检索结果-内蒙古大学图书馆

Correction to: Evolutionary optimization of large complex problems

学校读者我要写书评

暂无评论

Complex & Intelligent Systems 2022年第4期8卷 3593-3593页

作者： Wang, Handing Sun, Chaoli Ding, Jinliang Ong, Yew-soon School of Artificial Intelligence Xidian University Xi’an China School of Computer Science and Technology Taiyuan University of Science and Technology Taiyuan China State Key Laboratory of Synthetical Automation for Process Industries Northeastern University Shenyang China School of Computer Science and Engineering Nanyang Technological University Singapore Singapore

Correction to: Real-time object tracking in the wild with Siamese network

学校读者我要写书评

暂无评论

Multimedia Tools and Applications 2023年第16期82卷 24345-24345页

作者： Han, Feng Jiang, Shaokui Wu, Jianmin Xu, Baile Zhao, Jian Shen, Furao State Key Laboratory for Novel Software Technology Nanjing University Nanjing China Department of Computer Science and Technology Nanjing University Nanjing China State Grid Shanghai Maintenance Company Shanghai China School of Electronic Science and Engineering Nanjing University Nanjing China School of Artificial Intelligence Nanjing University Nanjing China

GSCo: Towards Generalizable AI in Medicine via Generalist-Specialist Collaboration

学校读者我要写书评

暂无评论

arXiv 2024年

作者： He, Sunan Nie, Yuxiang Wang, Hongmei Yang, Shu Wang, Yihui Cai, Zhiyuan Chen, Zhixuan Xu, Yingxue Luo, Luyang Xiang, Huiling Lin, Xi Wu, Mingxiang Peng, Yifan Shih, George Xu, Ziyang Wu, Xian Wang, Qiong Chan, Ronald Cheong Kin Vardhanabhuti, Varut Chu, Winnie Chiu Wing Zheng, Yefeng Rajpurkar, Pranav Zhang, Kang Chen, Hao Department of Computer Science and Engineering The Hong Kong University of Science and Technology Hong Kong Department of Biomedical Informatics Harvard University Boston United States Department of Ultrasound Sun Yat-sen University Cancer Center Guangzhou China Department of Radiology Shenzhen People’s Hospital Shenzhen China Population Health Sciences Weill Cornell Medicine New York United States Department of Radiology Weill Cornell Medicine New York United States Perelman Department of Dermatology New York Langone Health New York United States Jarvis Research Center Tencent YouTu Lab Shenzhen China Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China Department of Anatomical and Cellular Pathology The Chinese University of Hong Kong Hong Kong State Key Laboratory of Translational Oncology The Chinese University of Hong Kong Hong Kong Department of Diagnostic Radiology The University of Hong Kong Hong Kong Department of Imaging and Interventional Radiology The Chinese University of Hong Kong Hong Kong Medical Artificial Intelligence Laboratory Westlake University Hangzhou China Faculty of Medicine The Macau University of Science and Technology China Department of Chemical and Biological Engineering The Hong Kong University of Science and Technology Hong Kong Division of Life Science The Hong Kong University of Science and Technology Hong Kong State Key Laboratory of Molecular Neuroscience The Hong Kong University of Science and Technology Hong Kong Shenzhen-Hong Kong Collaborative Innovation Research Institute The Hong Kong University of Science and Technology Shenzhen China

Generalist foundation models (GFMs) are renowned for their exceptional capability and flexibility in effectively generalizing across diverse tasks and modalities. In the field of medicine, while GFMs exhibit superior generalizability based on their extensive intrinsic knowledge as well as proficiency in instruction following and in-context learning, specialist models excel in precision due to their in-depth domain-specific knowledge. In this work, for the first time, we explore the synergy between the GFM and specialist models, to enable precise medical image analysis on a broader scope. Specifically, we propose a novel cooperative framework, Generalist-Specialist Collaboration (GSCo), which consists of two stages, namely the construction of GFM and specialists, and collaborative inference on downstream tasks. In the construction stage, we develop MedDr, the largest open-source GFM tailored for medicine, showcasing exceptional instruction-following and in-context learning capabilities. Meanwhile, a series of lightweight specialists are crafted for specific downstream tasks with low computational overhead. In the collaborative inference stage, we introduce two cooperative mechanisms, Mixture-of-Expert Diagnosis (MoED) and Retrieval-Augmented Diagnosis (RAD), to harvest the generalist’s in-context learning abilities alongside the specialists’ domain expertise. Concretely, MoED incorporates predictions from specialists as references, while RAD employs specialists to retrieve similar cases, collectively providing MedDr with much more contextual guidance. For a comprehensive evaluation, we curate a large-scale benchmark featuring 28 datasets and about 250,000 images across a wide range of medical modalities, including radiology, pathology, dermatology, ophthalmology, and gastroenterology. Extensive experimental results demonstrate that MedDr consistently outperforms state-of-the-art GFMs on downstream datasets. Furthermore, GSCo exceeds both GFMs and specialists across a

关键词： Ophthalmology

RF-Based fall monitoring using convolutional neural networks

学校读者我要写书评

暂无评论

Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 2018年第3期2卷 1-24页

作者： Tian, Yonglong Lee, Guang-He He, Hao Hsu, Chen-Yu Katabi, Dina Computer Science and Artificial Intelligence Laboratory Massachusetts Institute of Technology CambridgeMA02139 United States

Falls are the top reason for fatal and non-fatal injuries among seniors. Existing solutions are based on wearable fall-alert sensors, but medical research has shown that they are ineffective, mostly because seniors do not wear them. These revelations have led to new passive sensors that infer falls by analyzing Radio Frequency (RF) signals in homes. Seniors can go about their lives as usual without the need to wear any device. While passive monitoring has made major advances, current approaches still cannot deal with the complexities of real-world scenarios. They typically train and test their classifiers on the same people in the same environments, and cannot generalize to new people or new environments. Further, they cannot separate motions from different people and can easily miss a fall in the presence of other motions. To overcome these limitations, we introduce Aryokee, an RF-based fall detection system that uses convolutional neural networks governed by a state machine. Aryokee works with new people and environments unseen in the training set. It also separates different sources of motion to increase robustness. Results from testing Aryokee with over 140 people performing 40 types of activities in 57 different environments show a recall of 94% and a precision of 92% in detecting falls. © 2018 Copyright held by the owner/author(s).

关键词： Fall detection

On-chip quantum interference between the origins of a multi-photon state

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Feng, Lan-Tian Zhang, Ming Liu, Di Cheng, Yu-Jie Guo, Guo-Ping Dai, Dao-Xin Guo, Guang-Can Krenn, Mario Ren, Xi-Feng CAS Key Laboratory of Quantum Information University of Science and Technology of China Hefei230026 China CAS Synergetic Innovation Center of Quantum Information & Quantum Physics University of Science and Technology of China Hefei230026 China Hefei National Laboratory University of Science and Technology of China Hefei230088 China State Key Laboratory for Modern Optical Instrumentation Centre for Optical and Electromagnetic Research Zhejiang Provincial Key Laboratory for Sensing Technologies Zhejiang University Zijingang Campus Hangzhou310058 China Erlangen Germany Department of Chemistry & Computer Science University of Toronto Toronto Canada Vector Institute for Artificial Intelligence Toronto Canada

Path identiy induces a broad interest in recent years due to the foundation for numerous novel quantum information applications. Here, we experimentally demonstrate quantum coherent superposition of two different origins of a four-photon state, where multi-photon frustrated interference emerges from the quantum indistinguishability by path identity. The quantum state is created in four probabilistic photon-pair sources on one integrated silicon photonic chip, two combinations of which can create photon quadruplets. Coherent elimination and revival of distributed four-photons are fully controlled by tuning phases. The experiment gives rise to peculiar quantum interference of two possible ways to create photon quadruplets rather than interference of different intrinsic properties of photons. Besides many known potential applications, this new kind of multi-photon nonlinear interference enables the possibility for various fundamental studies such as nonlocality with multiple spatially separated locations. © 2021, CC BY.

关键词： Photons

Lattice quantum chromodynamics at large isospin density

学校读者我要写书评

暂无评论

Physical Review D 2023年第11期108卷 114506-114506页

作者： Ryan Abbott William Detmold Fernando Romero-López Zohreh Davoudi Marc Illa Assumpta Parreño Robert J. Perry Phiala E. Shanahan Michael L. Wagman Center for Theoretical Physics Massachusetts Institute of Technology Cambridge Massachusetts 02139 USA The NSF AI Institute for Artificial Intelligence and Fundamental Interactions Department of Physics and Maryland Center for Fundamental Physics University of Maryland College Park Maryland 20742 USA Joint Center for Quantum Information and Computer Science NIST/University of Maryland College Park Maryland 20742 USA InQubator for Quantum Simulation (IQuS) Department of Physics University of Washington Seattle Washington 98195 USA Departament de Física Quàntica i Astrofísica and Institut de Ciències del Cosmos Universitat de Barcelona Martí i Franquès 1 E08028 Spain Fermi National Accelerator Laboratory Batavia Illinois 60510 USA

We present an algorithm to compute correlation functions for systems with the quantum numbers of many identical mesons from lattice quantum chromodynamics (QCD). The algorithm is numerically stable and allows for the computation of n-pion correlation functions for n∈{1,…,N} using a single N×N matrix decomposition, improving on previous algorithms. We apply the algorithm to calculations of correlation functions with up to 6144 charged pions using two ensembles of gauge field configurations generated with quark masses corresponding to a pion mass mπ=170 MeV and spacetime volumes of (4.43×8.8) fm4 and (5.83×11.6) fm4. We also discuss statistical techniques for the analysis of such systems, in which the correlation functions vary over many orders of magnitude. In particular, we observe that the many-pion correlation functions are well-approximated by log-normal distributions, allowing the extraction of the energies of these systems. Using these energies, the large-isospin-density, zero-baryon-density region of the QCD phase diagram is explored. A peak is observed in the energy density at an isospin chemical potential μI∼1.5mπ, signaling the transition into a Bose-Einstein condensed phase. The isentropic speed of sound, cs, in the medium is seen to exceed the ideal-gas (conformal) limit (cs2≤1/3) over a wide range of chemical potential before falling towards the asymptotic expectation at μI∼15mπ. These, and other thermodynamic observables, indicate that the isospin chemical potential must be large for the system to be well described by an ideal gas or perturbative QCD.

关键词： Lattice QCD QCD phase transitions

Adjacency Constraint for Efficient Hierarchical Reinforcement Learning

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Zhang, Tianren Guo, Shangqi Tan, Tian Hu, Xiaolin Chen, Feng The Department of Automation Tsinghua University Beijing100086 China The Beijing Innovation Center for Future Chip Beijing100086 China The LSBDPA Beijing Key Laboratory Beijing100084 China The Department of Civil and Environmental Engineering Stanford University Stanford CA94305 United States The Department of Computer Science and Technology Institute for Artificial Intelligence Beijing National Research Center for Information Science and Technology State Key Laboratory of Intelligent Technology and Systems Tsinghua University Beijing100084 China

Goal-conditioned Hierarchical Reinforcement Learning (HRL) is a promising approach for scaling up reinforcement learning (RL) techniques. However, it often suffers from training inefficiency as the action space of the high-level, i.e., the goal space, is large. Searching in a large goal space poses difficulty for both high-level subgoal generation and low-level policy learning. In this paper, we show that this problem can be effectively alleviated by restricting the high-level action space from the whole goal space to a k-step adjacent region of the current state using an adjacency constraint. We theoretically prove that in a deterministic Markov Decision Process (MDP), the proposed adjacency constraint preserves the optimal hierarchical policy, while in a stochastic MDP the adjacency constraint induces a bounded state-value suboptimality determined by the MDP’s transition structure. We further show that this constraint can be practically implemented by training an adjacency network that can discriminate between adjacent and non-adjacent subgoals. Experimental results on discrete and continuous control tasks including challenging simulated robot locomotion and manipulation tasks show that incorporating the adjacency constraint significantly boosts the performance of state-of-the-art goal-conditioned HRL approaches. Copyright © 2021, The Authors. All rights reserved.

关键词： Reinforcement learning

Use of Machine Learning for real-time antibiotic treatment adjustment in high-risk patients with CRGNB infection

学校读者我要写书评

暂无评论

computer Methods and Programs in Biomedicine 2025年

作者： Qian Zhang Yejun Wu Lei Zhao Jiao Xie Haotian Mao Chi Guo Xin Zheng Department of Infectious Diseases Union Hospital Tongji Medical College Huazhong University of Science and Technology Wuhan Hubei Province China School of Computer Science Wuhan University Wuhan Hubei Province China Institute of Artificial Intelligence School of Computer Science Wuhan University Wuhan Hubei Province China Health Management Center Union Hospital Tongji Medical College Huazhong University of Science and Technology Wuhan Hubei Province China Research Center of GNSS Wuhan University Wuhan Hubei Province China Joint International Laboratory of Infection and Immunity Huazhong University of Science and Technology Wuhan Hubei Province China Hubei Jiangxia Laboratory Wuhan Hubei Province China

Background : Infections caused by carbapenem resistant gram-negative bacilli (CRGNB) are associated with high mortality and pose a great challenge for clinical treatment. We aim to identify patients at high risk for CRGNB as early as possible and alert clinicians to make timely antibiotic treatment adjustments. Methods : In our study, we included a range of variables for analysis, such as demographic characteristics, past medical history, antibiotic use, and laboratory data. Two machine-learning (ML) models were employed. The baseline prediction model accurately identifies patients with a high likelihood of drug resistance, whereas the real-time ML model facilitates clinical decision-making by alerting practitioners to tailor antibiotic treatment regimens for these high-risk patients. Results : A total of 2139 patients with positive microbiological cultures were included in the study. We found that carbapenem and combination/sequential antibiotic therapy increased the likelihood of CRGNB infection (P < 0.001), and for patients treated with similar antibiotics, the risk increased with exposure time. Six basic models-SVM, DT, RF, GBDT, ANN, and bagging ANN-based were used to developed predictive models and achieved AUC values of 0.87, 0.83, 0.88, 0.89, 0.90 and 0.91. The bagging ANN model has the best clinical prediction performance. Invasive procedures, RBC, ALB, fungal infection and respiratory tract infection were the top 5 most important predictors. Conclusions : The integration of our innovative machine learning models into real-time decision-support tools provides clinicians with valuable insights, facilitating the timely modification of antibiotic treatment regimens for patients at high risk of CRGNB infection.

关键词：

Learning to Detect Unacceptable Machine Translations for Downstream Tasks

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Zhang, Meng Jiang, Xin Liu, Yang Liu, Qun Huawei Noah’s Ark Lab Institute for Artificial Intelligence State Key Laboratory of Intelligent Technology and Systems Department of Computer Science and Technology Tsinghua University Beijing China Beijing National Research Center for Information Science and Technology

The field of machine translation has progressed tremendously in recent years. Even though the translation quality has improved significantly, current systems are still unable to produce uniformly acceptable machine translations for the variety of possible use cases. In this work, we put machine translation in a cross-lingual pipeline and introduce downstream tasks to define task-specific acceptability of machine translations. This allows us to leverage parallel data to automatically generate acceptability annotations on a large scale, which in turn help to learn acceptability detectors for the downstream tasks. We conduct experiments to demonstrate the effectiveness of our framework for a range of downstream tasks and translation models. Copyright © 2020, The Authors. All rights reserved.

关键词： computer aided language translation