Harmonized Tabular-Image Fusion via Gradient-Aligned Alternating Learning
Not on the map. Why?
Abstract (arXiv)
Multimodal tabular-image fusion is an emerging task that has received increasing attention in various domains. However, existing methods may be hindered by gradient conflicts between modalities, misleading the optimization of the unimodal learner. In this paper, we propose a novel Gradient-Aligned Alternating Learning (GAAL) paradigm to address this issue by aligning modality gradients. Specifically, GAAL adopts an alternating unimodal learning and shared classifier to decouple the multimodal gradient and facilitate interaction. Furthermore, we design uncertainty-based cross-modal gradient surgery to selectively align cross-modal gradients, thereby steering the shared parameters to benefit all modalities. As a result, GAAL can provide effective unimodal assistance and help boost the overall fusion performance. Empirical experiments on widely used datasets reveal the superiority of our method through comparison with various state-of-the-art (SoTA) tabular-image fusion baselines and test-time tabular missing baselines. The source code is available at https://github.com/njustkmg/ICME26-GAAL.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 23 of its 34 references that are in the map. First posted on arXiv 2 Apr 2026.
Builds on
23 of 34 references.
References · 23
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2015Distilling the Knowledge in a Neural Networkcited by 10,324
- 2014Deep Learning Face Attributes in the Wildcited by 3,693
- 2017Gradient Episodic Memory for Continual Learningcited by 1,898
- 2018Efficient Lifelong Learning with A-GEMcited by 1,058
- 2021Revisiting Deep Learning Models for Tabular Datacited by 740
- 2021Deep Neural Networks and Tabular Data: A Surveycited by 491
- 2019What Makes Training Multi-Modal Classification Networks Hard?cited by 344
- 2022Balanced Multimodal Learning via On-the-fly Gradient Modulationcited by 264
- 2023Multimodal Representation Learning by Alternating Unimodal Adaptationcited by 75
- 2024MMPareto: Boosting Multimodal Learning with Innocent Unimodal Assistancecited by 73
- 2022M3Care: Learning with Missing Modalities in Multimodal Healthcare Datacited by 72
11 more
- 2023Best of Both Worlds: Multimodal Contrastive Learning with Tabular and Imaging Datacited by 69
- 2024ReconBoost: Boosting Can Achieve Modality Reconcilementcited by 60
- 2022The Modality Focusing Hypothesis: Towards Understanding Crossmodal Knowledge Distillationcited by 44
- 2024TIP: Tabular-Image Pre-training for Multimodal Classification with Incomplete Datacited by 32
- 2022ItemSage: Learning Product Embeddings for Shopping Recommendations at Pinterestcited by 23
- 2021DVM-CAR: A large-scale automotive dataset for visual marketing research and applicationscited by 17
- 2024MUNBa: Machine Unlearning via Nash Bargainingcited by 17
- 2024Learning to Rebalance Multi-Modal Optimization by Adaptively Masking Subnetworkscited by 16
- 2024TabEBM: A Tabular Data Augmentation Method with Distinct Class-Specific Energy-Based Modelscited by 16
- 2025Rethinking Multimodal Learning from the Perspective of Mitigating Classification Ability Disproportioncited by 13
- 2024Detached and Interactive Multimodal Learningcited by 11
Built on by
Citing papers per year: 2026: 1.
1 paper.