Leveraging vast training data, multimodal large language models (MLLMs) have demonstrated formidable general visual comprehension capabilities and achieved remarkable performance across various tasks. However, their performance in visual document understanding still leaves much room for improvement. This discrepancy is primarily attributed to the fact that visual document understanding is a fine-grained prediction task. In natural scenes, MLLMs typically use low-resolution images, leading to a substantial loss of visual information. Furthermore, general-purpose MLLMs do not excel in handling document-oriented instructions. In this paper, we propose a High-Resolution Visual Document Assistant (HRVDA), which bridges the gap between MLLMs and visual document understanding. This model employs a content filtering mechanism and an instruction filtering module to separately filter out the content-agnostic visual tokens and instruction-agnostic visual tokens, thereby achieving efficient model training and inference for high-resolution images. In addition, we construct a document-oriented visual instruction tuning dataset and apply a multi-stage training strategy to enhance the model's document modeling capabilities. Extensive experiments demonstrate that our model achieves state-of-the-art performance across multiple document understanding datasets, while maintaining training efficiency and inference speed comparable to low-resolution models.
We present the LM Transparency Tool (LM-TT), an open-source interactive toolkit for analyzing the internal workings of Transformer-based language models. Differently from previously existing tools that focus on isolated parts of the decision-making process, our framework is designed to make the entire prediction process transparent, and allows tracing back model behavior from the top-layer representation to very fine-grained parts of the model. Specifically, it (1) shows the important part of the whole input-to-output information flow, (2) allows attributing any changes done by a model block to individual attention heads and feed-forward neurons, (3) allows interpreting the functions of those heads or neurons. A crucial part of this pipeline is showing the importance of specific model components at each step. As a result, we are able to look at the roles of model components only in cases where they are important for a prediction. Since knowing which components should be inspected is key for analyzing large models where the number of these components is extremely high, we believe our tool will greatly support the interpretability community both in research settings and in practical applications.
This study focuses on a multi-user massive multiple-input multiple-output (MU-mMIMO) system by incorporating an unmanned aerial vehicle (UAV) as a decode-and-forward (DF) relay between the base station (BS) and multiple Internet-of-Things (IoT) devices. Our primary objective is to maximize the overall achievable rate (AR) by introducing a novel framework that integrates joint hybrid beamforming (HBF) and UAV localization in dynamic MU-mMIMO IoT systems. Particularly, HBF stages for BS and UAV are designed by leveraging slow time-varying angular information, whereas a deep reinforcement learning (RL) algorithm, namely deep deterministic policy gradient (DDPG) with continuous action space, is developed to train the UAV for its deployment. By using a customized reward function, the RL agent learns an optimal UAV deployment policy capable of adapting to both static and dynamic environments. The illustrative results show that the proposed DDPG-based UAV deployment (DDPG-UD) can achieve approximately 99.5% of the sum-rate capacity achieved by particle swarm optimization (PSO)-based UAV deployment (PSO-UD), while requiring a significantly reduced runtime at approximately 68.50% of that needed by PSO-UD, offering an efficient solution in dynamic MU-mMIMO environments.
Recent advances in Automated Theorem Proving have shown the effectiveness of leveraging a (large) language model that generates tactics (i.e. proof steps) to search through proof states. The current model, while trained solely on successful proof paths, faces a discrepancy at the inference stage, as it must sample and try various tactics at each proof state until finding success, unlike its training which does not incorporate learning from failed attempts. Intuitively, a tactic that leads to a failed search path would indicate that similar tactics should receive less attention during the following trials. In this paper, we demonstrate the benefit of training models that additionally learn from failed search paths. Facing the lack of such trial-and-error data in existing open-source theorem-proving datasets, we curate a dataset on intuitionistic propositional logic theorems and formalize it in Lean, such that we can reliably check the correctness of proofs. We compare our model trained on relatively short trial-and-error information (TrialMaster) with models trained only on the correct paths and discover that the former solves more unseen theorems with lower trial searches.
Hierarchical Task Network (HTN) planning usually requires a domain engineer to provide manual input about how to decompose a planning problem. Even HTN-MAKER, a well-known method-learning algorithm, requires a domain engineer to annotate the tasks with information about what to learn. We introduce CURRICULAMA, an HTN method learning algorithm that completely automates the learning process. It uses landmark analysis to compose annotated tasks and leverages curriculum learning to order the learning of methods from simpler to more complex. This eliminates the need for manual input, resolving a core issue with HTN-MAKER. We prove CURRICULAMA's soundness, and show experimentally that it has a substantially similar convergence rate in learning a complete set of methods to HTN-MAKER.
Recently, visually-situated text parsing (VsTP) has experienced notable advancements, driven by the increasing demand for automated document understanding and the emergence of Generative Large Language Models (LLMs) capable of processing document-based questions. Various methods have been proposed to address the challenging problem of VsTP. However, due to the diversified targets and heterogeneous schemas, previous works usually design task-specific architectures and objectives for individual tasks, which inadvertently leads to modal isolation and complex workflow. In this paper, we propose a unified paradigm for parsing visually-situated text across diverse scenarios. Specifically, we devise a universal model, called OmniParser, which can simultaneously handle three typical visually-situated text parsing tasks: text spotting, key information extraction, and table recognition. In OmniParser, all tasks share the unified encoder-decoder architecture, the unified objective: point-conditioned text generation, and the unified input & output representation: prompt & structured sequences. Extensive experiments demonstrate that the proposed OmniParser achieves state-of-the-art (SOTA) or highly competitive performances on 7 datasets for the three visually-situated text parsing tasks, despite its unified, concise design. The code is available at https://github.com/AlibabaResearch/AdvancedLiterateMachinery.
This paper proposes a semi-supervised methodology for training physics-informed machine learning methods. This includes self-training of physics-informed neural networks and physics-informed Gaussian processes in isolation, and the integration of the two via co-training. We demonstrate via extensive numerical experiments how these methods can ameliorate the issue of propagating information forward in time, which is a common failure mode of physics-informed machine learning.
By deploying a large number of antennas with sub-half-wavelength spacing in a compact space, dense array systems(DASs) can fully unleash the multiplexing-and-diversity gains of limited apertures. To acquire these gains, accurate channel state information acquisition is necessary but challenging due to the large antenna numbers. To overcome this obstacle, this paper reveals that exploiting the high spatial correlation of DAS channels is crucial while designing the observation matrix for optimal/near-optimal channel estimation. Firstly, we prove that the observation matrix design is equivalent to a time-domain duality of multiple-input multiple-output precoding, which can be ideally addressed by the water-filling principle. For practical realizations, a novel ice-filling algorithm is proposed to design amplitude-and-phase controllable observation matrices, and a majorization-minimization algorithm is proposed to address the phase-only controllable case. Particularly, we prove that the ice-filling algorithm can be viewed as a ``quantized" water-filling algorithm. To support the sub-optimality of the proposed designs, we provide comprehensive analyses on the achievable mean square errors and their asymptotic expressions. Finally, numerical simulations verify that our proposed channel estimation designs can achieve the near-optimal performance and outperform existing approaches significantly.
Multispectral photoacoustic tomography (PAT) is an imaging modality that utilizes the photoacoustic effect to achieve non-invasive and high-contrast imaging of internal tissues. However, the hardware cost and computational demand of a multispectral PAT system consisting of up to thousands of detectors are huge. To address this challenge, we propose an ultra-sparse spiral sampling strategy for multispectral PAT, which we named U3S-PAT. Our strategy employs a sparse ring-shaped transducer that, when switching excitation wavelengths, simultaneously rotates and translates. This creates a spiral scanning pattern with multispectral angle-interlaced sampling. To solve the highly ill-conditioned image reconstruction problem, we propose a self-supervised learning method that is able to introduce structural information shared during spiral scanning. We simulate the proposed U3S-PAT method on a commercial PAT system and conduct in vivo animal experiments to verify its performance. The results show that even with a sparse sampling rate as low as 1/30, our U3S-PAT strategy achieves similar reconstruction and spectral unmixing accuracy as non-spiral dense sampling. Given its ability to dramatically reduce the time required for three-dimensional multispectral scanning, our U3S-PAT strategy has the potential to perform volumetric molecular imaging of dynamic biological activities.
This study aims to establish a computer-aided diagnostic system for lung lesions using bronchoscope endobronchial ultrasound (EBUS) to assist physicians in identifying lesion areas. During EBUS-transbronchial needle aspiration (EBUS-TBNA) procedures, physicians rely on grayscale ultrasound images to determine the location of lesions. However, these images often contain significant noise and can be influenced by surrounding tissues or blood vessels, making interpretation challenging. Previous research has lacked the application of object detection models to EBUS-TBNA, and there has been no well-defined solution for annotating the EBUS-TBNA dataset. In related studies on ultrasound images, although models have been successful in capturing target regions for their respective tasks, their training and predictions have been based on two-dimensional images, limiting their ability to leverage temporal features for improved predictions. This study introduces a three-dimensional image-based object detection model. It utilizes an attention mechanism to capture temporal correlations and we will implements a filtering mechanism to select relevant information from previous frames. Subsequently, a teacher-student model training approach is employed to optimize the model further, leveraging unlabeled data. To mitigate the impact of poor-quality pseudo-labels on the student model, we will add a special Gaussian Mixture Model (GMM) to ensure the quality of pseudo-labels.