Search

Large-scale vision models have become integral in many applications due to their unprecedented performance and versatility across downstream tasks. However, the robustness of these foundation models has primarily been explored for a single task, namely image classification. The vulnerability of other common vision tasks, such as semantic segmentation and depth estimation, remains largely unknown. We present a comprehensive empirical evaluation of the adversarial robustness of self-supervised vision encoders across multiple downstream tasks. Our attacks operate in the encoder embedding space and at the downstream task output level. In both cases, current state-of-the-art adversarial fine-tuning techniques tested only for classification significantly degrade clean and robust performance on other tasks. Since the purpose of a foundation model is to cater to multiple applications at once, our findings reveal the need to enhance encoder robustness more broadly. Our code is available at ${github.com/layer6ai-labs/ssl-robustness}$., Comment: Accepted at the ICML 2024 Workshop on Foundation Models in the Wild

Contrastive learning is a model pre-training technique by first creating similar views of the original data, and then encouraging the data and its corresponding views to be close in the embedding space. Contrastive learning has witnessed success in image and natural language data, thanks to the domain-specific augmentation techniques that are both intuitive and effective. Nonetheless, in tabular domain, the predominant augmentation technique for creating views is through corrupting tabular entries via swapping values, which is not as sound or effective. We propose a simple yet powerful improvement to this augmentation technique: corrupting tabular data conditioned on class identity. Specifically, when corrupting a specific tabular entry from an anchor row, instead of randomly sampling a value in the same feature column from the entire table uniformly, we only sample from rows that are identified to be within the same class as the anchor row. We assume the semi-supervised learning setting, and adopt the pseudo labeling technique for obtaining class identities over all table rows. We also explore the novel idea of selecting features to be corrupted based on feature correlation structures. Extensive experiments show that the proposed approach consistently outperforms the conventional corruption method for tabular data classification tasks. Our code is available at https://github.com/willtop/Tabular-Class-Conditioned-SSL., Comment: 14 pages, 4 algorithms, 3 figures, 5 tables

In response to everyday queries, humans explicitly signal uncertainty and offer alternative answers when they are unsure. Machine learning models that output calibrated prediction sets through conformal prediction mimic this human behaviour; larger sets signal greater uncertainty while providing alternatives. In this work, we study the usefulness of conformal prediction sets as an aid for human decision making by conducting a pre-registered randomized controlled trial with conformal prediction sets provided to human subjects. With statistical significance, we find that when humans are given conformal prediction sets their accuracy on tasks improves compared to fixed-size prediction sets with the same coverage guarantee. The results show that quantifying model uncertainty with conformal prediction is helpful for human-in-the-loop decision making and human-AI teams., Comment: Published at ICML 2024. Code available at https://github.com/layer6ai-labs/hitl-conformal-prediction

360{\deg} spherical images have advantages of wide view field, and are typically projected on a planar plane for processing, which is known as equirectangular image. The object shape in equirectangular images can be distorted and lack translation invariance. In addition, there are few publicly dataset of equirectangular images with labels, which presents a challenge for standard CNNs models to process equirectangular images effectively. To tackle this problem, we propose a methodology for converting a perspective image into equirectangular image. The inverse transformation of the spherical center projection and the equidistant cylindrical projection are employed. This enables the standard CNNs to learn the distortion features at different positions in the equirectangular image and thereby gain the ability to semantically the equirectangular image. The parameter, {\phi}, which determines the projection position of the perspective image, has been analyzed using various datasets and models, such as UNet, UNet++, SegNet, PSPNet, and DeepLab v3+. The experiments demonstrate that an optimal value of {\phi} for effective semantic segmentation of equirectangular images is 6{\pi}/16 for standard CNNs. Compared with the other three types of methods (supervised learning, unsupervised learning and data augmentation), the method proposed in this paper has the best average IoU value of 43.76%. This value is 23.85%, 10.7% and 17.23% higher than those of other three methods, respectively.

Self-supervised representation learning~(SSRL) has advanced considerably by exploiting the transformation invariance assumption under artificially designed data augmentations. While augmentation-based SSRL algorithms push the boundaries of performance in computer vision and natural language processing, they are often not directly applicable to other data modalities, and can conflict with application-specific data augmentation constraints. This paper presents an SSRL approach that can be applied to any data modality and network architecture because it does not rely on augmentations or masking. Specifically, we show that high-quality data representations can be learned by reconstructing random data projections. We evaluate the proposed approach on a wide range of representation learning tasks that span diverse modalities and real-world applications. We show that it outperforms multiple state-of-the-art SSRL baselines. Due to its wide applicability and strong empirical results, we argue that learning from randomness is a fruitful research direction worthy of attention and further study., Comment: Published as a conference paper of ICLR 2024. https://openreview.net/pdf?id=EpYnZpDpsQ

We systematically study a wide variety of generative models spanning semantically-diverse image datasets to understand and improve the feature extractors and metrics used to evaluate them. Using best practices in psychophysics, we measure human perception of image realism for generated samples by conducting the largest experiment evaluating generative models to date, and find that no existing metric strongly correlates with human evaluations. Comparing to 17 modern metrics for evaluating the overall performance, fidelity, diversity, rarity, and memorization of generative models, we find that the state-of-the-art perceptual realism of diffusion models as judged by humans is not reflected in commonly reported metrics such as FID. This discrepancy is not explained by diversity in generated samples, though one cause is over-reliance on Inception-V3. We address these flaws through a study of alternative self-supervised feature extractors, find that the semantic information encoded by individual networks strongly depends on their training procedure, and show that DINOv2-ViT-L/14 allows for much richer evaluation of generative models. Next, we investigate data memorization, and find that generative models do memorize training examples on simple, smaller datasets like CIFAR10, but not necessarily on more complex datasets like ImageNet. However, our experiments show that current metrics do not properly detect memorization: none in the literature is able to separate memorization from other phenomena such as underfitting or mode shrinkage. To facilitate further development of generative models and their evaluation we release all generated image datasets, human evaluation data, and a modular library to compute 17 common metrics for 9 different encoders at https://github.com/layer6ai-labs/dgm-eval., Comment: NeurIPS 2023. 53 pages, 29 figures, 12 tables. Code at https://github.com/layer6ai-labs/dgm-eval, reviews at https://openreview.net/forum?id=08zf7kTOoh

In the traditional federated learning setting, a central server coordinates a network of clients to train one global model. However, the global model may serve many clients poorly due to data heterogeneity. Moreover, there may not exist a trusted central party that can coordinate the clients to ensure that each of them can benefit from others. To address these concerns, we present a novel decentralized framework, FedeRiCo, where each client can learn as much or as little from other clients as is optimal for its local data distribution. Based on expectation-maximization, FedeRiCo estimates the utilities of other participants' models on each client's data so that everyone can select the right collaborators for learning. As a result, our algorithm outperforms other federated, personalized, and/or decentralized approaches on several benchmark datasets, being the only approach that consistently performs better than training with local data only.

Spatiotemporal prediction of event data is a challenging task with a long history of research. While recent work in spatiotemporal prediction has leveraged deep sequential models that substantially improve over classical approaches, these models are prone to overfitting when the observation is extremely sparse, as in the task of crime event prediction. To overcome these sparsity issues, we present Multi-axis Attentive Prediction for Sparse Event Data (MAPSED). We propose a purely attentional approach to extract both short-term dynamics and long-term semantics of event propagation through two observation angles. Unlike existing temporal prediction models that propagate latent information primarily along the temporal dimension, the MAPSED simultaneously operates over all axes (time, 2D space, event type) of the embedded data tensor. We additionally introduce a novel Frobenius norm-based contrastive learning objective to improve latent representational generalization.Empirically, we validate MAPSED on two publicly accessible urban crime datasets for spatiotemporal sparse event prediction, where MAPSED outperforms both classical and state-of-the-art deep learning models. The proposed contrastive learning objective significantly enhances the MAPSED's ability to capture the semantics and dynamics of the events, resulting in better generalization ability to combat sparse observations.