Prior studies on 3D scene understanding have primarily developed specialized models for specific tasks or required task-specific fine-tuning. In this study, we propose Grounded 3D-LLM, which explores the potential of 3D large multi-modal models (3D LMMs) to consolidate various 3D vision tasks within a unified generative framework. The model uses scene referent tokens as special noun phrases to reference 3D scenes, enabling the handling of sequences that interleave 3D and textual data. It offers a natural approach for translating 3D vision tasks into language formats using task-specific instruction templates. To facilitate the use of referent tokens in subsequent language modeling, we have curated large-scale grounded language datasets that offer finer scene-text correspondence at the phrase level by bootstrapping existing object labels. Subsequently, we introduced Contrastive LAnguage-Scene Pre-training (CLASP) to effectively leverage this data, thereby integrating 3D vision with language models. Our comprehensive evaluation covers open-ended tasks like dense captioning and 3D QA, alongside close-ended tasks such as object detection and language grounding. Experiments across multiple 3D benchmarks reveal the leading performance and the broad applicability of Grounded 3D-LLM. Code and datasets will be released on the project page: //groundedscenellm.github.io/grounded_3d-llm.github.io.
相關內容
This paper introduces a new multi-model predictive control (MMPC) method for quadrotor attitude control with performance nearly on par with nonlinear model predictive control (NMPC) and computational efficiency similar to linear model predictive control (LMPC). Conventional NMPC, while effective, is computationally intensive, especially for attitude control that needs a high refresh rate. Conversely, LMPC offers computational advantages but suffers from poor performance and local stability. Our approach relies on multiple linear models of attitude dynamics, each accompanied by a linear model predictive controller, dynamically switching between them given flight conditions. We leverage gap metric analysis to minimize the number of models required to accurately predict the vehicle behavior in various conditions and incorporate a soft switching mechanism to ensure system stability during controller transitions. Our results show that with just 15 models, the vehicle attitude can be accurately controlled across various set points. Comparative evaluations with existing controllers such as incremental nonlinear dynamic inversion, sliding mode control, LMPC, and NMPC reveal that our approach closely matches the effectiveness of NMPC, outperforming other methods, with a running time comparable to LMPC.
We present Direct Reward Fine-Tuning (DRaFT), a simple and effective method for fine-tuning diffusion models to maximize differentiable reward functions, such as scores from human preference models. We first show that it is possible to backpropagate the reward function gradient through the full sampling procedure, and that doing so achieves strong performance on a variety of rewards, outperforming reinforcement learning-based approaches. We then propose more efficient variants of DRaFT: DRaFT-K, which truncates backpropagation to only the last K steps of sampling, and DRaFT-LV, which obtains lower-variance gradient estimates for the case when K=1. We show that our methods work well for a variety of reward functions and can be used to substantially improve the aesthetic quality of images generated by Stable Diffusion 1.4. Finally, we draw connections between our approach and prior work, providing a unifying perspective on the design space of gradient-based fine-tuning algorithms.
High-resolution electron microscopy (HREM) imaging technique is a powerful tool for directly visualizing a broad range of materials in real-space. However, it faces challenges in denoising due to ultra-low signal-to-noise ratio (SNR) and scarce data availability. In this work, we propose Noise2SR, a zero-shot self-supervised learning (ZS-SSL) denoising framework for HREM. Within our framework, we propose a super-resolution (SR) based self-supervised training strategy, incorporating the Random Sub-sampler module. The Random Sub-sampler is designed to generate approximate infinite noisy pairs from a single noisy image, serving as an effective data augmentation in zero-shot denoising. Noise2SR trains the network with paired noisy images of different resolutions, which is conducted via SR strategy. The SR-based training facilitates the network adopting more pixels for supervision, and the random sub-sampling helps compel the network to learn continuous signals enhancing the robustness. Meanwhile, we mitigate the uncertainty caused by random-sampling by adopting minimum mean squared error (MMSE) estimation for the denoised results. With the distinctive integration of training strategy and proposed designs, Noise2SR can achieve superior denoising performance using a single noisy HREM image. We evaluate the performance of Noise2SR in both simulated and real HREM denoising tasks. It outperforms state-of-the-art ZS-SSL methods and achieves comparable denoising performance with supervised methods. The success of Noise2SR suggests its potential for improving the SNR of images in material imaging domains.
Affine frequency division multiplexing (AFDM) is a promising new multicarrier technique based on discrete affine Fourier transform (DAFT). By properly tuning pre-chirp parameter and post-chirp parameter in the DAFT, the effective channel in the DAFT domain can completely avoid overlap of different paths, thus constitutes a full representation of delay-Doppler profile, which significantly improves the system performance in high mobility scenarios. However, AFDM has the crucial problem of high peak-to-average power ratio (PAPR) caused by phase randomness of modulated symbols. In this letter, an algorithm named grouped pre-chirp selection (GPS) is proposed to reduce the PAPR by changing the value of pre-chirp parameter on sub-carriers group by group. Specifically, it is demonstrated first that the important properties of AFDM system are maintained when implementing GPS. Secondly, we elaborate the operation steps of GPS algorithm, illustrating its effect on PAPR reduction and its advantage in terms of computational complexity compared with the ungrouped approach. Finally, simulation results of PAPR reduction in the form of complementary cumulative distribution function (CCDF) show the effectiveness of the proposed GPS algorithm.
The ability of CodeLLMs to generate executable and functionally correct code at the repository-level scale remains largely unexplored. We introduce RepoExec, a novel benchmark for evaluating code generation at the repository-level scale. RepoExec focuses on three main aspects: executability, functional correctness through automated test case generation with high coverage rate, and carefully crafted cross-file contexts to accurately generate code. Our work explores a controlled scenario where developers specify necessary code dependencies, challenging the model to integrate these accurately. Experiments show that while pretrained LLMs outperform instruction-tuned models in correctness, the latter excel in utilizing provided dependencies and demonstrating debugging capabilities. We also introduce a new instruction-tuned dataset that focuses on code dependencies and demonstrate that CodeLLMs fine-tuned on our dataset have a better capability to leverage these dependencies effectively. RepoExec aims to provide a comprehensive evaluation of code functionality and alignment with developer intent, paving the way for more reliable and applicable CodeLLMs in real-world scenarios. The dataset and source code can be found at~\url{//github.com/FSoft-AI4Code/RepoExec}.
Self-supervised speech (SSL) models have recently become widely adopted for many downstream speech processing tasks. The general usage pattern is to employ SSL models as feature extractors, and then train a downstream prediction head to solve a specific task. However, different layers of SSL models have been shown to capture different types of information, and the methods of combining them are not well studied. To this end, we extend the general framework for SSL model utilization by proposing the interface that connects the upstream and downstream. Under this view, the dominant technique of combining features via a layerwise weighted sum can be regarded as a specific interface. We propose several alternative interface designs and demonstrate that the weighted sum interface is suboptimal for many tasks. In particular, we show that a convolutional interface whose depth scales logarithmically with the depth of the upstream model consistently outperforms many other interface designs.
We study the visual semantic embedding problem for image-text matching. Most existing work utilizes a tailored cross-attention mechanism to perform local alignment across the two image and text modalities. This is computationally expensive, even though it is more powerful than the unimodal dual-encoder approach. This work introduces a dual-encoder image-text matching model, leveraging a scene graph to represent captions with nodes for objects and attributes interconnected by relational edges. Utilizing a graph attention network, our model efficiently encodes object-attribute and object-object semantic relations, resulting in a robust and fast-performing system. Representing caption as a scene graph offers the ability to utilize the strong relational inductive bias of graph neural networks to learn object-attribute and object-object relations effectively. To train the model, we propose losses that align the image and caption both at the holistic level (image-caption) and the local level (image-object entity), which we show is key to the success of the model. Our model is termed Composition model for Object Relations and Attributes, CORA. Experimental results on two prominent image-text retrieval benchmarks, Flickr30K and MSCOCO, demonstrate that CORA outperforms existing state-of-the-art computationally expensive cross-attention methods regarding recall score while achieving fast computation speed of the dual encoder.
Fine-grained domain generalization (FGDG) is a more challenging task than traditional DG tasks due to its small inter-class variations and relatively large intra-class disparities. When domain distribution changes, the vulnerability of subtle features leads to a severe deterioration in model performance. Nevertheless, humans inherently demonstrate the capacity for generalizing to out-of-distribution data, leveraging structured multi-granularity knowledge that emerges from discerning the commonality and specificity within categories. Likewise, we propose a Feature Structuralized Domain Generalization (FSDG) model, wherein features experience structuralization into common, specific, and confounding segments, harmoniously aligned with their relevant semantic concepts, to elevate performance in FGDG. Specifically, feature structuralization (FS) is accomplished through joint optimization of five constraints: a decorrelation function applied to disentangled segments, three constraints ensuring common feature consistency and specific feature distinctiveness, and a prediction calibration term. By imposing these stipulations, FSDG is prompted to disentangle and align features based on multi-granularity knowledge, facilitating robust subtle distinctions among categories. Extensive experimentation on three benchmarks consistently validates the superiority of FSDG over state-of-the-art counterparts, with an average improvement of 6.2% in FGDG performance. Beyond that, the explainability analysis on explicit concept matching intensity between the shared concepts among categories and the model channels, along with experiments on various mainstream model architectures, substantiates the validity of FS.
This paper introduces the Visual Inverse Kinematics problem (VIK) to fill the gap between robot Inverse Kinematics (IK) and visual servo control. Different from the IK problem, the VIK problem seeks to find robot configurations subject to vision-based constraints, in addition to kinematic constraints. In this work, we develop a formulation of the VIK problem with a Field of View (FoV) constraint, enforcing the visibility of an object from a camera on the robot. Our proposed solution is based on the idea of adding a virtual kinematic chain connecting the physical robot and the object; the FoV constraint is then equivalent to a joint angle kinematic constraint. Along the way, we introduce multiple vision-based cost functions to fulfill different objectives. We solve this formulation of the VIK problem using a method that involves a semidefinite program (SDP) constraint followed by a rank minimization algorithm. The performance of this method for solving the VIK problem is validated through simulations.
We propose a novel single shot object detection network named Detection with Enriched Semantics (DES). Our motivation is to enrich the semantics of object detection features within a typical deep detector, by a semantic segmentation branch and a global activation module. The segmentation branch is supervised by weak segmentation ground-truth, i.e., no extra annotation is required. In conjunction with that, we employ a global activation module which learns relationship between channels and object classes in a self-supervised manner. Comprehensive experimental results on both PASCAL VOC and MS COCO detection datasets demonstrate the effectiveness of the proposed method. In particular, with a VGG16 based DES, we achieve an mAP of 81.7 on VOC2007 test and an mAP of 32.8 on COCO test-dev with an inference speed of 31.5 milliseconds per image on a Titan Xp GPU. With a lower resolution version, we achieve an mAP of 79.7 on VOC2007 with an inference speed of 13.0 milliseconds per image.
小貼士
登錄享
相關主題
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191