and noisy supervision. CMA might cause conflicts and biases among modalities.
  arXiv  Detail & Related papers  (2022-11-03T18:12:32Z)
• Counterfactual Intervention Feature Transfer for Visible-Infrared Person Re-identification [69.45543438974963]
  We find graph-based methods in the visible-infrared person re-identification task (VI-ReID) suffer from bad generalization because of two issues. The well-trained input features weaken the learning of graph topology, making it not generalized enough during the inference process. We propose a Counterfactual Intervention Feature Transfer (CIFT) method to tackle these problems.
  arXiv  Detail & Related papers  (2022-08-01T16:15:31Z)
• On Modality Bias Recognition and Reduction [70.69194431713825]
  We study the modality bias problem in the context of multi-modal classification. We propose a plug-and-play loss function method, whereby the feature space for each label is adaptively learned. Our method yields remarkable performance improvements compared with the baselines.
  arXiv  Detail & Related papers  (2022-02-25T13:47:09Z)
• Diversity inducing Information Bottleneck in Model Ensembles [73.80615604822435]
  In this paper, we target the problem of generating effective ensembles of neural networks by encouraging diversity in prediction. We explicitly optimize a diversity inducing adversarial loss for learning latent variables and thereby obtain diversity in the output predictions necessary for modeling multi-modal data. Compared to the most competitive baselines, we show significant improvements in classification accuracy, under a shift in the data distribution.
  arXiv  Detail & Related papers  (2020-03-10T03:10:41Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.