Overview

Multimodal learning asks how information from different sources can be combined to improve tasks such as diagnosis, retrieval, forecasting, and robot perception. Common approaches include feature fusion and cross-attention, contrastive alignment, self-supervised pretraining, and mixture-of-experts models that balance shared information with modality-specific cues. Foundation models increasingly support understanding and generation across several modalities, while domain-specific systems use paired or partially paired data and may use extra inputs during training even when they are unavailable at deployment. A central tension is that modalities can complement one another but can also be missing, noisy, misaligned, or unevenly used, so research focuses on robust fusion and evaluations that test whether models truly rely on the relevant evidence.

Latest papers

358 papers21 papers in the last four weeks, down 53% on the four weeks before. 0.1% of all new papers.

Nov 12, 2026

Unbundling Multimodal Vision Encoding

How multimodal language models should encode visual input remains up for debate. Most models use a pretrained vision encoder, while some recent recipes train the encoder from scratch with the language model or remove it entirely. Vision encoders bundle together three design choices: a visual prior from upstream pretraining, dedicated vision capacity, and pre-fusion placement of that capacity (i.e., before the multimodal backbone). Because published models also differ in data, architecture, and compute, comparing them cannot isolate the effect of each choice. We disentangle these choices via a controlled framework where we pretrain all models from scratch on common data and compare them on visual question answering and language benchmarks at matched lifetime compute. We find that the visual prior is not necessary; an encoder trained jointly from scratch is on par with a pretrained one, without a separate training stage. What matters instead is dedicated vision capacity: an encoder-free model trails on visual understanding even when its backbone is enlarged by the encoder's parameter count. Yet that capacity need not precede multimodal fusion, as vision-only modules added in-fusion to early backbone layers perform comparably to a pre-fusion encoder and can lower time to first token. These results also hold across backbone sizes, when encoder pretraining is treated as free, and when starting from a pretrained language model. Overall, our findings favor dedicated vision capacity trained jointly with the backbone and placed early within it.
Nov 12, 2026

MusicRWKV: A Multimodal RWKV Framework for Music Representation Learning

Recent music foundation models have achieved remarkable progress in music representation learning, yet most existing approaches rely on Transformer architectures whose quadratic attention complexity introduces substantial computational and memory overhead when modeling long-duration music recordings. In this work, we propose MusicRWKV, a multimodal music representation learning framework built upon the RWKV architecture. MusicRWKV jointly models waveform, spectrogram, and metadata information within a unified self-supervised learning framework. To improve efficiency for long musical sequences, we introduce an invertible compression module that reduces sequence length while preserving reconstruction capability. A FiLM-based conditioning strategy is employed to facilitate modality-aware representation learning without introducing expensive cross-attention operations. We further investigate the integration of Soft ROSA as a differentiable retrieval mechanism for neural-symbolic interaction. Experimental evaluation across a diverse collection of Music Information Retrieval (MIR) benchmarks demonstrates that the proposed RWKV-based architecture can be effectively adapted to heterogeneous music understanding tasks. A controlled pretraining ablation shows that Soft ROSA improves validation performance, while layer-wise telemetry analysis reveals a progressive transition from distributed retrieval to selective retrieval in deeper network layers. These results suggest that RWKV provides a promising architectural alternative to Transformer-based models for multimodal music understanding.
Nov 12, 2026

Asymmetric Multimodal Learning with Decomposed JEPAs

Cross-modal alignment is the standard objective for building multimodal representations. We show that alignment as a sole objective is lossy wherever modalities are informationally asymmetric---where neither can express what the other carries. Visual question answering is an obvious case: a question is a selective relational operator probing a dense visual field whose fine-grained features no query recovers, so what alignment preserves is not what the task requires. We prove an Alignment Redundancy Bound: alignment-only Dirichlet energy minimization retains at most the redundant information atom \Red(Z;X1,X2)\Red(Z; X_1, X_2), annihilating private modality-specific structure \Unq(Z;Xm)\Unq(Z; X_m) and cross-modal synergy \Syn(Z;X1,X2)\Syn(Z; X_1, X_2). We introduce Decomposed JEPA (D-JEPA), which factorizes latents into an aligned shared core and private modality-specific fibers. The shared core retains most of the structure alignment targets; the private fibers recover what alignment discards, via contextual masked prediction and asymmetric cross-subspace independence. Under information asymmetry, alignment-only training undergoes catastrophic semantic collapse on compositional reasoning benchmarks, surviving only on binary queries. D-JEPA averts this collapse, substantially improving downstream accuracy over both alignment-only and alignment-with-reconstruction baselines, and restoring a positive multimodal synergy gap in place of the baseline's inverted gap. Information-theoretic audits confirm that D-JEPA recovers most of the mutual information available from a frozen backbone, and causal fiber-swap interventions demonstrate that the recovered information is causally localized in the private fibers.
Nov 10, 2026

Relaxing Cross-Modal Correspondence for Multimodal Remote Sensing Pretraining

Multimodal remote sensing pretraining commonly relies on exact instance-level correspondence across modalities, implicitly treating strictly paired observations as a prerequisite for learning transferable representations. In this work, we revisit this assumption by relaxing cross-modal correspondence from the instance level to the semantic level. We introduce Semantic Pairing (SemPair), a pairing-relaxed pretraining strategy that replaces exact instance pairs with class-consistent cross-instance pairs. To characterize the effect of correspondence relaxation, we formulate a latent-factor model that separates cross-modal dependence into shared semantic structure and instance-specific correlation, showing that SemPair preserves the former while marginalizing instance-specific cross-modal correlations in expectation. We further introduce Mixed Pairing (MixPair) as a controlled interpolation between semantic and exact pairing to study correspondence granularity along a continuous spectrum. Extensive experiments consistently show that SemPair achieves transfer performance comparable to exact pairing (ExaPair), while MixPair reveals a non-monotonic relationship between downstream transfer and the instance-pairing ratio. These results demonstrate that exact instance correspondence is not a prerequisite for effective multimodal remote sensing pretraining and establish cross-modal correspondence granularity as an important design dimension for transferable multimodal representation learning. More broadly, this finding suggests a practical route toward scaling multimodal pretraining with reduced dependence on costly instance-aligned data.
Nov 9, 2026

Structured Kernel Surrogate Models for Multimodal Attribution

We study the problem of cross-sample modality attribution, which involves measuring how one modality in one sample affects the learning outcome in other multimodal samples. This problem is crucial in a multimodal system because a data modality (such as a text description) may be redundant for one sample (a self-explanatory image), but the same data modality would otherwise be uniquely informative for another sample, or informative when combined with other modalities. Prior work focuses on influence estimation at the level of one sample, and it remains an open question how one should capture the interactions across both samples and modalities. To capture scenarios like the above, we propose structured kernel surrogate models, which are surrogate mappings from both samples and modalities to the prediction outcome when they are trained together. We design an algorithm, SKS, which applies kernel ridge regression with a structured, compositional kernel on the selection states of individual samples and modalities. Specifically, the kernel function multiplies kernel-based similarity scores between every pair of data modalities for one sample and a pair of two samples. Additionally, we design a gradient-based estimation algorithm to efficiently estimate the structured kernel surrogate model. We evaluate our approach on two multimodal benchmarks, spanning a range of data modalities including satellite images, road networks, text, audio, and tabular data. We find that SKS achieves 57.857.8% higher correlation with the true prediction outcomes than existing methods. In downstream data selection, SKS improves the average accuracy by 1.931.93% over baselines. Models with data selected by SKS have lower Hessian traces, indicating a flatter loss surface. Using SKS to select satellite images identifies road segments, where accident scenes commonly occur, as opposed to background images.
Nov 9, 2026

Operation-aware Multimodal Clinical Decision Support System for ICU Discharge Risk Assessment

Unplanned 30-day intensive care unit (ICU) readmission is a critical benchmark for evaluating healthcare quality. Most predictive models assume readmission risk depends solely on a patient's individual pathophysiological state, ignoring hospital operational context and resource capacity. In practice, discharge decisions are dynamic trade-offs between patient stability and bed turnover pressures. To address this gap, we develop an Operation-aware Multimodal Clinical Decision Support System (CDSS) for ICU discharge risk assessment. The system combines multi-agent collaborative extraction with self-evolving retrieval-augmented generation (RAG) to eliminate template noise from progress notes, while constructing departmental workload indicators alongside admission flow features, continuous vital signs, and sparse medical codes (ICD/DRG). Addressing severe class imbalance (about 10.5% readmissions) and alarm fatigue, we implement data normalization, class reweighting, and decision-threshold calibration. In 20 independent runs across 20,000 ICU stays from MIMIC-III, empirical findings demonstrate that: vital signs alone have limited predictive capacity under real-world class imbalance (PR-AUC of 0.1387 +/- 0.0086); directly concatenating raw text embeddings degrades decision tree performance (XGBoost and LightGBM PR-AUC drop to 0.3893 and 0.4030); whereas an end-to-end multimodal fine-tuning architecture (Transformer E2E FineTune) utilizing layer-wise parameter freezing and differential learning rates reverses text negative transfer, achieving the highest AUROC of 0.8288 +/- 0.0110, PR-AUC of 0.4373 +/- 0.0310, and F1-Score of 0.4757 +/- 0.0196. Operational indicators supply independent risk signals, laying an engineering foundation for high-utility clinical decision support tools.
Nov 9, 2026

Layer Roles Matter: Mitigating Modality Imbalance in Multimodal Emotion Reasoning via Adaptive Layer Profiling

Multimodal Emotion Reasoning (MER) requires perceiving and integrating emotion cues from multiple modalities. Recent advances increasingly leverage Multimodal Large Language Models (MLLMs) to enhance reasoning capacity. However, due to the inherent heterogeneity in conveying emotions across modalities, these models often suffer from modality imbalance and confront flawed emotion understanding. In this paper, we conduct an in-depth analysis on MER models and uncover two key empirical insights. (i) In MER, modality imbalance manifests two distinct failure modes: unimodal over-reliance and insufficient cross-modal utilization, which call for different intervention strategies. (ii) MLLM layers exhibit implicit functional hierarchies in multimodal emotion processing, playing separable roles in stabilizing unimodal representations and integrating cross-modal information. Motivated by these observations, we propose a Layer-Aware Modality Rebalancing (LAMR) framework by explicitly aligning intervention strategies with layer roles. Specifically, we first introduce two quantized metrics for adaptive layer role profiling. In each layer, Modality Entropy (ME) measures the dispersion of modality contributions, and Cross-modal Intensity (CMI) captures the strength of cross-modal interactions. Then, we respectively enhance modality equilibrium in stabilization layers and cross-modal propagation in integration layers. Extensive experiments on two comprehensive MER benchmarks demonstrate that LAMR effectively mitigates modality imbalance and improves emotion reasoning across diverse MLLM backbones, without additional training.
Nov 6, 2026

UniMore: Fine-Grained Supervision for Unified Understanding and Generation

Unified multimodal models combine image understanding and generation within one system, creating an opportunity to improve both capabilities through paired image--text supervision. We examine the outputs of a pretrained unified model in both directions and observe a loss of specificity: generated images can contain the requested objects with incorrect attributes, while captions identify a scene without expressing its distinguishing details. These failures call for supervision that specifies both the visual content to realize and the descriptive differences to distinguish. We propose UniMore, a post-training method that derives these complementary learning tasks from the fine-grained content of image--text pairs. The generation route predicts grounded crops from their corresponding phrases, matching the scope of the condition and visual target while using the native visual objective. The understanding route combines positive-caption fitting with ranking against captions that omit or alter details, conditioned on the full image. The two routes are optimized independently, and their learned updates are combined into a single unified model without changing its inference procedure. Experiments show gains across multiple understanding and generation benchmarks.
Nov 5, 2026

Multimodal Information Interactions Bottleneck for Multimodal Self-Supervised Learning

Multimodal self-supervised learning is challenged by heterogeneity, unequal informativeness, and complex information interactions across multimodal data sources such as images, text, audio, and tabular data. While contrastive methods effectively align shared semantics, they typically rely on strong multi-view redundancy assumptions, overlooking modality-specific and synergistic information, thus leading to suboptimal representations. We propose the Multimodal Information Interactions Bottleneck (MIIB), a principled framework grounded in Partial Information Decomposition that explicitly encodes redundant, unique, and synergistic information within multimodal interactions. MIIB employs a dual-path model architecture for modality-specific encoding and multimodal fusion, optimized using surrogate tractable entropy-based objectives that enforce minimal sufficient representations without requiring explicit mutual information estimation. By leveraging variational bounds and score-based entropy estimation, MIIB efficiently captures complex cross-modal interactions while suppressing superfluous information. Experiments across diverse benchmarks demonstrate consistent improvements over state-of-the-art multimodal self-supervised methods. Code is available at https://github.com/wanglingzaishi/MIIB.
Nov 5, 2026

InfMasking: A Unified Variational Framework for Synergistic Multimodal Learning

Multimodal models are expected to exploit synergistic information, that is, information available only when multiple modalities are observed together. In practice, however, a fusion module can fit the training objective by relying on a dominant modality or on information that modalities share, and the synergy between modalities is often left unused. We propose InfMasking, a unified multimodal learning framework based on Infinite Masking. Before fusion, InfMasking randomly masks the features of each modality; each sampled mask produces a different partial combination of cross-modal evidence and a fused representation of the corresponding modality interactions. All masked fused representations are required to remain consistent with a reference target: the unmasked fused view when training is unsupervised, and the class proxy of the task label when labels are available. Maintaining this consistency across diverse partial combinations discourages reliance on a dominant modality or on shared information alone, and drives the fusion module to learn cross-modal interaction patterns that remain meaningful under partial observations. Since the objective is an expectation over infinitely many masks, we model the masked fused representations as Gaussian and optimize closed-form variational bounds, whose population-moment gaps are controlled by the masking-induced covariance under this model. On the evaluated benchmarks, InfMasking improves predictive performance and yields higher mean PID estimates of task-label synergy in learned modality-specific representations, alongside higher InterSHAP interaction attribution in the trained predictors. Code is available at https://github.com/brightest66/InfMasking.
Nov 2, 2026

Multimodal Prompt Learning with Dual-Granularity Contrastive Alignment for Preoperative Prediction of Lung Adenocarcinoma Invasiveness Using 3D CT

Accurate preoperative prediction of lung adenocarcinoma invasiveness is important for surgical planning and individualized treatment. However, overlapping CT appearances make it challenging to distinguish invasive adenocarcinoma (IAD) from Pre-IAD lesions. We propose a multimodal prompt-learning framework with dual-granularity image-text alignment for 3D CT-based IAD classification. A 3D ResNet-CBAM encoder extracts global and spatial visual features. A frozen RadBERT backbone is adapted with separate learnable prompts for report-level and keyword-level text encoding. These prompts are initialized using semantic representations derived from real-world clinical reports. Global contrastive alignment connects whole-volume CT representations with reports generated from clinician-annotated radiological attributes to provide overall semantic supervision. A multiple instance learning-based local alignment module associates spatial CT features with vascular-focused radiological keywords, to capture local imaging characteristics potentially associated with tumor invasion. The framework jointly optimizes classification and alignment objectives, while inference requires only CT images. Experiments included 1,593 pathologically confirmed nodules. On the test set, the proposed method achieved an accuracy of 87.92%, F1-score of 87.78%, and AUC of 0.9499. Its AUC significantly exceeded that of the image-only Res-CBAM baseline with the same visual backbone (0.9297; DeLong test, p = 0.0078). Ablation analyses supported the complementary contributions of global and local alignment and the benefit of report-based prompt initialization. Qualitative visualizations showed activation within nodules and adjacent vascular structures. These findings suggest that dual-granularity radiological supervision can enhance CT-based invasiveness prediction while preserving a CT-only inference workflow.
Nov 2, 2026

Hybrid Federated Learning for Multimodal Fusion under Missing Modalities and Site Heterogeneity

Multimodal fusion is studied under hybrid federation, where data are partitioned across samples and feature blocks. The proposed two-dimensional federated smooth multiple-kernel large-margin learning method (2D-FedSMKL) integrates availability-corrected modality scores, group-kernel regularization, and smooth worst-site risk aggregation. For fixed maps and propensities, the objective is convex and smooth. Curvature bounds and exact synchronous block-gradient equivalence are established for binary and multiclass scores, giving a conditional accelerated rate. An oracle-propensity binary risk bound and a propensity-perturbation result are also derived. Simulations and two public activity-recognition benchmarks assess site heterogeneity and modality absence. Worst-site aggregation improves worst-site performance under heterogeneity, including reversed and size-balanced designs. On the benchmarks, validation selects the correction; the selected model is not significantly worse than the strongest reference at the worst site, and group-regularization benefits vary by dataset.
Nov 2, 2026

HistoTx: Early fusion of H&E images and spatial transcriptomics at varying spatial transcriptomics resolution with self-supervised learning

Spatial transcriptomics maps gene expression across tissue sections and, when integrated with whole slide images, jointly captures both morphological architecture and molecular activity to provide a rich, multimodal view of tissue biology. ML approaches combining spatial transcriptomics and WSIs typically learn a mapping from morphological features to molecular profiles via contrastive alignment or supervised fine-tuning of pathology foundation models. In both cases, transcriptomics serves as a training signal rather than a true input modality, leaving inference fundamentally image-driven. We propose an early-fusion vision transformer architecture in which transcript tokens are constructed at arbitrary spatial resolution, merged with image patch tokens, and jointly processed through a shared transformer stack for deep cross-modal interaction. Starting from a pretrained vision-only pathology foundation model, we train HistoTx by continuing self-supervised pretraining on paired image and spatial transcriptomics data and demonstrate that: 1) at inference with image only, performance remains on par with the baseline on most tasks while improving on tasks that inherently benefit from molecular knowledge, such as gene expression prediction; 2) when both modalities are available at inference, jointly providing image and transcript inputs outperforms either modality alone.
Oct 30, 2026

Do Multimodal Models Decide Consistently on Equivalent Optical Spectra?

A spectrum can be given to a model as a table in nanometers or as a plot in photon energy, and the decision should not depend on which one it sees. When the data needed for an answer are missing, the model should say so. We test representation invariance and abstention on measured optical spectra of silicon metasurfaces (MOCLIP) and gold nanoparticles, with labels computed from the data. Each spectrum is shown as four tables (nm, μμm, scientific notation, eV) and two plots (nm, eV), and partial spectra remove data inside or outside the queried range. We evaluate the decision models d3-nano and Jev 1.13, three GPT-6 Luna configurations and the Qwen3.5-2B base model. Changing only the representation moves accuracy by up to 41 percentage points within one model, mostly on photon-energy tables and plots. With the same evidence, asking in photon energy instead of wavelength helps some model-view pairs by 16-50 points and hurts others by up to 25. When the relevant extremum is removed, unadapted models rarely answer "undetermined". LoRA fine-tuning of d3-nano improves comparisons on the trained table formats, but much of its gain on unanswerable questions comes from a shortcut: it abstains whenever the input is truncated. Training with answerable truncations avoids the shortcut in all six seeds, with a 90-100 point gap between abstention after inside and outside cuts on held-out structures; with the same amount of extra original-task data, five of six seeds keep the shortcut. On gold spectra the two training sets perform similarly. Multimodal language models used for scientific decisions should be tested on equivalent representations and on missing evidence, not only on average accuracy. We release the tasks, per-response results and analysis code.
Oct 27, 2026

Zero-Shot Multimodal Event Spotting in Soccer Broadcasts

Automatic soccer event spotting supports highlight generation, video indexing, and sports analytics, but existing methods depend on large annotated datasets and soccer-specific training. We present a zero-shot framework that detects, classifies, and temporally localizes events in full-length soccer broadcasts without task-specific parameter training. Each match is segmented into overlapping clips represented by a compact visual storyboard and the aligned commentary transcript, which a general-purpose multimodal model processes jointly. On three manually annotated matches containing 130 events, the best configuration achieves 72.0% F1@5s and outperforms SoccerNet-trained CALF and NetVLAD++ in a cross-dataset setting, while the storyboard representation reduces inference cost by 72% compared with individual frames. We further analyze the effects of model family, input representation, and prompt optimization, and introduce an evidence-grounded excitement-scoring stage that ranks detected events with natural-language explanations for highlight selection.
Oct 27, 2026

DocXEval: A Systematic Evaluation of Explainability in Multimodal Document Understanding

Multimodal transformer models have driven unprecedented performance breakthroughs in modern document understanding. However, the inherent black-box nature of these models remains a major obstacle to their safe deployment in high-stakes sectors. Despite significant progress in explainability for vision and language tasks, explainable AI (XAI) research tailored to multimodal document understanding has received little attention. To address this gap, we present \ourframework{}, the first comprehensive explainability framework that adapts existing XAI methods and evaluation metrics to the multimodal (text, layout, and image) document understanding context, while introducing efficient multi-target adaptations of both methods and metrics for domain-specific tasks such as semantic entity recognition. Using \ourframework{}, we systematically evaluate 13 XAI methods across 3 document understanding tasks, 8 datasets, 4 models, and 17 evaluation metrics, and find that only a few methods perform consistently across all evaluation criteria, with Occlusion emerging as the most faithful and interpretable method overall. Through a dedicated meta-evaluation, we observe moderate to strong agreement among several faithfulness and complexity metrics while identifying certain ones that produce unreliable results. Finally, through multiple dataset-wide diagnostic analyses, we uncover key insights into model behavior, including the contribution of individual modalities across tasks and labels, spatial attribution patterns around target entities, and potential dataset biases. Overall, this work serves as both a benchmark and a practical guide for future XAI research in multimodal document understanding. To support reproducibility, we publicly release our code at https://github.com/saifullah3396/docxeval/.
Oct 27, 2026

MM-Net: A Multimodal Multi-Task Network for Sleep Staging and Sleep Apnea Assessment After Ischemic Stroke

Automated polysomnography after ischemic stroke is challenging because of altered sleep-related neural activity and concurrent respiratory disturbances. Existing stroke-specific approaches mainly address sleep-stage classification, leaving joint analysis of sleep architecture and respiratory events underexplored. Moreover, adapting pretrained EEG encoders introduces a representation inconsistency when downstream models are trained on embeddings from previously seen patients but evaluated on unseen individuals. We propose MM-Net, a multimodal multi-task framework that resolves this inconsistency through cross-fitting and ridge-based embedding alignment. Sleep-adapted EEG-EOG representations and engineered neural descriptors are fused with learned cardiorespiratory features through cross-attention and shared temporal modeling for simultaneous sleep staging and respiratory-event detection. On iSLEEPS (99 patients with subacute ischemic stroke), ten-fold patient-independent evaluation achieved 76.3% staging accuracy (Cohen's κκ = 0.656) and 0.825 respiratory-event AUC, outperforming the evaluated baselines under matched protocols. The sleep-adapted encoder improved the downstream model when its embeddings were aligned, raising staging accuracy by 1.4 percentage points over the frozen encoder (pp = 0.020) versus 0.2 points without alignment. MM-Net also achieved 86.4% staging accuracy on Sleep-EDF, above every single-EEG architecture reimplemented on the same recordings, and 80.3% staging accuracy with 0.889 respiratory AUC on ISRUC-I, exceeding its retrained staging baselines. Without retraining, the iSLEEPS-trained model retained respiratory AUCs of 0.833 on ISRUC-I and 0.841 on UCD. These findings support representation-consistent neural adaptation and joint physiological modeling across stroke, healthy-sleep, and sleep-disorder cohorts.
Oct 22, 2026

ConsMSA: Semantic Distribution Consistency Learning for Multimodal Sentiment Analysis

Multimodal sentiment analysis (MSA) aims to predict human sentiments by integrating signals from different modalities such as text, video, and audio. However, raw multimodal sequences often suffer from semantic inconsistency--exhibiting redundancy or conflicts within and across modalities--which hinders robust understanding and increases computational cost. To this end, we introduce ConsMSA, which explicitly formalizes semantic distribution consistency across both intra\textit{intra}- and inter\textit{inter}-modality, providing a consistency-aware mechanism for robust and efficient multimodal sentiment prediction. Specifically, ConsMSA projects multimodal token features into a shared sentiment space to compute an Intra- and Inter-modality Consistency Score (I2CSI^2CS). By coupling this score with predictive relevance, we formulate consistency-aware importance signals that are utilized: (i) as a consistency regularizer to align latent distributions during training, (ii) to derive semantic-aware weights for adaptive multimodal token reweighting, and (iii) as a practical criterion to prune redundant or conflicting tokens. Extensive experiments on CMU-MOSI, CMU-MOSEI, and CH-SIMS demonstrate that ConsMSA achieves state-of-the-art performance while remaining robust under aggressive token compression--retaining only 10% of tokens yields comparable accuracy. These results establish semantic distribution consistency as a promising foundation for synergizing predictive robustness with computational efficiency.
Oct 22, 2026

AGF-Net: A Robust Multimodal Emotion Recognition Model for Asynchronous Audio-Visual Signals Based on Predictive Dynamic Fusion with Negative Learning

This study develops an audio-visual emotion recognition system that integrates automatic utterance segmentation with multimodal fusion. Unlike conventional methods that rely on manually annotated utterance boundaries, the proposed system automatically generates candidate segments from continuous dialogue and extracts the corresponding audio and visual features for emotion classification. To address the challenges introduced by automatic segmentation, the proposed AGF-Net integrates Audio-Guided Fusion (AGF), Multimodal Negative Learning (MNL), and Predictive Dynamic Fusion (PDF) to improve cross-modal fusion, modality balance, and robustness to varying modality quality. Experimental results show that AGF-Net achieves 64.92% Accuracy and 65.47% F1-score under clean conditions, 62.13% Accuracy and 63.08% F1-score under moderate noise, and 58.19% Accuracy and 59.41% F1-score under higher noise. Overall, the proposed framework reduces reliance on manually annotated utterance boundaries and maintains competitive performance under automatically segmented and noisy audio-visual inputs.
Oct 17, 2026

Cross-Dimension and Cross-Modality Relational Tensor Representation Learning for Multimodal Health Data Modeling

Multimodal health data integrate information from heterogeneous clinical sources, creating substantial opportunities for data-driven decision support. Realizing these opportunities, however, requires effectively modeling complex dependencies that arise both within individual modalities and across modalities. This paper proposes Relational Tensor Representation (RTR), a unified framework that jointly learns dimension-specific tensor representations and cross-dimensional and cross-modality relationships. RTR introduces learnable relational bridge matrices that maintain relational information while preserving the structural identities of tensor dimensions and modalities. A structured fusion mechanism further integrates factor-level and relationship-level information into compact representations for downstream prediction. Experiments on simulated data and two multimodal neurophysiological datasets, covering emotion recognition and brain activity classification, demonstrate that RTR consistently outperforms representative tensor learning, multimodal fusion, and Transformer-based baselines. These results highlight the value of explicitly modeling cross-dimensional and cross-modality relationships for learning informative representations from heterogeneous multimodal tensor data. The source code is planned for public release to support reproducibility and future research.
Oct 16, 2026

CIIR - Class-Indexed Interface Representation Learning

Multimodal learning commonly assumes paired or synchronized observations, yet independently collected datasets rarely provide such correspondence. We introduce CIIR, a modular encoder--decoder framework that connects heterogeneous datasets through a class-indexed address space while preserving each dataset's native input format and output module. A known local index selects address kk, and an input-dependent 128-dimensional address representation Z_k(x) is written to that address. To demonstrate CIIR, we learn this address space from handwritten digits, clothing images, spoken-digit spectrograms, and hand signs without synchronized or instance-matched samples. Newly trained label-free receivers recover the requested addresses with 94.7% macro-average Address Realization Accuracy. We then attach the unseen SVHN dataset after freezing the base address space and train only lightweight adapters, achieving 90.2% MNIST>SVHN and 82.3% SVHN>MNIST end-to-end Address Transfer Accuracy without target-side decoder labels. To demonstrate generalizability of CIIR, we reuse the ten addresses as code elements in collision-free arithmetic and hierarchical address codes for all 100 CIFAR-100 classes and evaluate direct arithmetic-code realization through the existing receivers. These results establish CIIR as a compact interface for modular cross-dataset communication and post-hoc dataset attachment without exhaustive paired multimodal supervision.
Oct 15, 2026

TWIN-MoE: Two-stage Intra-Expert Imputation Network for Incomplete Multimodal Learning

Recent Mixture-of-Experts (MoE) approaches address missing modalities either by routing each missingness pattern to a dedicated expert, which scales poorly as the number of modalities grows, or by imputing the missing modality from representations contributed by multiple experts, which requires a costly retrieval mechanism. We show that routing induces expert-conditioned representation neighborhoods and derive an imputation-error bound controlled by their local feature dispersion. When these expert-conditioned neighborhoods are more compact than the global representation space, intra-expert aggregation admits a tighter worst-case reconstruction bound than unconstrained global aggregation. Guided by this analysis, we propose TWIN, a two-stage intra-expert imputation framework. Pre-imputation generates a context-conditioned estimate in the encoder embedding space and radially soft-clips it using modality-specific statistics maintained in an observed FIFO queue, this ensures that the pre-imputed modality stays in empirically calibrated region of the target modality space. Post-imputation refines the missing representation after MoE by aggregating expert-conditioned observed representations within the same sample. Across four missing-modality benchmarks under natural, random, and asymmetric dropout protocols, TWIN achieves state-of-the-art performance while using substantially fewer parameters than competing MoE-based imputation methods.
Oct 15, 2026

FlexUMM: Flexible Representation Learning for Unified Multimodal Models

Unified multimodal models (UMMs) can integrate understanding and generation within a single backbone. However, it is commonly known that these two capabilities can cause conflicting representations, leading to a central question: how representation and computation should be shared across functionalities? Existing UMMs often rely on a largely fixed structure, where different inputs and functionalities follow the same depth and parameter-sharing pattern. We argue that this assumption is too restrictive: more effective unification requires sharing the computation that supports cross-functional communication while specializing the capacity that becomes task- or input-specific. We introduce FlexUMM, a flexible UMM framework that dynamically decides when to preserve shared computation, when to route specialized capacity, and when to stop further processing. Evidence shows that this flexible allocation improves UMMs along two complementary axes. It improves accuracy by reducing harmful competition between functionalities, and it improves efficiency by avoiding unnecessary computation once an input representation is sufficiently mature. Beyond simple architectural unification from the pre-trained models, FlexUMM works as a model-agnostic, functional plug-in to make existing UMMs more unified, stronger, and efficient. Code is available at: https://github.com/AIFrontierLab/TorchUMM/tree/main/src/umm/post_training/flexumm.
Oct 15, 2026

FusionAudit: Pathway-Conditioned Robustness Auditing for Native Multimodal Models

As vision-language models (VLMs) transition from frozen-encoder pipelines to native multimodal architectures, the visual channel increasingly serves as both a perceptual input and a carrier of rendered text that competes with user prompts. Current single-attack, pixel-centric evaluations fail to capture this complexity: they collapse distinct visual attack surfaces into a single scalar and ignore how architectural shifts alter vulnerability. We introduce PaRMA, an evidence-based auditing framework that defines pathway-conditioned risk over two primary image-level interventions: pixel perturbation and rendered-text insertion. Instead of arbitrary attack enumeration, PaRMA uses pathway-specific diagnostics to expose architecture-specific blind spots. Notably, we uncover a see-but-not-deceived signature where native VLMs attend heavily to typographic overlays but visually reject them, and we identify an objective mismatch between encoder-level and answer-level attacks in deep-fusion models. To exploit these findings, we instantiate FusionAudit, introducing two diagnostic-motivated attacks: MANTIS (margin-aware pixel perturbation) and Nat-Evid (host-congruent semantic insertion). Evaluating 18 open-weight VLMs across a novel nativeness taxonomy, we demonstrate that legacy single-attack rankings (e.g., PGD) are unstable for native models. Our diagnostic-driven attacks substantially raise the estimated residual risk, showing that PaRMA provides a falsifiable, evidence-driven auditing methodology for uncovering vulnerabilities missed by standard benchmarks.
Oct 15, 2026

TimeCap: A Scalable Open-Source Pipeline for Captioning Wearable Health Time-Series Data

Time-series data generated continuously across industrial, environmental, and especially health domains produce multimodal signals whose scale and heterogeneity make their processing and interpretation particularly challenging. Time-Series Language Models (TSLMs) promise scalable interpretation of these signals, yet remain bottlenecked by a fundamental lack of text-annotated time series pairs that are signal-grounded, reproducible, and extensible across modalities and schemas. Existing paired corpora are either closed or limited in scale and modality coverage. Annotation with large language models could scale them up, but its outputs cannot be regenerated exactly, which prevents others from re-deriving, extending, or auditing the resulting corpus. To address this, we release TimeCap, an open deterministic captioning pipeline with three reusable extractor tiers and a cross-channel synthesis layer that operates directly on the signal with no language model in the loop, extensible to new schemas by writing channel definitions alone. This positions TimeCap as a reproducible mechanism for future dataset generation rather than a one-off release. Applying TimeCap to OpenMHC, the MyHeart Counts wearable dataset, yields MHCap, the largest open caption layer for multimodal wearable time-series data. MHCap spans 19 Apple HealthKit channels and 7,005,763 captions over 908,944 person-days. To showcase the impact of the corpus, we train OpenTSLM, a time-series language model, on MHCap and benchmark on health-outcome prediction. Caption-augmented training (OpenTSLM-Cap) outperforms a label-only baseline on 29 of 32 per-user binary evaluations across 8 tasks.
Oct 14, 2026

MSP: Modality Self-Play for Training Multimodal LLMs Without Paired Cross-Modal Data

Large multimodal models have achieved remarkable progress by scaling model capacity and training on massive paired multimodal data. However, this paradigm introduces a fundamental bottleneck: aligned multimodal datasets are expensive to curate, difficult to scale across diverse modality combinations, and increasingly constrain further progress. We introduce Modality Self-Play (MSP), a training framework for multimodal foundation models based on proxy-mediated modality composition. Instead of relying on paired cross-modal data, MSP uses anchor-supporting proxy textual descriptions of unseen modalities as semantic bridges during training. Given an observed modality (e.g., image or audio), the model generates proxy descriptions for complementary unseen modalities, which are then encoded and aligned with the observed modality through learned bridges, enabling cross-modal interaction without real paired non-text supervision. Our central hypothesis is that proxy text, combined with contrastively trained pretrained encoders, provides sufficient semantic structure for multimodal composition to emerge. We show that MSP enables zero-shot modality composition: models trained only with a real anchor modality and accompanying proxy descriptions can integrate and reason over multiple modalities at inference time despite never observing real paired non-text data during training. Empirically, MSP achieves state-of-the-art or competitive performance on multiple multimodal benchmarks. Ablation studies further demonstrate that learned bridges improve alignment, while anchor-supporting proxy descriptions enable effective cross-modal composition. MSP removes the need for real paired non-text (e.g., audio-vision) supervision, while continuing to rely on modality-text pairs, pretrained contrastive encoders, and modality-distribution bridge calibration.
Oct 14, 2026

Task-Centric Multimodal Learning for Cognitive Impairment Detection

Dementia affects millions of older adults worldwide, yet early detection remains difficult outside clinical settings. Traditional models primarily rely on text or voice-based responses, capturing only a limited view of human cognition despite the inherently multimodal nature of human expression. This work proposes a contactless, multimodal screening framework that leverages video, audio, and text signals collected during structured voice-based interactions with Amazon Alexa. By incorporating these modalities, our approach captures a more holistic representation of human behavior encompassing speech patterns, vocal characteristics, facial expressions, and linguistic structure, which may enable more accurate detection of cognitive decline. Different participants completed three conversational tasks designed to probe spontaneous speech, comprehension, and recall. We pursued three objectives: (1) evaluate which of the three Alexa tasks provides the most informative cognitive signals, (2) compare the effectiveness of unimodal models, and (3) assess the benefits of multimodal fusion models. Task-level analysis shows that the video model achieves its best performance on Command-Reading, with an accuracy of 0.65 (AUC 0.69), while the audio model performs best on Question-Answering, reaching an accuracy of 0.67 (AUC 0.73). In contrast, the multimodal model demonstrates consistently stable performance across tasks, achieving an accuracy of 0.76 (AUC 0.75) on Command-Reading and 0.76 (AUC 0.73) on Question-Answering. These findings demonstrate that multimodal behavioral cues captured through naturalistic interactions can reliably distinguish MCI from healthy aging, highlighting the limitations of traditional diagnostic paradigms and supporting the shift toward multimodal, real-world cognitive assessment frameworks.
Oct 14, 2026

CogDrift-4D: Hierarchical Multimodal Modeling of Temporal Behavioral Decay for MCI Detection

Early detection of mild cognitive impairment remains a critical bottleneck in dementia care, demanding scalable tools capable of capturing subtle behavioral decay. To capture the subtle temporal changes of cognitive decline, we introduce a behavioral dataset collected across seven unscripted multimodal interactions over an 18-month horizon. This dataset formalizes continuous, multi-session interactions into a dense 4D tensor (Batch * Session * Time * Feature) that explicitly preserves concurrent local dynamics alongside long-term changes. To effectively process these complex behavioral patterns, we propose CogDrift-4D: a novel, hierarchical 4D Transformer framework. To capture the complex trajectory of cognitive decline, CogDrift-4D implements a dual-level behavioral fatigue and drift modeling architecture. At the intra-session level, the model computes localized least-squares regression slopes to quantify the acute ``fading'' of frame-level facial action units and acoustic energy. At the inter-session level, a recursive cross-attention feedback encoder acts as a cumulative fatigue accumulator, mapping the patient's continuous behavioral drift across successive sessions. To mitigate environmental noise and track failures across sessions, a localized Quality Gate dynamically filters low-integrity features before final fusion, also an early fusion that integrates semantic insights with raw physical dynamics. Evaluated on structured verbal and linguistic interactions, our architecture successfully detects MCI with a classification accuracy of 0.733 +- 0.13 and an F1-score of 0.716 +- 0.14, outperforming advanced baselines. The framework provides a mathematically rigorous, scalable, and robust foundation for home-based, real-time dementia screening.
Oct 11, 2026

Learning Multimodal Sensor Fusion with JEPA Representations for Breathing Activity and Pose Recognition

Predictive video representations do not by themselves specify how a multimodal system should handle missing sensors. We transfer frozen V-JEPA2 features to a set of heterogeneous sensors comprised by RGB video, depth video, and radar time series. A shared fusion model uses modality queries, fixed sensor slots, and explicit availability to classify breathing activity and pose from all seven non-empty sensor subsets. The contribution is a controlled evaluation of subset supervision, pooling capacity, and representation transfer on 39 OMuSense-23 subjects with five subject-group folds and three seeds. With matched downstream parameter counts, the final model reaches recording-level activity macro-F1 of 75.2375.23%\pm0.60 averaged across seven subsets. Supervising all subsets in each batch improves the matched query-pooling control by 1.923 percentage points (conditional 95% subject-bootstrap interval [0.647, 3.209]), while query pooling shows no clear advantage over matched mean pooling under balanced supervision. Radar probes support transfer beyond pooled wavelet pixels, 64-token inference approximately preserves mean activity performance, and the tested Consensus-JEPA loss adds no measured benefit against its exact control.
Oct 10, 2026

Prior-Fitted and Frozen Scorers for Multimodal Contrastive Learning

Multimodal contrastive learning typically relies on manually specified scoring functions, such as cosine similarity, to quantify the similarity between representations from different modalities. Several such functions have been recently proposed to extend contrastive learning beyond two modalities. In principle, however, any function that assigns high compatibility to matched samples and low compatibility to mismatched ones could serve as a scorer. Motivated by this broader view, we instantiate the scorer with a Transformer and find that even a randomly initialized model, when kept frozen, can outperform strong handcrafted multimodal contrastive baselines, whereas optimizing the scorer jointly with the encoders is less effective. Building on this observation, we introduce the Prior-Fitted Scorer (PFS): a Transformer scorer pre-trained on synthetic multimodal data generated from structural causal model (SCM) priors before transferring it to real-world data. The SCM exposes the scorer to diverse dependency structures, including shared, pairwise, and private information, providing a structured prior over higher-order cross-modal similarity. Across four real-world datasets, frozen Transformer scorers are competitive to contrastive baselines, and prior-fitting further improves the random initialization. More broadly, our results suggest that multimodal contrastive learning need not rely on manual similarity functions: Transformers can provide strong inductive biases, while causal pretraining can further shape these biases.