Multimodal Pretraining

Monocular 3D object detection typically relies on pseudo-labeling techniques to reduce dependency on real-world annotations. Recent advances demonstrate that deterministic linguistic cues can serve as effective auxiliary weak supervision signals, providing complementary semantic context. However, hand-crafted textual descriptions struggle to capture the inherent visual diversity of individuals across scenes, limiting the model's ability to learn scene-aware representations. To address this challenge, we propose Visual-referred Probabilistic Prompt Learning (VirPro), an adaptive multi-modal pretraining paradigm that can be seamlessly integrated into diverse weakly supervised monocular 3D detection frameworks. Specifically, we generate a diverse set of learnable, instance-conditioned prompts across scenes and store them in an Adaptive Prompt Bank (APB). Subsequently, we introduce Multi-Gaussian Prompt Modeling (MGPM), which incorporates scene-based visual features into the corresponding textual embeddings, allowing the text prompts to express visual uncertainties. Then, from the fused vision-language embeddings, we decode a prompt-targeted Gaussian, from which we derive a unified object-level prompt embedding for each instance. RoI-level contrastive matching is employed to enforce modality alignment, bringing embeddings of co-occurring objects within the same scene closer in the latent space, thus enhancing semantic coherence. Extensive experiments on the KITTI benchmark demonstrate that integrating our pretraining paradigm consistently yields substantial performance gains, achieving up to a 4.8% average precision improvement than the baseline. Code is available at https://github.com/AustinLCP/VirPro.

Multimodal large language models (MLLMs) have achieved rapid progress, yet their scaling behavior remains less clearly characterized and often less predictable than that of text-only LLMs. Increasing model size and task diversity often yields diminishing returns. In this work, we argue that the primary bottleneck in multimodal scaling is not task format, but knowledge density in training data. We first show that task-specific supervision such as Visual Question Answering (VQA) contributes little incremental semantic information beyond image captions: VQA signals can be reconstructed from captions with negligible performance loss. We then demonstrate that increasing knowledge density -- through structured caption enrichment and cross-modal knowledge injection -- leads to consistent performance improvements across multimodal and downstream benchmarks. Across controlled experiments, performance correlates more strongly with semantic coverage than with task diversity. These findings suggest that current MLLMs fail to scale primarily because training data lacks sufficient knowledge coverage. We advocate for knowledge-centric multimodal training as a principled foundation for scalable multimodal models.

The current expressway operation relies on rule-based and isolated models, which limits the ability to jointly analyze knowledge across different systems. Meanwhile, Large Language Models (LLMs) are increasingly applied in intelligent transportation, advancing traffic models from algorithmic to cognitive intelligence. However, general LLMs are unable to effectively understand the regulations and causal relationships of events in unconventional scenarios in the expressway field. Therefore, this paper constructs a pre-trained multimodal large language model (MLLM) for expressways, ExpressMind, which serves as the cognitive core for intelligent expressway operations. This paper constructs the industry's first full-stack expressway dataset, encompassing traffic knowledge texts, emergency reasoning chains, and annotated video events to overcome data scarcity. This paper proposes a dual-layer LLM pre-training paradigm based on self-supervised training and unsupervised learning. Additionally, this study introduces a Graph-Augmented RAG framework to dynamically index the expressway knowledge base. To enhance reasoning for expressway incident response strategies, we develop a RL-aligned Chain-of-Thought (RL-CoT) mechanism that enforces consistency between model reasoning and expert problem-solving heuristics for incident handling. Finally, ExpressMind integrates a cross-modal encoder to align the dynamic feature sequences under the visual and textual channels, enabling it to understand traffic scenes in both video and image modalities. Extensive experiments on our newly released multi-modal expressway benchmark demonstrate that ExpressMind comprehensively outperforms existing baselines in event detection, safety response generation, and complex traffic analysis. The code and data are available at: https://wanderhee.github.io/ExpressMind/.

Multimodal Large Language Models (MLLMs) demonstrate exceptional semantic reasoning but struggle with 3D spatial perception when restricted to pure RGB inputs. Despite leveraging implicit geometric priors from 3D reconstruction models, image-based methods still exhibit a notable performance gap compared to methods using explicit 3D data. We argue that this gap does not arise from insufficient geometric priors, but from a misalignment in the training paradigm: text-dominated fine-tuning fails to activate geometric representations within MLLMs. Existing approaches typically resort to naive feature concatenation and optimize directly for downstream tasks without geometry-specific supervision, leading to suboptimal structural utilization. To address this limitation, we propose GAP-MLLM, a Geometry-Aligned Pre-training paradigm that explicitly activates structural perception before downstream adaptation. Specifically, we introduce a visual-prompted joint task that compels the MLLMs to predict sparse pointmaps alongside semantic labels, thereby enforcing geometric awareness. Furthermore, we design a multi-level progressive fusion module with a token-level gating mechanism, enabling adaptive integration of geometric priors without suppressing semantic reasoning. Extensive experiments demonstrate that GAP-MLLM significantly enhances geometric feature fusion and consistently enhances performance across 3D visual grounding, 3D dense captioning, and 3D video object detection tasks.

Unified Multimodal Models (UMMs) are often constrained by the pre-training of their $\textbf{visual generation components}$, which typically relies on inefficient paradigms and scarce, high-quality text-image paired data. In this paper, we systematically analyze pre-training recipes for $\textbf{UMM visual generation}$ and identify these two issues as the major bottlenecks. To address them, we propose $\textbf{Image-Only Training for UMMs (IOMM)}$, a data-efficient two-stage training framework. The first stage pre-trains the visual generative component $\textbf{exclusively}$ using abundant unlabeled image-only data, thereby removing the dependency on paired data $\textbf{for this costly phase}$. The second stage fine-tunes the model using a mixture of unlabeled images and a small curated set of text-image pairs, leading to improved instruction alignment and generative quality. Extensive experiments show that IOMM not only improves training efficiency but also achieves state-of-the-art (SOTA) performance. For example, our IOMM-B (3.6B) model was trained from scratch using only $\sim \textbf{1050}$ H800 GPU hours (with the vast majority, $\textbf{1000}$ hours, dedicated to the efficient $\textbf{image-only pre-training stage}$). It achieves $\textbf{0.89}$ on GenEval and $\textbf{0.55}$ on WISE--surpassing strong baselines such as BAGEL-7B (0.82 & 0.55) and BLIP3-o-4B (0.84 & 0.50). Code is available https://github.com/LINs-lab/IOMM.

Context-aided forecasting (CAF) holds promise for integrating domain knowledge and forward-looking information, enabling AI systems to surpass traditional statistical methods. However, recent empirical studies reveal a puzzling gap: multimodal models often fail to outperform their unimodal counterparts. We hypothesize that this underperformance stems from poor context quality in existing datasets, as verification is challenging. To address these limitations, we introduce a semi-synthetic data augmentation method that generates contexts both descriptive of temporal dynamics and verifiably complementary to numerical histories. This approach enables massive-scale dataset creation, resulting in CAF-7M, a corpus of 7 million context-augmented time series windows, including a rigorously verified test set. We demonstrate that semi-synthetic pre-training transfers effectively to real-world evaluation, and show clear evidence of context utilization. Our results suggest that dataset quality, rather than architectural limitations, has been the primary bottleneck in context-aided forecasting.

Multimodal language models have demonstrated strong semantic understanding and reasoning over physiological and behavioral signals captured from diverse healthcare sensors. However, existing sensor-language models can only process a fixed number of sensor channels at a fixed sampling rate, making them difficult to adapt to new or different sensor configurations. This inflexibility prevents models from generalizing across heterogeneous health sensing ecosystems, where data spans high-frequency wearable streams to sparse, day-scale clinical signals. To bridge this gap, we introduce SLIP (Sensor Language-Informed Pretraining), an open-source framework for learning language-aligned representations that generalize across diverse sensor setups. SLIP integrates contrastive alignment with sensor-conditioned captioning, facilitating both discriminative understanding and generative reasoning. By repurposing a pretrained decoder-only language model via cross-attention and introducing a flexible patch embedder, SLIP transfers to new sensor configurations at inference without retraining, regardless of their native sampling rate or input length. Across 11 datasets, SLIP demonstrates superior performance in retrieval, signal captioning, and question answering. It achieves a 77.14% average linear-probing accuracy, a 3.56% improvement over the strongest baseline (SensorLM). Beyond classification, SLIP supports open-vocabulary sensor captioning and question answering without any architectural modifications. Our work offers broad implications for the development of sensor-language models, utilizing cross-modal supervision to ensure more robust generalizability in health sensing applications.

The paradigm of Multimodal Large Language Models (MLLMs) offers a promising blueprint for advancing the electromagnetic (EM) domain. However, prevailing approaches often deviate from the native MLLM paradigm, instead using task-specific or pipelined architectures that lead to fundamental limitations in model performance and generalization. Fully realizing the MLLM potential in EM domain requires overcoming three main challenges: (1) Data. The scarcity of high-quality datasets with paired EM signals and descriptive text annotations used for MLLMs pre-training; (2) Benchmark. The absence of comprehensive benchmarks to systematically evaluate and compare the performance of models on EM signal-to-text tasks; (3) Model. A critical fragility in low Signal-to-Noise Ratio (SNR) environments, where critical signal features can be obscured, leading to significant performance degradation. To address these challenges, we introduce a tripartite contribution to establish a foundation for MLLMs in the EM domain. First, to overcome data scarcity, we construct and release EM-100k, a large-scale dataset comprising over 100,000 EM signal-text pairs. Second, to enable rigorous and standardized evaluation, we propose EM-Bench, the most comprehensive benchmark featuring diverse downstream tasks spanning from perception to reasoning. Finally, to tackle the core modeling challenge, we present MERLIN, a novel training framework designed not only to align low-level signal representations with high-level semantic text, but also to explicitly enhance model robustness and performance in challenging low-SNR environments. Comprehensive experiments validate our method, showing that MERLIN is state-of-the-art in the EM-Bench and exhibits remarkable robustness in low-SNR settings.

Contrastive Language-Image Pre-training (CLIP) has shown impressive performance in aligning visual and textual representations. Recent studies have extended this paradigm to 3D vision to improve scene understanding for autonomous driving. A common strategy is to employ pairwise cosine similarity between modalities to guide the training of a 3D encoder. However, considering the similarity between individual modality pairs rather than all modalities jointly fails to ensure consistent and unified alignment across the entire multimodal space. In this paper, we propose a Contrastive Tensor Pre-training (CTP) framework that simultaneously aligns multiple modalities in a unified embedding space to enhance end-to-end autonomous driving. Compared with pairwise cosine similarity alignment, our method extends the 2D similarity matrix into a multimodal similarity tensor. Furthermore, we introduce a tensor loss to enable joint contrastive learning across all modalities. For experimental validation of our framework, we construct a text-image-point cloud triplet dataset derived from existing autonomous driving datasets. The results show that our proposed unified multimodal alignment framework achieves favorable performance for both scenarios: (i) aligning a 3D encoder with pretrained CLIP encoders, and (ii) pretraining all encoders from scratch.

We present DeepEarth, a self-supervised multi-modal world model with Earth4D, a novel planetary-scale 4D space-time positional encoder. Earth4D extends 3D multi-resolution hash encoding to include time, efficiently scaling across the planet over centuries with sub-meter, sub-second precision. Multi-modal encoders (e.g. vision-language models) are fused with Earth4D embeddings and trained via masked reconstruction. We demonstrate Earth4D's expressive power by achieving state-of-the-art performance on an ecological forecasting benchmark. Earth4D with learnable hash probing surpasses a multi-modal foundation model pre-trained on substantially more data. Access open source code and download models at: https://github.com/legel/deepearth

Image-based shape retrieval (IBSR) aims to retrieve 3D models from a database given a query image, hence addressing a classical task in computer vision, computer graphics, and robotics. Recent approaches typically rely on bridging the domain gap between 2D images and 3D shapes based on the use of multi-view renderings as well as task-specific metric learning to embed shapes and images into a common latent space. In contrast, we address IBSR through large-scale multi-modal pretraining and show that explicit view-based supervision is not required. Inspired by pre-aligned image--point-cloud encoders from ULIP and OpenShape that have been used for tasks such as 3D shape classification, we propose the use of pre-aligned image and shape encoders for zero-shot and standard IBSR by embedding images and point clouds into a shared representation space and performing retrieval via similarity search over compact single-embedding shape descriptors. This formulation allows skipping view synthesis and naturally enables zero-shot and cross-domain retrieval without retraining on the target database. We evaluate pre-aligned encoders in both zero-shot and supervised IBSR settings and additionally introduce a multi-modal hard contrastive loss (HCL) to further increase retrieval performance. Our evaluation demonstrates state-of-the-art performance, outperforming related methods on $Acc_{Top1}$ and $Acc_{Top10}$ for shape retrieval across multiple datasets, with best results observed for OpenShape combined with Point-BERT. Furthermore, training on our proposed multi-modal HCL yields dataset-dependent gains in standard instance retrieval tasks on shape-centric data, underscoring the value of pretraining and hard contrastive learning for 3D shape retrieval. The code will be made available via the project website.

Radiology foundation models (RFMs) have largely inherited the scale-first recipe of natural-image vision--language pre-training. This recipe is difficult to deploy in 3D radiology, where training corpora are smaller, reports vary across institutions, and receiving hospitals often need local adaptation under privacy and compute constraints. We ask whether routine radiology reports can instead be converted into auditable diagnostic supervision that shapes the image encoder, text encoder, aligned space, and local-adaptation procedure. We develop GreenRFM, a supervision-centric pre-training framework organized around four empirical principles: More distilled, Ubiquitous, Semantic-enforcing, and Task-aligning (MUST) supervision. These principles convert noisy reports into structured diagnostic signals and use them to learn discriminative unimodal encoders plus an aligned image--text space for diagnosis-centered multimodal use. GreenRFM requires 24 GPU-hours on a single 24GB GPU (lightweight variant: 6GB VRAM, 4~hours) and reaches a zero-shot CT-RATE AUC of 84.8. Evaluations using more than 200,000 volumes from six institutions and two modalities show transfer to private clinical cohorts and to musculoskeletal MRI. On a local institutional cohort, computationally feasible retraining raises macro-AUC from 70.5 to 82.1. The aligned space also improves hepatocellular-carcinoma microvascular-invasion prediction and trans-arterial chemoembolization response analysis over established clinical scores. These results support supervision-centric pre-training as a practical route to resource-efficient, locally adaptable, diagnosis-centered radiology vision--language representations.

Multimodal research has predominantly focused on single-image reasoning, with limited exploration of multi-image scenarios. Recent models have sought to enhance multi-image understanding through large-scale pretraining on interleaved image-text datasets. However, most Vision-Language Models (VLMs) are trained primarily on English datasets, leading to inadequate representation of Indian languages. To address this gap, we introduce the Chitrakshara dataset series, covering 11 Indian languages sourced from Common Crawl. It comprises (1) Chitrakshara-IL, a large-scale interleaved pretraining dataset with 193M images, 30B text tokens, and 50M multilingual documents, and (2) Chitrakshara-Cap, which includes 44M image-text pairs with 733M tokens. This paper details the data collection pipeline, including curation, filtering, and processing methodologies. Additionally, we present a comprehensive quality and diversity analysis to assess the dataset's representativeness across Indic languages and its potential for developing more culturally inclusive VLMs.

Zero-shot MRI reconstruction relies on generative priors, but single-modality unconditional priors produce hallucinations under severe ill-posedness. In many clinical workflows, complementary MRI acquisitions (e.g. high-quality structural scans) are routinely available, yet existing reconstruction methods lack mechanisms to leverage this additional information. We propose MPFlow, a zero-shot multi-modal reconstruction framework built on rectified flow that incorporates auxiliary MRI modalities at inference time without retraining the generative prior to improve anatomical fidelity. Cross-modal guidance is enabled by our proposed self-supervised pretraining strategy, Patch-level Multi-modal MR Image Pretraining (PAMRI), which learns shared representations across modalities. Sampling is jointly guided by data consistency and cross-modal feature alignment using pre-trained PAMRI, systematically suppressing intrinsic and extrinsic hallucinations. Extensive experiments on HCP and BraTS show that MPFlow matches diffusion baselines on image quality using only 20% of sampling steps while reducing tumor hallucinations by more than 15% (segmentation dice score). This demonstrates that cross-modal guidance enables more reliable and efficient zero-shot MRI reconstruction.

The visual world offers a critical axis for advancing foundation models beyond language. Despite growing interest in this direction, the design space for native multimodal models remains opaque. We provide empirical clarity through controlled, from-scratch pretraining experiments, isolating the factors that govern multimodal pretraining without interference from language pretraining. We adopt the Transfusion framework, using next-token prediction for language and diffusion for vision, to train on diverse data including text, video, image-text pairs, and even action-conditioned video. Our experiments yield four key insights: (i) Representation Autoencoder (RAE) provides an optimal unified visual representation by excelling at both visual understanding and generation; (ii) visual and language data are complementary and yield synergy for downstream capabilities; (iii) unified multimodal pretraining leads naturally to world modeling, with capabilities emerging from general training; and (iv) Mixture-of-Experts (MoE) enables efficient and effective multimodal scaling while naturally inducing modality specialization. Through IsoFLOP analysis, we compute scaling laws for both modalities and uncover a scaling asymmetry: vision is significantly more data-hungry than language. We demonstrate that the MoE architecture harmonizes this scaling asymmetry by providing the high model capacity required by language while accommodating the data-intensive nature of vision, paving the way for truly unified multimodal models.

Heterogeneous multi-modal remote sensing object detection aims to accurately detect objects from diverse sensors (e.g., RGB, SAR, Infrared). Existing approaches largely adopt a late alignment paradigm, in which modality alignment and task-specific optimization are entangled during downstream fine-tuning. This tight coupling complicates optimization and often results in unstable training and suboptimal generalization. To address these limitations, we propose BabelRS, a unified language-pivoted pretraining framework that explicitly decouples modality alignment from downstream task learning. BabelRS comprises two key components: Concept-Shared Instruction Aligning (CSIA) and Layerwise Visual-Semantic Annealing (LVSA). CSIA aligns each sensor modality to a shared set of linguistic concepts, using language as a semantic pivot to bridge heterogeneous visual representations. To further mitigate the granularity mismatch between high-level language representations and dense detection objectives, LVSA progressively aggregates multi-scale visual features to provide fine-grained semantic guidance. Extensive experiments demonstrate that BabelRS stabilizes training and consistently outperforms state-of-the-art methods without bells and whistles. Code: https://github.com/zcablii/SM3Det.

Multimodal embedding models, rooted in multimodal large language models (MLLMs), have yielded significant performance improvements across diverse tasks such as retrieval and classification. However, most existing approaches rely heavily on large-scale contrastive learning, with limited exploration of how the architectural and training paradigms of MLLMs affect embedding quality. While effective for generation, the causal attention and next-token prediction paradigm of MLLMs does not explicitly encourage the formation of globally compact representations, limiting their effectiveness as multimodal embedding backbones. To address this, we propose CoCoA, a Content reconstruction pre-training paradigm based on Collaborative Attention for multimodal embedding optimization. Specifically, we restructure the attention flow and introduce an EOS-based reconstruction task, encouraging the model to reconstruct input from the corresponding <EOS> embeddings. This drives the multimodal model to compress the semantic information of the input into the <EOS> token, laying the foundations for subsequent contrastive learning. Extensive experiments on MMEB-V1 demonstrate that CoCoA built upon Qwen2-VL and Qwen2.5-VL significantly improves embedding quality. Results validate that content reconstruction serves as an effective strategy to maximize the value of existing data, enabling multimodal embedding models generate compact and informative representations, raising their performance ceiling.

Large Language Models have achieved remarkable success in language understanding and reasoning, and their multimodal extensions enable comprehension of images, video, and audio. Inspired by this, foundation models for brain functional connectivity networks derived from resting-state fMRI have shown promise in clinical tasks. However, existing methods do not align FCNs with the text modality, limiting the ability of LLMs to directly understand FCNs. To address this, we propose FCN-LLM, a framework that enables LLMs to understand FCNs through graph-level, multi-task instruction tuning. Our approach employs a multi-scale FCN encoder capturing brain-region, functional subnetwork, and whole-brain features, projecting them into the semantic space of LLM. We design multi-paradigm instruction tasks covering 19 subject-specific attributes across demographics, phenotypes, and psychiatric conditions. A multi-stage learning strategy first aligns FCN embeddings with the LLM and then jointly fine-tunes the entire model to capture high-level semantic information. Experiments on a large-scale, multi-site FCN database show that FCN-LLM achieves strong zero-shot generalization on unseen datasets, outperforming conventional supervised and foundation models. This work introduces a new paradigm for integrating brain functional networks with LLMs, offering a flexible and interpretable framework for neuroscience.

MultiPUFFIN is a domain-informed multimodal foundation model for predicting thermophysical properties of small molecules, addressing a critical gap in chemical engineering, drug discovery, and materials science. Existing molecular foundation models pretrain on millions of molecules to learn general-purpose representations, but their standard MLP output layers impose no physical constraints, vapor pressure predictions may violate monotonic temperature dependence, and viscosity curves may lack the functional form required by process simulators. Domain-informed approaches that guarantee thermodynamic consistency have remained limited to single properties and small datasets, whereas multimodal foundation models have focused on biological activity rather than thermophysical properties. MultiPUFFIN fills this gap by fusing SMILES sequences, 2D molecular graphs, and 3D conformer geometries through bidirectional cross-modal attention and gated fusion, supplemented by auxiliary encoders for experimental conditions and molecular descriptors. The backbone is pretrained on 500,000 unlabelled PubChem molecules using three complementary self-supervised objectives. A condition-aware refinement stack of five conditioners (temperature, pH, pressure, polymorph, and measurement method) routes each property to a four-head tournament that selects the best-performing thermodynamically informed head for that property. MultiPUFFIN achieves a mean test R2 of 0.784 and outperforms fine-tuned ChemBERTa-2 on all nine properties despite training on roughly 2,000x fewer labeled molecules.

Biomedical multimodal assistants have the potential to unify radiology, pathology, and clinical-text reasoning, yet a critical deployment gap remains: top-performing systems are either closed-source or computationally prohibitive, precluding the on-premises deployment required for patient privacy and PHI compliance. We introduce MEDGPT-OSS, an open-weight, 20B-parameter generalist vision-language model designed to facilitate open research in clinical AI. Rather than relying on architectural complexity, MEDGPT-OSS pairs the GPT-oss language backbone with a visual front-end via a optimized, three-stage training curriculum. By progressively domain-adapting these modules through rigorous data curation and long-context multimodal alignment, we demonstrate that a 20B model can bridge the capacity gap. It successfully outperforms larger open medical models on out-of-distribution (OOD) multimodal reasoning and complex text-only clinical tasks. By unifying diverse modalities under a single instruction-following interface, MEDGPT-OSS maintains a parameter-efficient footprint fully compatible with commodity GPUs. We release the complete training recipe, open-weight checkpoints, and a rigorous evaluation harness to serve as a verifiable foundation for privacy-preserving, institution-specific clinical AI research.

Electroencephalography (EEG) foundation models have shown promise for learning generalizable representations, yet they remain sensitive to channel heterogeneity, such as changes in channel composition or ordering. We propose channel-aware multimodal EEG-text alignment contrastive language-image pretraining (CAMEL-CLIP), a contrastive EEG-text multimodal foundation model designed to be robust to heterogeneous channel configurations and widely applicable to diverse downstream tasks. CAMEL-CLIP introduces three key components: (1) channel attribute-based positional encoding, which identifies channels through semantic information; (2) dynamic channel projection, which generates variable-length embeddings by independently projecting each channel without feature compression; and (3) dual-level contrastive learning, which jointly performs channel-level and sample-level contrastive learning to capture both channel-specific and global signal characteristics. Experimental results demonstrate that CAMEL-CLIP achieves state-of-the-art performance under linear-probing and outperforms existing foundation models that rely on full-finetuning.

We present SleepLM, a family of sleep-language foundation models that enable human sleep alignment, interpretation, and interaction with natural language. Despite the critical role of sleep, learning-based sleep analysis systems operate in closed label spaces (e.g., predefined stages or events) and fail to describe, query, or generalize to novel sleep phenomena. SleepLM bridges natural language and multimodal polysomnography, enabling language-grounded representations of sleep physiology. To support this alignment, we introduce a multilevel sleep caption generation pipeline that enables the curation of the first large-scale sleep-text dataset, comprising over 100K hours of data from more than 10,000 individuals. Furthermore, we present a unified pretraining objective that combines contrastive alignment, caption generation, and signal reconstruction to better capture physiological fidelity and cross-modal interactions. Extensive experiments on real-world sleep understanding tasks verify that SleepLM outperforms state-of-the-art in zero-shot and few-shot learning, cross-modal retrieval, and sleep captioning. Importantly, SleepLM also exhibits intriguing capabilities including language-guided event localization, targeted insight generation, and zero-shot generalization to unseen tasks. All code and data will be open-sourced.

Medical diagnosis requires the effective synthesis of visual manifestations and clinical metadata. However, existing methods often treat metadata as isolated tags, failing to exploit the rich semantic knowledge embedded in clinical descriptions. We propose PRIMA (Pre-training with Risk-integrated Image-Metadata Alignment), a framework that integrates domain-specific knowledge into multi-modal representation learning. We first curate a corpus of risk--disease correlations via Retrieval-Augmented Generation (RAG) to refine Clinical ModernBERT, embedding diagnostic priors into the text encoder. To bridge the modality gap, we introduce a dual-encoder pre-training strategy utilizing DINOv3 and our refined Clinical ModernBERT, optimized by a suite of four complementary loss functions. These losses are designed to capture multi-granular semantic alignment and handle the ambiguity of clinical correlations through soft labels. Finally, we leverage Qwen3 to fuse these aligned features for precise disease classification. Extensive experiments demonstrate that PRIMA effectively harmonizes pixel-level features with abstract clinical expertise. Across PAD-UFES-20 and AQUA, PRIMA achieves average F1-scores of 72.04% and 85.22%, outperforming strong image-only, metadata-fusion, and medical vision-language baselines. Notably, our framework achieves strong performance without requiring massive data collection or exhaustive computational resources. Our code is available at https://github.com/yqwang01/PRIMA.

Brain foundation models have achieved remarkable advances across a wide range of neuroscience tasks. However, most existing models are limited to a single functional modality, restricting their ability to exploit complementary spatiotemporal dynamics and the collective data scale across different neuroimaging techniques. This limitation largely arises from severe semantic heterogeneity and resolution discrepancies among modalities. To address these challenges, we propose Brain-OF, an omnifunctional brain foundation model jointly pretrained on fMRI, EEG and MEG, capable of handling both unimodal and multimodal inputs within a unified framework. To reconcile heterogeneous spatiotemporal resolutions, we introduce the Any-Resolution Neural Signal Sampler, which projects diverse brain signals into a shared semantic space. To further manage semantic shifts, the Brain-OF backbone integrates DINT attention with a Sparse Mixture of Experts, where shared experts capture modality-invariant representations and routed experts specialize in modality-specific semantics. Furthermore, to explicitly internalize the characteristics of neural activity through self-supervised learning, we propose Masked Temporal-Frequency Modeling, a dual-domain pretraining objective that jointly reconstructs brain signals in both the time and frequency domains. Brain-OF is pretrained on a large-scale corpus comprising around 40 datasets and demonstrates superior performance across diverse downstream tasks, highlighting the benefits of joint multimodal integration and dual-domain pretraining.

Discrete diffusion models have emerged as strong alternatives to autoregressive language models, with recent work initializing and fine-tuning a base unimodal model for bimodal generation. Diverging from previous approaches, we introduce the first tri-modal masked diffusion model pretrained from scratch on text, image-text, and audio-text data. We systematically analyze multimodal scaling laws, modality mixing ratios, noise schedules, and batch-size effects, and we provide optimized inference sampling defaults. Our batch-size analysis yields a novel stochastic differential equation (SDE)-based reparameterization that eliminates the need for tuning the optimal batch size as reported in recent work. This reparameterization decouples the physical batch size, often chosen based on compute constraints (GPU saturation, FLOP efficiency, wall-clock time), from the logical batch size, chosen to balance gradient variance during stochastic optimization. Finally, we pretrain a preliminary 3B-parameter tri-modal model on 6.4T tokens, demonstrating the capabilities of a unified design and achieving strong results in text generation, text-to-image tasks, and text-to-speech tasks. Our work represents the largest-scale systematic open study of multimodal discrete diffusion models conducted to date, providing insights into scaling behaviors across multiple modalities.

Foundation models are transforming Earth Observation (EO), yet the diversity of EO sensors and modalities makes a single universal model unrealistic. Multiple specialized EO foundation models (EOFMs) will likely coexist, making efficient knowledge transfer across modalities essential. Most existing EO pretraining relies on masked image modeling, which emphasizes local reconstruction but provides limited control over global semantic structure. To address this, we propose a dual-teacher contrastive distillation framework for multispectral imagery that aligns the student's pretraining objective with the contrastive self-distillation paradigm of modern optical vision foundation models (VFMs). Our approach combines a multispectral teacher with an optical VFM teacher, enabling coherent cross-modal representation learning. Experiments across diverse optical and multispectral benchmarks show that our model adapts to multispectral data without compromising performance on optical-only inputs, achieving state-of-the-art results in both settings, with an average improvement of 3.64 percentage points in semantic segmentation, 1.2 in change detection, and 1.31 in classification tasks. This demonstrates that contrastive distillation provides a principled and efficient approach to scalable representation learning across heterogeneous EO data sources. Project page: \textcolor{magenta}{https://wolfilip.github.io/DEO/}.

We present Xray-Visual, a unified vision model architecture for large-scale image and video understanding trained on industry-scale social media data. Our model leverages over 15 billion curated image-text pairs and 10 billion video-hashtag pairs from Facebook and Instagram, employing robust data curation pipelines that incorporate balancing and noise suppression strategies to maximize semantic diversity while minimizing label noise. We introduce a three-stage training pipeline that combines self-supervised MAE, semi-supervised hashtag classification, and CLIP-style contrastive learning to jointly optimize image and video modalities. Our architecture builds on a Vision Transformer backbone enhanced with efficient token reorganization (EViT) for improved computational efficiency. Extensive experiments demonstrate that Xray-Visual achieves state-of-the-art performance across diverse benchmarks, including ImageNet for image classification, Kinetics and HMDB51 for video understanding, and MSCOCO for cross-modal retrieval. The model exhibits strong robustness to domain shift and adversarial perturbations. We further demonstrate that integrating large language models as text encoders (LLM2CLIP) significantly enhances retrieval performance and generalization capabilities, particularly in real-world environments. Xray-Visual establishes new benchmarks for scalable, multimodal vision models, while maintaining superior accuracy and computational efficiency.

We present TokaMind, to our knowledge the first open-source foundation model for tokamak plasma dynamics, based on a Multi-Modal Transformer (MMT) and pretrained on heterogeneous diagnostics from the publicly available MAST dataset. TokaMind supports multiple data modalities (time-series, 2D profiles, and videos) with different sampling rates, robust missing-signal handling, and efficient task adaptation via selectively loading and freezing four model components. To represent multi-modal signals, we use a lightweight fixed-basis Discrete Cosine Transform embedding (DCT3D) and provide a clean interface for alternative embeddings (e.g., Variational Autoencoders). We evaluate TokaMind on the recently introduced MAST benchmark TokaMark, which comprises 14 tasks with heterogeneous reconstruction and forecasting objectives. Our results show that fine-tuned TokaMind outperforms the strongest benchmark baseline on all but one task. Compared with training the same architecture from scratch under a matched epoch budget, warm-start adaptation is most beneficial on demanding downstream settings, including long-horizon forecasting and high-dimensional equilibrium objectives. These findings highlight the value of multi-modal pretraining for tokamak plasma dynamics and provide a practical, extensible foundation for future fusion modeling tasks. Training code and model weights are publicly available at github.com/UKAEA-IBM-STFC-Fusion-FMs/tokamind and huggingface.co/UKAEA-IBM-STFC, respectively.

Industrial-scale user representation learning requires balancing robust universality with acute task-sensitivity. However, existing paradigms primarily yield static, task-agnostic embeddings that struggle to reconcile the divergent requirements of downstream scenarios within unified vector spaces. Furthermore, heterogeneous multi-source data introduces inherent noise and modality conflicts, degrading representation. We propose Query-as-Anchor, a framework shifting user modeling from static encoding to dynamic, query-aware synthesis. To empower Large Language Models (LLMs) with deep user understanding, we first construct UserU, an industrial-scale pre-training dataset that aligns multi-modal behavioral sequences with user understanding semantics, and our Q-Anchor Embedding architecture integrates hierarchical coarse-to-fine encoders into dual-tower LLMs via joint contrastive-autoregressive optimization for query-aware user representation. To bridge the gap between general pre-training and specialized business logic, we further introduce Cluster-based Soft Prompt Tuning to enforce discriminative latent structures, effectively aligning model attention with scenario-specific modalities. For deployment, anchoring queries at sequence termini enables KV-cache-accelerated inference with negligible incremental latency. Evaluations on 10 Alipay industrial benchmarks show consistent SOTA performance, strong scalability, and efficient deployment. Large-scale online A/B testing in Alipay's production system across two real-world scenarios further validates its practical effectiveness. Our code is prepared for public release and will be available at: https://github.com/JhCircle/Q-Anchor.

We present PrismSSL, a Python library that unifies state-of-the-art self-supervised learning (SSL) methods across audio, vision, graphs, and cross-modal settings in a single, modular codebase. The goal of the demo is to show how researchers and practitioners can: (i) install, configure, and run pretext training with a few lines of code; (ii) reproduce compact benchmarks; and (iii) extend the framework with new modalities or methods through clean trainer and dataset abstractions. PrismSSL is packaged on PyPI, released under the MIT license, integrates tightly with HuggingFace Transformers, and provides quality-of-life features such as distributed training in PyTorch, Optuna-based hyperparameter search, LoRA fine-tuning for Transformer backbones, animated embedding visualizations for sanity checks, Weights & Biases logging, and colorful, structured terminal logs for improved usability and clarity. In addition, PrismSSL offers a graphical dashboard - built with Flask and standard web technologies - that enables users to configure and launch training pipelines with minimal coding. The artifact (code and data recipes) will be publicly available and reproducible.