We address the problem of sufficient dimension reduction for feature matrices, which arises often in sensor network localization, brain neuroimaging, and electroencephalography analysis. In general, feature matrices have both row- and column-wise interpretations and contain structural information that can be lost with naive vectorization approaches. To address this, we propose a method called principal support matrix machine (PSMM) for the matrix sufficient dimension reduction. The PSMM converts the sufficient dimension reduction problem into a series of classification problems by dividing the response variables into slices. It effectively utilizes the matrix structure by finding hyperplanes with rank-1 normal matrix that optimally separate the sliced responses. Additionally, we extend our approach to the higher-order tensor case. Our numerical analysis demonstrates that the PSMM outperforms existing methods and has strong interpretability in real data applications.
Multi-task learning has attracted much attention due to growing multi-purpose research with multiple related data sources. Moreover, transduction with matrix completion is a useful method in multi-label learning. In this paper, we propose a transductive matrix completion algorithm that incorporates a calibration constraint for the features under the multi-task learning framework. The proposed algorithm recovers the incomplete feature matrix and target matrix simultaneously. Fortunately, the calibration information improves the completion results. In particular, we provide a statistical guarantee for the proposed algorithm, and the theoretical improvement induced by calibration information is also studied. Moreover, the proposed algorithm enjoys a sub-linear convergence rate. Several synthetic data experiments are conducted, which show the proposed algorithm out-performs other existing methods, especially when the target matrix is associated with the feature matrix in a nonlinear way.
Multi-modality image fusion aims to combine different modalities to produce fused images that retain the complementary features of each modality, such as functional highlights and texture details. To leverage strong generative priors and address challenges such as unstable training and lack of interpretability for GAN-based generative methods, we propose a novel fusion algorithm based on the denoising diffusion probabilistic model (DDPM). The fusion task is formulated as a conditional generation problem under the DDPM sampling framework, which is further divided into an unconditional generation subproblem and a maximum likelihood subproblem. The latter is modeled in a hierarchical Bayesian manner with latent variables and inferred by the expectation-maximization algorithm. By integrating the inference solution into the diffusion sampling iteration, our method can generate high-quality fused images with natural image generative priors and cross-modality information from source images. Note that all we required is an unconditional pre-trained generative model, and no fine-tuning is needed. Our extensive experiments indicate that our approach yields promising fusion results in infrared-visible image fusion and medical image fusion. The code will be released.
The field of natural language processing (NLP) has made significant strides in recent years, particularly in the development of large-scale vision-language models (VLMs). These models aim to bridge the gap between text and visual information, enabling a more comprehensive understanding of multimedia data. However, as these models become larger and more complex, they also become more challenging to train and deploy. One approach to addressing this challenge is the use of sparsely-gated mixture-of-experts (MoE) techniques, which divide the model into smaller, specialized sub-models that can jointly solve a task. In this paper, we explore the effectiveness of MoE in scaling vision-language models, demonstrating its potential to achieve state-of-the-art performance on a range of benchmarks over dense models of equivalent computational cost. Our research offers valuable insights into stabilizing the training of MoE models, understanding the impact of MoE on model interpretability, and balancing the trade-offs between compute performance when scaling VLMs. We hope our work will inspire further research into the use of MoE for scaling large-scale vision-language models and other multimodal machine learning applications.
We study sequential probability assignment in the Gaussian setting, where the goal is to predict, or equivalently compress, a sequence of real-valued observations almost as well as the best Gaussian distribution with mean constrained to a given subset of $\mathbf{R}^n$. First, in the case of a convex constraint set $K$, we express the hardness of the prediction problem (the minimax regret) in terms of the intrinsic volumes of $K$; specifically, it equals the logarithm of the Wills functional from convex geometry. We then establish a comparison inequality for the Wills functional in the general nonconvex case, which underlines the metric nature of this quantity and generalizes the Slepian-Sudakov-Fernique comparison principle for the Gaussian width. Motivated by this inequality, we characterize the exact order of magnitude of the considered functional for a general nonconvex set, in terms of global covering numbers and local Gaussian widths. This implies metric isomorphic estimates for the log-Laplace transform of the intrinsic volume sequence of a convex body. As part of our analysis, we also characterize the minimax redundancy for a general constraint set. We finally relate and contrast our findings with classical asymptotic results in information theory.
With the recent development of autonomous driving technology, as the pursuit of efficiency for repetitive tasks and the value of non-face-to-face services increase, mobile service robots such as delivery robots and serving robots attract attention, and their demands are increasing day by day. However, when something goes wrong, most commercial serving robots need to return to their starting position and orientation to operate normally again. In this paper, we focus on end-to-end relocalization of serving robots to address the problem. It is to predict robot pose directly from only the onboard sensor data using neural networks. In particular, we propose a deep neural network architecture for the relocalization based on camera-2D LiDAR sensor fusion. We call the proposed method FusionLoc. In the proposed method, the multi-head self-attention complements different types of information captured by the two sensors. Our experiments on a dataset collected by a commercial serving robot demonstrate that FusionLoc can provide better performances than previous relocalization methods taking only a single image or a 2D LiDAR point cloud as well as a straightforward fusion method concatenating their features.
This paper explores minimum sensing navigation of robots in environments cluttered with obstacles. The general objective is to find a path plan to a goal region that requires minimal sensing effort. In [1], the information-geometric RRT* (IG-RRT*) algorithm was proposed to efficiently find such a path. However, like any stochastic sampling-based planner, the computational complexity of IG-RRT* grows quickly, impeding its use with a large number of nodes. To remedy this limitation, we suggest running IG-RRT* with a moderate number of nodes, and then using a smoothing algorithm to adjust the path obtained. To develop a smoothing algorithm, we explicitly formulate the minimum sensing path planning problem as an optimization problem. For this formulation, we introduce a new safety constraint to impose a bound on the probability of collision with obstacles in continuous-time, in contrast to the common discrete-time approach. The problem is amenable to solution via the convex-concave procedure (CCP). We develop a CCP algorithm for the formulated optimization and use this algorithm for path smoothing. We demonstrate the efficacy of the proposed approach through numerical simulations.
In today's society, where independent living is becoming increasingly important, it can be extremely constricting for those who are blind. Blind and visually impaired (BVI) people face challenges because they need manual support to prompt information about their environment. In this work, we took our first step towards developing an affordable and high-performing eye wearable assistive device, DRISHTI, to provide visual navigation assistance for BVI people. This system comprises a camera module, ESP32 processor, Bluetooth module, smartphone and speakers. Using artificial intelligence, this system is proposed to detect and understand the nature of the users' path and obstacles ahead of the user in that path and then inform BVI users about it via audio output to enable them to acquire directions by themselves on their journey. This first step discussed in this paper involves establishing a proof-of-concept of achieving the right balance of affordability and performance by testing an initial software integration of a currency detection algorithm on a low-cost embedded arrangement. This work will lay the foundation for our upcoming works toward achieving the goal of assisting the maximum of BVI people around the globe in moving independently.
Positron Emission Tomography (PET) and Computer Tomography (CT) are routinely used together to detect tumors. PET/CT segmentation models can automate tumor delineation, however, current multimodal models do not fully exploit the complementary information in each modality, as they either concatenate PET and CT data or fuse them at the decision level. To combat this, we propose Mirror U-Net, which replaces traditional fusion methods with multimodal fission by factorizing the multimodal representation into modality-specific branches and an auxiliary multimodal decoder. At these branches, Mirror U-Net assigns a task tailored to each modality to reinforce unimodal features while preserving multimodal features in the shared representation. In contrast to previous methods that use either fission or multi-task learning, Mirror U-Net combines both paradigms in a unified framework. We explore various task combinations and examine which parameters to share in the model. We evaluate Mirror U-Net on the AutoPET PET/CT and on the multimodal MSD BrainTumor datasets, demonstrating its effectiveness in multimodal segmentation and achieving state-of-the-art performance on both datasets. Our code will be made publicly available.
Accurately segmenting fluid in 3D volumetric optical coherence tomography (OCT) images is a crucial yet challenging task for detecting eye diseases. Traditional autoencoding-based segmentation approaches have limitations in extracting fluid regions due to successive resolution loss in the encoding phase and the inability to recover lost information in the decoding phase. Although current transformer-based models for medical image segmentation addresses this limitation, they are not designed to be applied out-of-the-box for 3D OCT volumes, which have a wide-ranging channel-axis size based on different vendor device and extraction technique. To address these issues, we propose SwinVFTR, a new transformer-based architecture designed for precise fluid segmentation in 3D volumetric OCT images. We first utilize a channel-wise volumetric sampling for training on OCT volumes with varying depths (B-scans). Next, the model uses a novel shifted window transformer block in the encoder to achieve better localization and segmentation of fluid regions. Additionally, we propose a new volumetric attention block for spatial and depth-wise attention, which improves upon traditional residual skip connections. Consequently, utilizing multi-class dice loss, the proposed architecture outperforms other existing architectures on the three publicly available vendor-specific OCT datasets, namely Spectralis, Cirrus, and Topcon, with mean dice scores of 0.72, 0.59, and 0.68, respectively. Additionally, SwinVFTR outperforms other architectures in two additional relevant metrics, mean intersection-over-union (Mean-IOU) and structural similarity measure (SSIM).