Recent advancements in Multimodal Large Language Models (LLMs) have focused primarily on scaling by increasing text-image pair data and enhancing LLMs to improve performance on multimodal tasks. However, these scaling approaches are computationally expensive and overlook the significance of efficiently improving model capabilities from the vision side. Inspired by the successful applications of Mixture-of-Experts (MoE) in LLMs, which improves model scalability during training while keeping inference costs similar to those of smaller models, we propose CuMo, which incorporates Co-upcycled Top-K sparsely-gated Mixture-of-experts blocks into both the vision encoder and the MLP connector, thereby enhancing the multimodal LLMs with neglectable additional activated parameters during inference. CuMo first pre-trains the MLP blocks and then initializes each expert in the MoE block from the pre-trained MLP block during the visual instruction tuning stage, with auxiliary losses to ensure a balanced loading of experts. CuMo outperforms state-of-the-art multimodal LLMs across various VQA and visual-instruction-following benchmarks within each model size group, all while training exclusively on open-sourced datasets.

A great deal of progress has been made in image captioning, driven by research into how to encode the image using pre-trained models. This includes visual encodings (e.g. image grid features or detected objects) and more recently textual encodings (e.g. image tags or text descriptions of image regions). As more advanced encodings are available and incorporated, it is natural to ask: how to efficiently and effectively leverage the heterogeneous set of encodings? In this paper, we propose to regard the encodings as augmented views of the input image. The image captioning model encodes each view independently with a shared encoder efficiently, and a contrastive loss is incorporated across the encoded views in a novel way to improve their representation quality and the model's data efficiency. Our proposed hierarchical decoder then adaptively weighs the encoded views according to their effectiveness for caption generation by first aggregating within each view at the token level, and then across views at the view level. We demonstrate significant performance improvements of +5.6% CIDEr on MS-COCO and + 12.9% CIDEr on Flickr30k compared to state of the arts, and conduct rigorous analyses to demonstrate the importance of each part of our design.

In Vision-and-Language Navigation (VLN), researchers typically take an image encoder pre-trained on ImageNet without fine-tuning on the environments that the agent will be trained or tested on. However, the distribution shift between the training images from ImageNet and the views in the navigation environments may render the ImageNet pre-trained image encoder suboptimal. Therefore, in this paper, we design a set of structure-encoding auxiliary tasks (SEA) that leverage the data in the navigation environments to pre-train and improve the image encoder. Specifically, we design and customize (1) 3D jigsaw, (2) traversability prediction, and (3) instance classification to pre-train the image encoder. Through rigorous ablations, our SEA pre-trained features are shown to better encode structural information of the scenes, which ImageNet pre-trained features fail to properly encode but is crucial for the target navigation task. The SEA pre-trained features can be easily plugged into existing VLN agents without any tuning. For example, on Test-Unseen environments, the VLN agents combined with our SEA pre-trained features achieve absolute success rate improvement of 12% for Speaker-Follower [14], 5% for Env-Dropout [37], and 4% for AuxRN [50].

Significant progress has been made on visual captioning, largely relying on pre-trained features and later fixed object detectors that serve as rich inputs to auto-regressive models. A key limitation of such methods, however, is that the output of the model is conditioned only on the object detector's outputs. The assumption that such outputs can represent all necessary information is unrealistic, especially when the detector is transferred across datasets. In this work, we reason about the graphical model induced by this assumption, and propose to add an auxiliary input to represent missing information such as object relationships. We specifically propose to mine attributes and relationships from the Visual Genome dataset and condition the captioning model on them. Crucially, we propose (and show to be important) the use of a multi-modal pre-trained model (CLIP) to retrieve such contextual descriptions. Further, object detector models are frozen and do not have sufficient richness to allow the captioning model to properly ground them. As a result, we propose to condition both the detector and description outputs on the image, and show qualitatively and quantitatively that this can improve grounding. We validate our method on image captioning, perform thorough analyses of each component and importance of the pre-trained multi-modal model, and demonstrate significant improvements over the current state of the art, specifically +7.5% in CIDEr and +1.3% in BLEU-4 metrics.

In this paper, we propose the problem of collaborative perception, where robots can combine their local observations with those of neighboring agents in a learnable way to improve accuracy on a perception task. Unlike existing work in robotics and multi-agent reinforcement learning, we formulate the problem as one where learned information must be shared across a set of agents in a bandwidth-sensitive manner to optimize for scene understanding tasks such as semantic segmentation. Inspired by networking communication protocols, we propose a multi-stage handshake communication mechanism where the neural network can learn to compress relevant information needed for each stage. Specifically, a target agent with degraded sensor data sends a compressed request, the other agents respond with matching scores, and the target agent determines who to connect with (i.e., receive information from). We additionally develop the AirSim-CP dataset and metrics based on the AirSim simulator where a group of aerial robots perceive diverse landscapes, such as roads, grasslands, buildings, etc. We show that for the semantic segmentation task, our handshake communication method significantly improves accuracy by approximately 20% over decentralized baselines, and is comparable to centralized ones using a quarter of the bandwidth.

3D reconstruction, inferring 3D shape information from a single 2D image, has drawn attention from learning and vision communities. In this paper, we propose a framework for learning pose-aware 3D shape reconstruction. Our proposed model learns deep representation for recovering the 3D object, with the ability to extract camera pose information but without any direct supervision of ground truth camera pose. This is realized by exploitation of 2D-3D self-consistency between 2D masks and 3D voxels. Experiments qualitatively and quantitatively demonstrate the effectiveness and robustness of our model, which performs favorably against state-of-the-art methods.

This paper presents a challenging computer vision task, namely the detection of generic components on a PCB, and a novel set of deep-learning methods that are able to jointly leverage the appearance of individual components and the propagation of information across the structure of the board to accurately detect and identify various types of components on a PCB. Due to the expense of manual data labeling, a highly unbalanced distribution of component types, and significant domain shift across boards, most earlier attempts based on traditional image processing techniques fail to generalize well to PCB images with various quality, lighting conditions, etc. Newer object detection pipelines such as Faster R-CNN, on the other hand, require a large amount of labeled data, do not deal with domain shift, and do not leverage structure. To address these issues, we propose a three stage pipeline in which a class-agnostic region proposal network is followed by a low-shot similarity prediction classifier. In order to exploit the data dependency within a PCB, we design a novel Graph Network block to refine the component features conditioned on each PCB. To the best of our knowledge, this is one of the earliest attempts to train a deep learning based model for such tasks, and we demonstrate improvements over recent graph networks for this task. We also provide in-depth analysis and discussion for this challenging task, pointing to future research.

In this paper, a service-oriented multiagent system (SoMAS) for the control and analysis of the cyber–physical system (CPS) in manufacturing automation utilizing a noncontact dynamic obstacle avoidance seven-DoF robot arm is presented. The interfaces of the services which the robot arm subsystem should provide to fully exploit its capability are identified. Specifically, the services of moving, object recognition, object fetching, and safety of human–robot interaction are considered as the fundamental functionalities that the robot arm should provide. The way to evaluate the quality of services (QoS) for the robot arm subsystem is also explained. To build such a robot arm subsystem, the system architecture is proposed. Also, implementation for the subsystem which includes: 3-D model-based object recognition, grasp database for object fetching, and online noncontact obstacle avoidance for the safety of human–robot interaction is provided. The experimental results demonstrate that the capability of 3-D model-based object recognition, object fetching, and dynamic collision avoidance are successfully implemented.

The objective of this paper is to present the model-based 3D object recognition and fetching by a 7-DoF robot with online obstacle avoidance for factory automation. The robot can fetch the random type of 3D objects with arbitrary pose using a 3D visual camera. Object recognition pipeline based on different descriptors are compared in order to decide the best choice for this scenario. A correct grasp is taught by human operator using programming by touch. As a result, the time of configuring and setting up the robot drastically decreases and thus saves a large amount of preparation time before the robot can actually get to work. Considering the safety issue of the human operators that cooperate with the robot, as well as the uncertainty of the environment such that a moving obstacle might present, or the environment might change for some reason, an online obstacle avoidance algorithm is also integrated into the system. Experimental proof of principles has been successfully implemented.

In this paper, we propose a modularized architecture for a robot arm object fetching system integrated with 3D CAD-model based object recognition system that can cope with objects in random types and poses. The interface and the core functionality of each module in the architecture are discussed in detail. Implementation of each module is also conducted. Furthermore, the assumptions and the working conditions behind each module are carefully examined. We develop the system based on our previous work and enhance the recognition module. To proof the feasibility of our architecture, 3D object recognition and fetching demonstration are successfully implemented, and the result of object recognition, teaching by touching, and fast grasp synthesis are successfully demonstrated.