检索结果-内蒙古大学图书馆

DLPO : Towards a Robust, Efficient, and Generalizable Prompt Optimization Framework from a Deep-Learning Perspective

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Peng, Dengyun Zhou, Yuhang Chen, Qiguang Liu, Jinhao Chen, Jingjing Qin, Libo Research Center for Social Computing and Interactive Robotics Harbin Institute of Technology China School of Computer Science and Engineering Central South University China Fudan University China

Large Language Models (LLMs) have achieved remarkable success across diverse tasks, largely driven by well-designed prompts. However, crafting and selecting such prompts often requires considerable human effort, significantly limiting its scalability. To mitigate this, recent studies have explored automated prompt optimization as a promising solution. Despite these efforts, existing methods still face critical challenges in robustness, efficiency, and generalization. To systematically address these challenges, we first conduct an empirical analysis to identify the limitations of current reflection-based prompt optimization paradigm. Building on these insights, we propose 7 innovative approaches inspired by traditional deep learning paradigms for prompt optimization (DLPO), seamlessly integrating these concepts into text-based gradient optimization. Through these advancements, we progressively tackle the aforementioned challenges and validate our methods through extensive experimentation. We hope our study not only provides valuable guidance for future research but also offers a comprehensive understanding of the challenges and potential solutions in prompt optimization. Our code is available at https://***/sfasfaffa/DLPO. Copyright © 2025, The Authors. All rights reserved.

关键词： Contrastive Learning

SelfMedHPM: Self Pre-Training with Hard Patches Mining Masked Autoencoders for Medical Image Segmentation

学校读者我要写书评

暂无评论

SelfMedHPM: Self Pre-Training with Hard Patches Mining Maske...

IEEE International Symposium on Biomedical Imaging

作者： Yunhao Lv Lingyu Chen Jian Wang Yangxi Li Fang Chen Department of Computer Science and Engineering Nanjing University of Aeronautics and Astronautics Nanjing China School of Biomedical Engineering and the Institute of Medical Robotics Shanghai JiaoTong University Shanghai China

ISBN: (数字)9798331520526

ISBN: (纸本)9798331520533

In recent years, deep learning methods such as convolutional neural network (CNN) and transformers have made significant progress in CT multi-organ segmentation. However, CT multi-organ segmentation methods based on masked image modeling (MIM) are very limited. There are already methods using MAE for CT multi-organ segmentation task, we believe that the existing methods do not identify the most difficult areas to reconstruct. To this end, we propose a MIM self-training framework with hard patches mining masked autoencoders for CT multi-organ segmentation tasks (self-MedHPM). The method performs ViT self-pretraining on the training set of the target data and introduces an auxiliary loss predictor, which first predicts the patch loss and determines the location of the next mask. SelfMedHPM implementation is better than various competitive methods in abdominal CT multi-organ segmentation and body CT multi-organ segmentation. We have validated the performance of our method on the Multi Atlas Labeling Beyond The Cranial Vault (BTCV) dataset for abdomen mult-organ segmentation and the SinoMed Whole Body (SMWB) dataset for body multi-organ segmentation tasks.

关键词： Training Image segmentation Three-dimensional displays Computed tomography Autoencoders Transformers Convolutional neural networks Prognostics and health management Image reconstruction Biomedical imaging

Iodine-enhanced Liver Vessel Segmentation in Photon Counting Detector-based Computed Tomography using Deep Learning 7

学校读者我要写书评

暂无评论

Iodine-enhanced Liver Vessel Segmentation in Photon Counting...

7th International Conference on Image Formation in X-Ray Computed Tomography

作者： Baek, Sumin Lee, Okkyun Ye, Donghye The Department of Robotics and Mechatronics Engineering Daegu Gyeongbuk Institute of Science and Technology Daegu42988 Korea Republic of The Department of Electrical and Computer Engineering Marquette University MilwaukeeWI53233 United States

ISBN: (纸本)9781510656697

Liver vessel segmentation is important in diagnosing and treating liver diseases. Iodine-based contrast agents are typically used to improve liver vessel segmentation by enhancing vascular structure contrast. However, conventional computed tomography (CT) is still limited with low contrast due to energy-integrating detectors. Photon counting detector-based computed tomography (PCD-CT) shows the high vascular structure contrast in CT images using multi-energy information, thereby allowing accurate liver vessel segmentation. In this paper, we propose a deep learning-based liver vessel segmentation method which takes advantages of the multi-energy information from PCD-CT. We develop a 3D UNet to segment vascular structures within the liver from 4 multi-energy bin images which separates iodine contrast agents. The experimental results on simulated abdominal phantom dataset demonstrated that our proposed method for the PCD-CT outperformed the standard deep learning segmentation method with conventional CT in terms of dice overlap score and 3D vascular structure visualization. © 2022 SPIE.

关键词： Iodine

Research Progress in Cell Classification Detection Based on Deep Learning 14

学校读者我要写书评

暂无评论

Research Progress in Cell Classification Detection Based on ...

14th IEEE International Conference on Cyber Technology in Automation, Control, and Intelligent Systems, CYBER 2024

作者： Gai, Dongze Zhang, Chuang Zhang, Yiwei Guo, Jiaduo Zhang, Qi Qin, Hengshen Dang, Dan School of Computer Science and Engineering Shenyang Jianzhu University Shenyang110168 China Shenyang Institute of Automation Chinese Academy of Sciences State Key Laboratory of Robotics Shenyang110016 China University of Chinese Academy of Sciences Beijing100049 China School of Mechanical Engineering Shenyang Jianzhu University Shenyang110168 China School of Science Shenyang Jianzhu University Shenyang110168 China

ISBN: (纸本)9798331506056

This paper reviews the application of deep learning in cell classification detection, highlights the importance of convolutional neural networks in image analysis, and discusses the application of image preprocessing, feature extraction and deep learning models. Despite the progress made by deep learning techniques in improving cell classification accuracy, challenges such as data quality, model generalization and interpretability remain. This article suggests future research directions, including combining medical knowledge with deep learning and developing multimodal learning to improve the accuracy and efficiency of cell detection. Meanwhile, this article also explores the potential of label-free cell classification methods combining optically induced optically-induced dielectrophoresis (ODEP) techniques with deep learning to achieve cell classification without labeling, providing a new perspective on cell analysis. optically induced. © 2024 IEEE.

关键词： Contrastive Learning

PhysORD: A Neuro-Symbolic Approach for Physics-infused Motion Prediction in Off-road Driving

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhao, Zhipeng Li, Bowen Du, Yi Fu, Taimeng Wang, Chen Lab Institute for Artificial Intelligence and Data Science Department of Computer Science and Engineering University at Buffalo BuffaloNY14260 United States Robotics Institute Carnegie Mellon University PittsburghPA15213 United States

Motion prediction is critical for autonomous off-road driving, however, it presents significantly more challenges than on-road driving because of the complex interaction between the vehicle and the terrain. Traditional physics-based approaches encounter difficulties in accurately modeling dynamic systems and external disturbance. In contrast, data-driven neural networks require extensive datasets and struggle with explicitly capturing the fundamental physical laws, which can easily lead to poor generalization. By merging the advantages of both methods, neuro-symbolic approaches present a promising direction. These methods embed physical laws into neural models, potentially significantly improving generalization capabilities. However, no prior works were evaluated in real-world settings for off-road driving. To bridge this gap, we present PhysORD, a neural-symbolic approach integrating the conservation law, i.e., the Euler-Lagrange equation, into data-driven neural models for motion prediction in off-road driving. Our experiments showed that PhysORD can accurately predict vehicle motion and tolerate external disturbance by modeling uncertainties. The learned dynamics model achieves 46.7% higher accuracy using only 3.1% of the parameters compared to data-driven methods, demonstrating the data efficiency and superior generalization ability of our neural-symbolic method. Copyright © 2024, The Authors. All rights reserved.

关键词： Off road vehicles

A Survey on Multimodal Wearable Sensor-based Human Action Recognition

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Ni, Jianyuan Tang, Hao Haque, Syed Tousiful Yan, Yan Ngu, Anne H.H. The Department of Computer Science Texas State University San MarcosTX78666 United States The Robotics Institute Carnegie Mellon University PittsburghPA15213 United States The Department of Computer Science Illinois Institute of Technology ChicagoIL60616 United States

The combination of increased life expectancy and falling birth rates is resulting in an aging population. Changes associated with aging, can impact an individual's quality of life, potentially leading to injuries, mental health issues, or reduced physical activity. Wearable Sensor-based Human Activity Recognition (WSHAR) emerges as a promising assistive technology to the healthy living of older individuals, unlocking vast potential for human-centric applications. However, recent surveys in WSHAR have been limited, focusing either solely on deep learning approaches or on a single sensor modality. In real life, our human interact with the world in a multi-sensory way, where diverse information sources are intricately processed and interpreted to accomplish a complex and unified sensing system. To give machines similar intelligence, multimodal machine learning, which merges data from various sources, has become a popular research area with recent advancements. In this study, we present a comprehensive survey from a novel perspective on how to leverage multimodal learning to WSHAR domain for newcomers and researchers. We begin by presenting the recent sensor modalities as well as deep learning approaches in HAR. Subsequently, we explore the techniques used in present multimodal systems for WSHAR. This includes inter-multimodal systems which utilize sensor modalities from both visual and non-visual systems and intra-multimodal systems that simply take modalities from non-visual systems. After that, we focus on current multimodal learning approaches that have applied to solve some of the challenges existing in WSHAR. Specifically, we make extra efforts by connecting the existing multimodal literature from other domains, such as computer vision and natural language processing, with current WSHAR area. Finally, we identify the corresponding challenges and potential research direction in current WSHAR area for further improvement. Copyright © 2024, The Authors. All rights re

关键词： Wearable sensors

Complexity based investigation in collaborative assembly scenarios via non intrusive techniques

学校读者我要写书评

暂无评论

Procedia computer science 2023年 217卷 478-485页

作者： Sotirios Panagou Monica Sileo Konstantinos Papoutsakis Fabio Fruggiero Ammar Qammaz Antonis Argyros School of Engineering University of Basilicata 85100 Potenza Italy Computational Vision and Robotics Laboratory Institute of Computer Science Foundation for Research and Technology – Hellas 70013 Heraklion Greece

Human and robot collaboration in assembly tasks is an integral part in modern manufactories. Robots provide advantages in both process and productivity with their repeatability and usability in different tasks, while human operators provide flexibility and can act as safeguards. However, process complexity increases which can lower the overall quality. Increased complexity can negatively influence decision making due to cognitive load on human operators, which can lead to lower quality, be it product, process or human work. Moreover, it can lead to safety risks, human-system error and accidents. In this work, we present the preliminary results on an experiment performed with student-participants, based on an assembly task. The experiment was set up to emulate an industrial assembly, and data collection was performed through qualitative and non-intrusive quantitative methods. Questionnaires were used to assess perceptual task complexity and cognitive load, while a stereo camera provided recordings for after-task analysis on process errors and human work quality based on a 3D skeleton-based human pose estimation and tracking method. The aim of the study is to investigate causes of errors and implications on quality. Future direction of the work is discussed.

关键词： collaboration assembly task complexity cognitive load human pose estimation tracking

MQADet: A Plug-and-Play Paradigm for Enhancing Open-Vocabulary Object Detection via Multimodal Question Answering

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Li, Caixiong Zhao, Xiongwei Zhang, Jinhang Zhang, Xing Sun, Qihao Wu, Zhou School of Computer and Information Science Qinghai Institute of Technology China School of Information Science and Technology Harbin Institute of Technology Shen Zhen China The State Key Laboratory of Robotics and System Harbin Institute of Technology China School of Computer Science and Technology Qinghai University China Eryuan Digital Technology Co. Ltd. China

Open-vocabulary detection (OVD) is a challenging task to detect and classify objects from an unrestricted set of categories, including those unseen during training. Existing open-vocabulary detectors are limited by complex visual-textual misalignment and long-tailed category imbalances, leading to suboptimal performance in challenging scenarios. To address these limitations, we introduce MQADet, a universal paradigm for enhancing existing open-vocabulary detectors by leveraging the cross-modal reasoning capabilities of multimodal large language models (MLLMs). MQADet functions as a plug-and-play solution that integrates seamlessly with pre-trained object detectors without substantial additional training costs. Specifically, we design a novel three-stage Multimodal Question Answering (MQA) pipeline to guide the MLLMs to precisely localize complex textual and visual targets while effectively enhancing the focus of existing object detectors on relevant objects. To validate our approach, we present a new benchmark for evaluating our paradigm on four challenging open-vocabulary datasets, employing three state-of-the-art object detectors as baselines. Experimental results demonstrate that our proposed paradigm significantly improves the performance of existing detectors, particularly in unseen complex categories, across diverse and challenging scenarios. To facilitate future research, we will publicly release our code. © 2025, CC BY.

关键词： Object detection

AI-based Framework for Robust Model-Based Connector Mating in Robotic Wire Harness Installation

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Kienle, Claudius Alt, Benjamin Schneider, Finn Pertlwieser, Tobias Jäkel, Rainer Rayyes, Rania ArtiMinds Robotics Karlsruhe Germany IAS Lab Computer Science Department TU Darmstadt Germany AICOR Institute for Artificial Intelligence University of Bremen Germany Karlsruhe Germany

Despite the widespread adoption of industrial robots in automotive assembly, wire harness installation remains a largely manual process, as it requires precise and flexible manipulation. To address this challenge, we design a novel AI-based framework that automates cable connector mating by integrating force control with deep visuotactile learning. Our system optimizes search-and-insertion strategies using first-order optimization over a multimodal transformer architecture trained on visual, tactile, and proprioceptive data. Additionally, we design a novel automated data collection and optimization pipeline that minimizes the need for machine learning expertise. The framework optimizes robot programs that run natively on standard industrial controllers, permitting human experts to audit and certify them. Experimental validations on a center console assembly task demonstrate significant improvements in cycle times and robustness compared to conventional robot programming approaches. Videos are available under https: //***/*** Codes 68T40 © 2025, CC BY.

关键词： Robot programming