The paper reexamines the belief that more complex multimodal neural architectures lead to better performance. By reimplementing 19 influential multimodal methods across nine datasets with up to 23 modalities, the authors find that increased architectural complexity often causes confusion rather than effective fusion, and such models rarely surpass unimodal baselines or a simple baseline (SimBaMM). They also highlight methodological flaws in top-tier multimodal studies and call for a shift toward standardized evaluation and methodological rigor.
arXiv:2505. 19614v2 Announce Type: replace Abstract: Multimodal learning has seen remarkable progress, particularly with large-scale pre-training across various modalities.
The paper identifies that in multimodal learning, optimization often produces asymmetric certainty gains, with the stronger modality becoming more confident than the weaker one, which leads to imbalanced contributions and suboptimal performance. The authors attribute this issue to unimodal characteristics and propose a Max Confidence Regularization (MaxCR) method that tracks each modality’s semantic confidence via a nonlinear sparsity measure and applies max suppression and excitation to balance confidence levels. Experiments on standard datasets demonstrate that MaxCR improves overall performance compared to state‑of‑the‑art multimodal baselines.
arXiv:2607. 05019v1 Announce Type: new Abstract: In multimodal classification, late-fusion approaches classify concatenated modality-specific features extracted by unimodal neural networks.
The paper introduces a generalized multimodal foundation model that can handle arbitrary combinations of modalities and prediction tasks. It trains on large-scale synthetic multimodal datasets with diverse causal structures to learn transferable multimodal correlations. Experiments on 18 real-world datasets across 12 modalities and 11 tasks show competitive performance compared to specialized models without task-specific adaptation.
The survey titled "Tackling Multimodal Learning Challenges with Mixture-of-Expert: A Survey" reviews how Mixture-of-Experts (MoE) frameworks address multimodal learning problems. It categorizes MoE’s contributions into three roles: an efficient multimodal engine that scales without excessive computation, a representation learner that fuses diverse expert knowledge, and an adapter that handles imperfect data such as missing modalities. The paper also highlights research gaps like interpretable routing, expert communication, and lifelong multimodal learning.
arXiv:2608. 05000v1 Announce Type: cross Abstract: Vision offers a critical axis for advancing foundation models, driving a shift towards natively unified multimodal pretraining.
arXiv:2606. 26348v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) can process diverse inputs, e.
arXiv:2606. 11614v1 Announce Type: cross Abstract: Multimodal learning hinges on capturing redundant, unique, and synergistic information across modalities, which collectively constitute multimodal interactions.
The paper discusses the rapid progress of machine unlearning for multimodal large language models (MLLMs) and the need for rigorous evaluation. It identifies three pitfalls in current benchmarks: lack of held‑out evaluation, reliance on synthetic fine‑tuned knowledge, and unfair comparisons across methods. To address these issues, the authors propose GEMMU, a benchmark using real‑world knowledge and a principled protocol, and show that re‑evaluating existing methods under GEMMU uncovers significant generalization gaps.
The paper investigates how to scale a Transformer-based vision‑language model trained from scratch on multimodal inputs under a fixed compute budget. It finds that minimal loss follows a predictable compute law while optimal model size and token count obey power‑law scaling, with language and multimodal objectives showing distinct allocation trends. The study also proposes a budget‑compensation rule that allows additional multimodal tokens to offset text‑efficiency penalties, and demonstrates that native multimodal pre‑training improves spatial reasoning and few‑shot learning.
arXiv:2606. 09853v1 Announce Type: new Abstract: A central objective in multimodal learning is to capture synergy: task-relevant information that arises only from the joint use of multiple modalities, and is not available from any single modality alone.
The paper introduces Fusion Anything Model (FAM), a foundation model designed for generalized multimodal data fusion that can handle arbitrary modality combinations and prediction tasks. FAM is trained on large-scale synthetic multimodal datasets generated via Structural Multimodal Causal Models (SMCMs), enabling it to encode transferable multimodal correlations. Experiments on 18 real-world datasets across 12 modalities and 11 tasks show that FAM performs competitively with specialized models without requiring task-specific adaptation.