Self-supervised learning shows promise in harnessing extensive unlabeled data, but it also confronts significant privacy concerns, especially in vision. In this paper, we perform membership inference on visual self-supervised models in a more realistic setting: self-supervised training method and details are unknown for an adversary when attacking as he usually faces a black-box system in practice. In this setting, considering that self-supervised model could be trained by completely different self-supervised paradigms, e.g., masked image modeling and contrastive learning, with complex training details, we propose a unified membership inference method called PartCrop. It is motivated by the shared part-aware capability among models and stronger part response on the training data. Specifically, PartCrop crops parts of objects in an image to query responses within the image in representation space. We conduct extensive attacks on self-supervised models with different training protocols and structures using three widely used image datasets. The results verify the effectiveness and generalization of PartCrop. Moreover, to defend against PartCrop, we evaluate two common approaches, i.e., early stop and differential privacy, and propose a tailored method called shrinking crop scale range. The defense experiments indicate that all of them are effective. Finally, besides prototype testing on toy visual encoders and small-scale image datasets, we quantitatively study the impacts of scaling from both data and model aspects in a realistic scenario and propose a scalable PartCrop-v2 by introducing two structural improvements to PartCrop. Our code is at https://github.com/JiePKU/PartCrop.
Accurate and robust localization is a critical enabler for emerging 5G and 6G applications, including autonomous driving, extended reality (XR), and smart manufacturing. While data-driven approaches have shown promise, most existing models require large amounts of labeled data and struggle to generalize across deployment scenarios and wireless configurations. To address these limitations, we propose a foundation-model-based solution tailored for wireless localization. We first analyze how different self-supervised learning (SSL) tasks acquire general-purpose and task-specific semantic features based on information bottleneck (IB) theory. Building on this foundation, we design a pretraining methodology for the proposed Large Wireless Localization Model (LWLM). Specifically, we propose an SSL framework that jointly optimizes three complementary objectives: (i) spatial-frequency masked channel modeling (SF-MCM), (ii) domain-transformation invariance (DTI), and (iii) position-invariant contrastive learning (PICL). These objectives jointly capture the underlying semantics of wireless channel from multiple perspectives. We further design lightweight decoders for key downstream tasks, including time-of-arrival (ToA) estimation, angle-of-arrival (AoA) estimation, single base station (BS) localization, and multiple BS localization. Comprehensive experimental results confirm that LWLM consistently surpasses both model-based and supervised learning baselines across all localization tasks. In particular, LWLM achieves 26.0%--87.5% improvement over transformer models without pretraining, and exhibits strong generalization under label-limited fine-tuning and unseen BS configurations, confirming its potential as a foundation model for wireless localization.
The success of deep learning in computer vision over the past decade has hinged on large labeled datasets and strong pretrained models. In data-scarce settings, the quality of these pretrained models becomes crucial for effective transfer learning. Image classification and self-supervised learning have traditionally been the primary methods for pretraining CNNs and transformer-based architectures. Recently, the rise of text-to-image generative models, particularly those using denoising diffusion in a latent space, has introduced a new class of foundational models trained on massive, captioned image datasets. These models' ability to generate realistic images of unseen content suggests they possess a deep understanding of the visual world. In this work, we present Marigold, a family of conditional generative models and a fine-tuning protocol that extracts the knowledge from pretrained latent diffusion models like Stable Diffusion and adapts them for dense image analysis tasks, including monocular depth estimation, surface normals prediction, and intrinsic decomposition. Marigold requires minimal modification of the pre-trained latent diffusion model's architecture, trains with small synthetic datasets on a single GPU over a few days, and demonstrates state-of-the-art zero-shot generalization. Project page: https://marigoldcomputervision.github.io
Scaling up model and data size have demonstrated impressive performance improvement over a wide range of tasks. Despite extensive studies on scaling behaviors for general-purpose tasks, medical images exhibit substantial differences from natural data. It remains unclear the key factors in developing medical vision foundation models at scale due to the absence of an extensive understanding of scaling behavior in the medical domain. In this paper, we explored the scaling behavior across model sizes, training algorithms, data sizes, and imaging modalities in developing scalable medical vision foundation models by self-supervised learning. To support scalable pretraining, we introduce BioVFM-21M, a large-scale biomedical image dataset encompassing a wide range of biomedical image modalities and anatomies. We observed that scaling up does provide benefits but varies across tasks. Additional analysis reveals several factors correlated with scaling benefits. Finally, we propose BioVFM, a large-scale medical vision foundation model pretrained on 21 million biomedical images, which outperforms the previous state-of-the-art foundation models across 12 medical benchmarks. Our results highlight that while scaling up is beneficial for pursuing better performance, task characteristics, data diversity, pretraining methods, and computational efficiency remain critical considerations for developing scalable medical foundation models.
Self-supervised learning (SSL) models have significantly advanced speech processing tasks, and several benchmarks have been proposed to validate their effectiveness. However, previous benchmarks have primarily focused on single-speaker scenarios, with less exploration of target-speaker tasks in noisy, multi-talker conditions -- a more challenging yet practical case. In this paper, we introduce the Target-Speaker Speech Processing Universal Performance Benchmark (TS-SUPERB), which includes four widely recognized target-speaker processing tasks that require identifying the target speaker and extracting information from the speech mixture. In our benchmark, the speaker embedding extracted from enrollment speech is used as a clue to condition downstream models. The benchmark result reveals the importance of evaluating SSL models in target speaker scenarios, demonstrating that performance cannot be easily inferred from related single-speaker tasks. Moreover, by using a unified SSL-based target speech encoder, consisting of a speaker encoder and an extractor module, we also investigate joint optimization across TS tasks to leverage mutual information and demonstrate its effectiveness.
Intelligent spectrum management is crucial for improving spectrum efficiency and achieving secure utilization of spectrum resources. However, existing intelligent spectrum management methods, typically based on small-scale models, suffer from notable limitations in recognition accuracy, convergence speed, and generalization, particularly in the complex and dynamic spectrum environments. To address these challenges, this paper proposes a novel spectrum foundation model, termed SpectrumFM, establishing a new paradigm for spectrum management. SpectrumFM features an innovative encoder architecture that synergistically exploits the convolutional neural networks and the multi-head self-attention mechanisms to enhance feature extraction and enable robust representation learning. The model is pre-trained via two novel self-supervised learning tasks, namely masked reconstruction and next-slot signal prediction, which leverage large-scale in-phase and quadrature (IQ) data to achieve comprehensive and transferable spectrum representations. Furthermore, a parameter-efficient fine-tuning strategy is proposed to enable SpectrumFM to adapt to various downstream spectrum management tasks, including automatic modulation classification (AMC), wireless technology classification (WTC), spectrum sensing (SS), and anomaly detection (AD). Extensive experiments demonstrate that SpectrumFM achieves superior performance in terms of accuracy, robustness, adaptability, few-shot learning efficiency, and convergence speed, consistently outperforming conventional methods across multiple benchmarks. Specifically, SpectrumFM improves AMC accuracy by up to 12.1% and WTC accuracy by 9.3%, achieves an area under the curve (AUC) of 0.97 in SS at -4 dB signal-to-noise ratio (SNR), and enhances AD performance by over 10%.
Downstream probing has been the dominant method for evaluating model representations, an important process given the increasing prominence of self-supervised learning and foundation models. However, downstream probing primarily assesses the availability of task-relevant information in the model's latent space, overlooking attributes such as equivariance, invariance, and disentanglement, which contribute to the interpretability, adaptability, and utility of representations in real-world applications. While some attempts have been made to measure these qualities in representations, no unified evaluation framework with modular, generalizable, and interpretable metrics exists. In this paper, we argue for the importance of representation evaluation beyond downstream probing. We introduce a standardized protocol to quantify informativeness, equivariance, invariance, and disentanglement of factors of variation in model representations. We use it to evaluate representations from a variety of models in the image and speech domains using different architectures and pretraining approaches on identified controllable factors of variation. We find that representations from models with similar downstream performance can behave substantially differently with regard to these attributes. This hints that the respective mechanisms underlying their downstream performance are functionally different, prompting new research directions to understand and improve representations.
The motion planning problem involves finding a collision-free path from a robot's starting to its target configuration. Recently, self-supervised learning methods have emerged to tackle motion planning problems without requiring expensive expert demonstrations. They solve the Eikonal equation for training neural networks and lead to efficient solutions. However, these methods struggle in complex environments because they fail to maintain key properties of the Eikonal equation, such as optimal value functions and geodesic distances. To overcome these limitations, we propose a novel self-supervised temporal difference metric learning approach that solves the Eikonal equation more accurately and enhances performance in solving complex and unseen planning tasks. Our method enforces Bellman's principle of optimality over finite regions, using temporal difference learning to avoid spurious local minima while incorporating metric learning to preserve the Eikonal equation's essential geodesic properties. We demonstrate that our approach significantly outperforms existing self-supervised learning methods in handling complex environments and generalizing to unseen environments, with robot configurations ranging from 2 to 12 degrees of freedom (DOF).
Self-supervised learning has transformed 2D computer vision by enabling models trained on large, unannotated datasets to provide versatile off-the-shelf features that perform similarly to models trained with labels. However, in 3D scene understanding, self-supervised methods are typically only used as a weight initialization step for task-specific fine-tuning, limiting their utility for general-purpose feature extraction. This paper addresses this shortcoming by proposing a robust evaluation protocol specifically designed to assess the quality of self-supervised features for 3D scene understanding. Our protocol uses multi-resolution feature sampling of hierarchical models to create rich point-level representations that capture the semantic capabilities of the model and, hence, are suitable for evaluation with linear probing and nearest-neighbor methods. Furthermore, we introduce the first self-supervised model that performs similarly to supervised models when only off-the-shelf features are used in a linear probing setup. In particular, our model is trained natively in 3D with a novel self-supervised approach based on a Masked Scene Modeling objective, which reconstructs deep features of masked patches in a bottom-up manner and is specifically tailored to hierarchical 3D models. Our experiments not only demonstrate that our method achieves competitive performance to supervised models, but also surpasses existing self-supervised approaches by a large margin. The model and training code can be found at our Github repository (https://github.com/phermosilla/msm).
Masked video modeling, such as VideoMAE, is an effective paradigm for video self-supervised learning (SSL). However, they are primarily based on reconstructing pixel-level details on natural videos which have substantial temporal redundancy, limiting their capability for semantic representation and sufficient encoding of motion dynamics. To address these issues, this paper introduces a novel SSL approach for video representation learning, dubbed as SMILE, by infusing both spatial and motion semantics. In SMILE, we leverage image-language pretrained models, such as CLIP, to guide the learning process with their high-level spatial semantics. We enhance the representation of motion by introducing synthetic motion patterns in the training data, allowing the model to capture more complex and dynamic content. Furthermore, using SMILE, we establish a new self-supervised video learning paradigm capable of learning strong video representations without requiring any natural video data. We have carried out extensive experiments on 7 datasets with various downstream scenarios. SMILE surpasses current state-of-the-art SSL methods, showcasing its effectiveness in learning more discriminative and generalizable video representations. Code is available: https://github.com/fmthoker/SMILE
Pancreatic cancer, characterized by its notable prevalence and mortality rates, demands accurate lesion delineation for effective diagnosis and therapeutic interventions. The generalizability of extant methods is frequently compromised due to the pronounced variability in imaging and the heterogeneous characteristics of pancreatic lesions, which may mimic normal tissues and exhibit significant inter-patient variability. Thus, we propose a generalization framework that synergizes pixel-level classification and regression tasks, to accurately delineate lesions and improve model stability. This framework not only seeks to align segmentation contours with actual lesions but also uses regression to elucidate spatial relationships between diseased and normal tissues, thereby improving tumor localization and morphological characterization. Enhanced by the reciprocal transformation of task outputs, our approach integrates additional regression supervision within the segmentation context, bolstering the model's generalization ability from a dual-task perspective. Besides, dual self-supervised learning in feature spaces and output spaces augments the model's representational capability and stability across different imaging views. Experiments on 594 samples composed of three datasets with significant imaging differences demonstrate that our generalized pancreas segmentation results comparable to mainstream in-domain validation performance (Dice: 84.07%). More importantly, it successfully improves the results of the highly challenging cross-lesion generalized pancreatic cancer segmentation task by 9.51%. Thus, our model constitutes a resilient and efficient foundational technological support for pancreatic disease management and wider medical applications. The codes will be released at https://github.com/SJTUBME-QianLab/Dual-Task-Seg.
Multiple Instance Learning (MIL) has emerged as the best solution for Whole Slide Image (WSI) classification. It consists of dividing each slide into patches, which are treated as a bag of instances labeled with a global label. MIL includes two main approaches: instance-based and embedding-based. In the former, each patch is classified independently, and then the patch scores are aggregated to predict the bag label. In the latter, bag classification is performed after aggregating patch embeddings. Even if instance-based methods are naturally more interpretable, embedding-based MILs have usually been preferred in the past due to their robustness to poor feature extractors. However, recently, the quality of feature embeddings has drastically increased using self-supervised learning (SSL). Nevertheless, many authors continue to endorse the superiority of embedding-based MIL. To investigate this further, we conduct 710 experiments across 4 datasets, comparing 10 MIL strategies, 6 self-supervised methods with 4 backbones, 4 foundation models, and various pathology-adapted techniques. Furthermore, we introduce 4 instance-based MIL methods never used before in the pathology domain. Through these extensive experiments, we show that with a good SSL feature extractor, simple instance-based MILs, with very few parameters, obtain similar or better performance than complex, state-of-the-art (SOTA) embedding-based MIL methods, setting new SOTA results on the BRACS and Camelyon16 datasets. Since simple instance-based MIL methods are naturally more interpretable and explainable to clinicians, our results suggest that more effort should be put into well-adapted SSL methods for WSI rather than into complex embedding-based MIL methods.
Video quality assessment (VQA) is essential for quantifying perceptual quality in various video processing workflows, spanning from camera capture systems to over-the-top streaming platforms. While recent supervised VQA models have made substantial progress, the reliance on manually annotated datasets -- a process that is labor-intensive, costly, and difficult to scale up -- has hindered further optimization of their generalization to unseen video content and distortions. To bridge this gap, we introduce a self-supervised learning framework for VQA to learn quality assessment capabilities from large-scale, unlabeled web videos. Our approach leverages a \textbf{learning-to-rank} paradigm to train a large multimodal model (LMM) on video pairs automatically labeled via two manners, including quality pseudo-labeling by existing VQA models and relative quality ranking based on synthetic distortion simulations. Furthermore, we introduce a novel \textbf{iterative self-improvement training strategy}, where the trained model acts an improved annotator to iteratively refine the annotation quality of training data. By training on a dataset $10\times$ larger than the existing VQA benchmarks, our model: (1) achieves zero-shot performance on in-domain VQA benchmarks that matches or surpasses supervised models; (2) demonstrates superior out-of-distribution (OOD) generalization across diverse video content and distortions; and (3) sets a new state-of-the-art when fine-tuned on human-labeled datasets. Extensive experimental results validate the effectiveness of our self-supervised approach in training generalized VQA models. The datasets and code will be publicly released to facilitate future research.
Audiovisual emotion recognition (AVER) aims to infer human emotions from nonverbal visual-audio (VA) cues, offering modality-complementary and language-agnostic advantages. However, AVER remains challenging due to the inherent ambiguity of emotional expressions, cross-modal expressive disparities, and the scarcity of reliably annotated data. Recent self-supervised AVER approaches have introduced strong multimodal representations, yet they predominantly rely on modality-specific encoders and coarse content-level alignment, limiting fine-grained emotional semantic modeling. To address these issues, we propose VAEmo, an efficient two-stage framework for emotion-centric joint VA representation learning with external knowledge injection. In Stage 1, a unified and lightweight representation network is pre-trained on large-scale speaker-centric VA corpora via masked reconstruction and contrastive objectives, mitigating the modality gap and learning expressive, complementary representations without emotion labels. In Stage 2, multimodal large language models automatically generate detailed affective descriptions according to our well-designed chain-of-thought prompting for only a small subset of VA samples; these rich textual semantics are then injected by aligning their corresponding embeddings with VA representations through dual-path contrastive learning, further bridging the emotion gap. Extensive experiments on multiple downstream AVER benchmarks show that VAEmo achieves state-of-the-art performance with a compact design, highlighting the benefit of unified cross-modal encoding and emotion-aware semantic guidance for efficient, generalizable VA emotion representations.
Self-supervised learning (SSL) in graphs has garnered significant attention, particularly in employing Graph Neural Networks (GNNs) with pretext tasks initially designed for other domains, such as contrastive learning and feature reconstruction. However, it remains uncertain whether these methods effectively reflect essential graph properties, precisely representation similarity with its neighbors. We observe that existing methods position opposite ends of a spectrum driven by the graph embedding smoothness, with each end corresponding to outperformance on specific downstream tasks. Decomposing the SSL objective into three terms via an information-theoretic framework with a neighbor representation variable reveals that this polarization stems from an imbalance among the terms, which existing methods may not effectively maintain. Further insights suggest that balancing between the extremes can lead to improved performance across a wider range of downstream tasks. A framework, BSG (Balancing Smoothness in Graph SSL), introduces novel loss functions designed to supplement the representation quality in graph-based SSL by balancing the derived three terms: neighbor loss, minimal loss, and divergence loss. We present a theoretical analysis of the effects of these loss functions, highlighting their significance from both the SSL and graph smoothness perspectives. Extensive experiments on multiple real-world datasets across node classification and link prediction consistently demonstrate that BSG achieves state-of-the-art performance, outperforming existing methods. Our implementation code is available at https://github.com/steve30572/BSG.
Graph self-supervised learning has gained significant attention recently. However, many existing approaches heavily depend on perturbations, and inappropriate perturbations may corrupt the graph's inherent information. The Vector Quantized Variational Autoencoder (VQ-VAE) is a powerful autoencoder extensively used in fields such as computer vision; however, its application to graph data remains underexplored. In this paper, we provide an empirical analysis of vector quantization in the context of graph autoencoders, demonstrating its significant enhancement of the model's capacity to capture graph topology. Furthermore, we identify two key challenges associated with vector quantization when applying in graph data: codebook underutilization and codebook space sparsity. For the first challenge, we propose an annealing-based encoding strategy that promotes broad code utilization in the early stages of training, gradually shifting focus toward the most effective codes as training progresses. For the second challenge, we introduce a hierarchical two-layer codebook that captures relationships between embeddings through clustering. The second layer codebook links similar codes, encouraging the model to learn closer embeddings for nodes with similar features and structural topology in the graph. Our proposed model outperforms 16 representative baseline methods in self-supervised link prediction and node classification tasks across multiple datasets.
Recently, learning-based robotic navigation systems have gained extensive research attention and made significant progress. However, the diversity of open-world scenarios poses a major challenge for the generalization of such systems to practical scenarios. Specifically, learned systems for scene measurement and state estimation tend to degrade when the application scenarios deviate from the training data, resulting to unreliable depth and pose estimation. Toward addressing this problem, this work aims to develop a visual odometry system that can fast adapt to diverse novel environments in an online manner. To this end, we construct a self-supervised online adaptation framework for monocular visual odometry aided by an online-updated depth estimation module. Firstly, we design a monocular depth estimation network with lightweight refiner modules, which enables efficient online adaptation. Then, we construct an objective for self-supervised learning of the depth estimation module based on the output of the visual odometry system and the contextual semantic information of the scene. Specifically, a sparse depth densification module and a dynamic consistency enhancement module are proposed to leverage camera poses and contextual semantics to generate pseudo-depths and valid masks for the online adaptation. Finally, we demonstrate the robustness and generalization capability of the proposed method in comparison with state-of-the-art learning-based approaches on urban, in-house datasets and a robot platform. Code is publicly available at: https://github.com/jixingwu/SOL-SLAM.
Current self-supervised methods, such as contrastive learning, predominantly focus on global discrimination, neglecting the critical fine-grained anatomical details required for accurate radiographic analysis. To address this challenge, we propose an Anatomy-driven self-supervised framework for enhancing Fine-grained Representation in radiographic image analysis (AFiRe). The core idea of AFiRe is to align the anatomical consistency with the unique token-processing characteristics of Vision Transformer. Specifically, AFiRe synergistically performs two self-supervised schemes: (i) Token-wise anatomy-guided contrastive learning, which aligns image tokens based on structural and categorical consistency, thereby enhancing fine-grained spatial-anatomical discrimination; (ii) Pixel-level anomaly-removal restoration, which particularly focuses on local anomalies, thereby refining the learned discrimination with detailed geometrical information. Additionally, we propose Synthetic Lesion Mask to enhance anatomical diversity while preserving intra-consistency, which is typically corrupted by traditional data augmentations, such as Cropping and Affine transformations. Experimental results show that AFiRe: (i) provides robust anatomical discrimination, achieving more cohesive feature clusters compared to state-of-the-art contrastive learning methods; (ii) demonstrates superior generalization, surpassing 7 radiography-specific self-supervised methods in multi-label classification tasks with limited labeling; and (iii) integrates fine-grained information, enabling precise anomaly detection using only image-level annotations.
Heart diseases constitute the main cause of international human defunction. According to the World Health Organization (WHO), approximately 18 million deaths happen each year due to precisely heart diseases. In particular, heart failures (HF) press the healthcare industry to develop systems for their early, rapid, and effective prediction. This work presents an automatic system based on a novel deep learning framework which analyses in real-time echocardiography video sequences for the challenging and more specific task of heart failure time prediction. This system works in two stages. The first one transforms the data from a database of echocardiography video sequences into a machine learning-compatible collection of annotated images which can be used in the training phase of any machine learning-based framework, including a deep learning-based one. This stage includes the use of the Higher Order Dynamic Mode Decomposition (HODMD) algorithm for both data augmentation and feature extraction. The second stage builds and trains a Vision Transformer (ViT). Self-supervised learning (SSL) methods, so far barely explored in the literature about heart failure prediction, are adopted to effectively train the ViT from scratch, even with scarce databases. The designed neural network analyses images from echocardiography sequences to estimate the time in which a heart failure will happen. The results obtained show the efficacy of the HODMD algorithm and the superiority of the proposed system with respect to several established ViT and Convolutional Neural Network (CNN) architectures. The source code will be incorporated into the next version release of the ModelFLOWs-app software (https://github.com/modelflows/ModelFLOWs-app).
The rapid advancement of audio generation technologies has escalated the risks of malicious deepfake audio across speech, sound, singing voice, and music, threatening multimedia security and trust. While existing countermeasures (CMs) perform well in single-type audio deepfake detection (ADD), their performance declines in cross-type scenarios. This paper is dedicated to studying the alltype ADD task. We are the first to comprehensively establish an all-type ADD benchmark to evaluate current CMs, incorporating cross-type deepfake detection across speech, sound, singing voice, and music. Then, we introduce the prompt tuning self-supervised learning (PT-SSL) training paradigm, which optimizes SSL frontend by learning specialized prompt tokens for ADD, requiring 458x fewer trainable parameters than fine-tuning (FT). Considering the auditory perception of different audio types,we propose the wavelet prompt tuning (WPT)-SSL method to capture type-invariant auditory deepfake information from the frequency domain without requiring additional training parameters, thereby enhancing performance over FT in the all-type ADD task. To achieve an universally CM, we utilize all types of deepfake audio for co-training. Experimental results demonstrate that WPT-XLSR-AASIST achieved the best performance, with an average EER of 3.58% across all evaluation sets. The code is available online.