27,206 results on '"Zhao Jian"'
Search Results
2. The effects of self-management intervention based on Information-Motivation-Behavioral model in patients receiving targeted therapy for lung cancer (基于IMB模型的自我管理干预在肺癌靶向治疗患者中的应用)
- Author
-
ZHANG Junli (张俊丽), ZHAO Jian (赵健), JING Dan (井丹), AN Juan (安娟), and HAN Lijun (韩丽军)
- Subjects
lung cancer ,information-motivation-behavioral model ,self-management ,quality of life ,oncology nursing ,肺癌 ,信息-动机-行为技巧模型 ,自我管理 ,生活质量 ,肿瘤护理 ,Nursing ,RT1-120 - Abstract
Objective To investigate the effect of self-management intervention based on Information-Motivation-Behavioral (IMB) model in the patients with advanced lung cancer receiving targeted therapy. Methods A total of 102 patients with advanced lung cancer receiving targeted therapy in a tertiary grade A hospital were divided into study group (n=51) and control group (n=51) according to the sequence of hospital admission. The patients in the control group were provided with routine intervention, while patients in the study group were treated with the self-management interventions based on the IMB model. The disease knowledge level, self-management skills and quality of life were evaluated before intervention and one month after discharge. Results The scores of disease knowledge questionnaire, Strategies Used by People to Promote Health (SUPPH)and Generic Quality of Life Inventory-74 (GQOLI-74) were decreased in both two groups after intervention, and scores of relevant assessment in the study group were higher than those in the control group(P<0. 05). Conclusion The self-management interventions based on IMB model can effectively improve lung cancer patients' self- management ability and quality of life during targeted therapy. (目的 分析基于信息-动机-行为技巧(IMB)模型的自我管理干预在肺癌靶向治疗患者中的应用效果。方法 选取2021年3月—2022年3月医院住院的肺癌靶向治疗患者102例, 按患者住院时间分为对照组和干预组, 其中2021年3月—9月收治的51例患者为对照组, 2021年10月—2022年3月收治的51例患者为干预组。对照组患者接受常规护理干预, 干预组患者在对照组基础上实施基于 IMB 模型的自我管理干预, 两组患者干预时间均为1个月。分别于入组时、干预后1个月时评估两组患者疾病知识水平、自我管理能力和生活质量。结果 干预后, 两组疾病知识调查问卷、癌症自我管理效能感量表(SUPPH)和生活质量综合评定问卷(GQOLI-74)得分较干预前上升, 且干预组疾病知识调查问卷、SUPPH量表和GQOLI-74问卷得分高于对照组, 差异有统计学意义(P<0. 05)。结论 基于IMB模型的自我管理干预可改善肺癌靶向治疗患者自我管理水平, 提升患者生活质量。)
- Published
- 2024
- Full Text
- View/download PDF
3. Application of balloon guide catheter in emergency endovascular thrombectomy for large vessel occlusion
- Author
-
LUO Lei⁃lei, ZHAO Jian⁃bin, and WEI Ming
- Subjects
ischemic stroke ,thrombectomy ,balloon guide catheter (not in mesh) ,review ,Neurology. Diseases of the nervous system ,RC346-429 - Abstract
Endovascular thrombectomy is a first ⁃ line treatment option for acute large vessel occlusive ischemic stroke after strict selection, which is an effective method to improve the good prognosis rate of patients. More and more evidence⁃based medicine proves that balloon guide catheter (BGC) can shorten the reperfusion time, increase the rate of initial reperfusion, and reduce the risk of distal embolism due to it's proximal blood flow occlusion, but the application of BGC in emergency endovascular thrombectomy is still limited. This review summarizes the clinical application progress of BGC to provide theoretical basis for it's application in emergency endovascular thrombectomy.
- Published
- 2024
- Full Text
- View/download PDF
4. Identification of Antemortem and Postmortem Injuries in Nude Mice Based on Microbial Communities
- Author
-
ZHENG Xin, QIU Yue, LI Zhi-gang, XIANG Qing-qing, WANG Guan-san, SHI He, XU Qu-yi, SUI Peng, MA Yan-bing, LIU Chao, CHEN Li-fang, and ZHAO Jian
- Subjects
forensic pathology ,microbiology ,colony ,antemortem injury ,postmortem injury ,16s rrna ,balb/c nude mice ,Medicine - Abstract
ObjectiveTo establish antemortem and postmortem injury models in nude mice, observe the morphological changes of the wounds and the changes of the microbial communities in the wounds at different time points after the injury, and analyze the differences between antemortem and postmortem wounds.MethodsAbdominal injury models were established in 48 BALB/c nude mice, which were classified into antemortem injury, 4 h and 72 h postmortem injury groups, and the gross manifestations and histopathological changes were observed on days 1, 3, 5, 8, 11 and 15 after injury. The microbial communities in the wounds were analyzed by 16S rRNA sequencing technology. QIIME 2 software was used to calculate Shannon and Observed species indices. The Kruskal-Wallis test was used to determine statistical differences in α-diversity between groups. Jaccard similarity coefficients were calculated by using R v4.3.0 software and applied to the principal co-ordinates analysis to demonstrate inter-sample differences. Permutational multivariate analysis of variance (PERMANOVA) was used to analyze the differences between groups in the composition of bacterial colonies, and R2 values were calculated.ResultsOn days 8, 11 and 15 after injury, the antemortem and postmortem injuries could not be differentiated by morphological examination; the Shannon index and Observed species index were statistically different between the antemortem injury group and the 72 h postmortem injury group; the Jaccard similarity coefficient of the microbial community was statistically different between the antemortem injury group and the 72 h postmortem injury group. The PERMANOVA R2 value gradually increased with the extension of time (0.22-0.61).ConclusionThrough the analysis of the wound microbial community, the microbial composition of wounds at different time points can be identified and compared, which provides a new perspective and method for the differentiation of antemortem injuries from postmortem injuries, with good application prospects.
- Published
- 2024
- Full Text
- View/download PDF
5. Changes of serum high mobility group box 1 and soluble triggering receptor expressed on myeloid cells-1 in patients with multiple injuries and their prognostic significance
- Author
-
WANG Guijie, DU Chuanchong, LU Ye, ZHAO Jian, SHEN Xie, JIN Donglin, and GENG Jiacai
- Subjects
multiple injury ,high mobility group box 1 (hmgb1) ,soluble triggering receptor expressed on myeloid cell-1 (strem-1) ,prognosis ,Medicine - Abstract
Objective·To detect the serum levels of high mobility group box 1 (HMGB1) and soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) in patients with multiple injuries at different time points, and to analyze their correlation with disease severity, complications and prognosis.Methods·Ninety-two patients with multiple injuries admitted to the Department of Emergency Medicine of the Suzhou Ninth People′s Hospital from December 2020 to December 2022 were selected. According to the injury severity scores of the patients at admission, the patients were divided into light injury group (n=24), grave injury group (n=58) and severe injury group (n=10). According to whether there was multiple organ dysfunction syndrome (MODS) after admission, the patients were divided into MODS group (n=20) and non-MODS group (n=72). According to the outcome within 28 d after trauma, the patients were divided into death group (n=13) and survival group (n=79). Inflammatory factor indicators in venous blood of patients after admission were detected. Enzyme linked immunosorbent assay (ELISA) was used to detect the serum HMGB1 and sTREM-1 levels at 24 h, 72 h and 7 d after trauma, and the differences of serum HMGB1 and sTREM-1 levels among different groups were analyzed. Multiple Logistic regression was used to analyze the influencing factors of adverse outcomes in patients with multiple injuries. The receiver operating characteristic (ROC) curve was used to evaluate the predictive value of HMGB1 and sTREM-1 for adverse outcomes.Results·The levels of HMGB1 and sTREM-1 in the grave injury and severe injury groups were significantly higher than those in the light injury group (P
- Published
- 2024
- Full Text
- View/download PDF
6. Two way workable microchanneled hydrogel suture to diagnose, treat and monitor the infarcted heart
- Author
-
Fangchao Xue, Shanlan Zhao, Hao Tian, Haoxiang Qin, Xiaochen Li, Zhao Jian, Jiahui Du, Yanzhao Li, Yanhong Wang, Lin Lin, Chen Liu, Yongning Shang, Lang He, Malcolm Xing, and Wen Zeng
- Subjects
Science - Abstract
Abstract During myocardial infarction, microcirculation disturbance in the ischemic area can cause necrosis and formation of fibrotic tissue, potentially leading to malignant arrhythmia and myocardial remodeling. Here, we report a microchanneled hydrogel suture for two-way signal communication, pumping drugs on demand, and cardiac repair. After myocardial infarction, our hydrogel suture monitors abnormal electrocardiogram through the mobile device and triggers nitric oxide on demand via the hydrogel sutures’ microchannels, thereby inhibiting inflammation, promoting microvascular remodeling, and improving the left ventricular ejection fraction in rats and minipigs by more than 60% and 50%, respectively. This work proposes a suture for bidirectional communication that acts as a cardio-patch to repair myocardial infarction, that remotely monitors the heart, and can deliver drugs on demand.
- Published
- 2024
- Full Text
- View/download PDF
7. Prognostic Nutritional Index Enhances the Discriminatory Ability of Procalcitonin for Predicting Pediatric Sepsis
- Author
-
Deng Hongya MBBS, Deng Linfan MBBS, He Chunyuan MBBS, Jiang Jun MD, Liu Bin MD, Zhao Jian MMed, and Li Gang MD
- Subjects
Pediatrics ,RJ1-570 - Abstract
Objective. Improving diagnostic ability of pediatric sepsis is of great significance for reducing the mortality of sepsis. This study explored the discriminatory capacity of nutritional index (PNI) in pediatric sepsis. Methods. We retrospectively enrolled 134 children with suspected sepsis and collected their clinical and laboratory data. Receiver operating characteristic curves (ROC), decision curve analysis (DCA) and net reclassification improvement (NRI) were performed to compare the predictive significance of the PNI, procalcitonin (PCT) and their combination. Results. Among 134 patients, 65 children were diagnosed with sepsis and 69 children with non-sepsis. PCT and PNI were independently associated with pediatric sepsis. PCT was superior to PNI to predict pediatric sepsis. The model based on PCT + PNI improved the predictive capacity than them alone, as demonstrated by ROC, DCA and NRI, respectively. Conclusion. PNI was independently associated with pediatric sepsis, and addition of PNI could improve the capacity of PCT to predict pediatric sepsis.
- Published
- 2024
- Full Text
- View/download PDF
8. Research Progress on the Effect of Small Molecular Sugars on Starch Properties
- Author
-
SONG Jin-hong, CHEN Long, ZHAO Jian-wei, and JI Hang-yan
- Subjects
small molecular sugar ,starch ,gelatinization properties ,freeze-thaw stability ,aging properties ,Food processing and manufacture ,TP368-456 ,Nutrition. Foods and food supply ,TX341-641 - Abstract
Small molecular sugars are a group of sugars with low molecular weight, mainly including monosaccharides, disaccharides, and other oligosaccharides with low degree of polymerization. Small molecular sugars are commonly used as food additives in the food industy, including providing sweetness, improving food texture and optimizeing food processing adaptability. In addition, some small molecular sugars also have unique physiological functions that can be used to regulate the nutritional quality of food. Starch is the most common staple component in food, and starch and its modified derivatives are also widely used as food ingredients in food categories such as sauces, dairy products, beverages, and meat products. The composite system of small molecular sugar and starch is widely used in food. The addition of small molecular sugar has an important impact on the properties of natural starch, and plays an important role in regulating the quality of starch-based foods. This article mainly introduces the common small molecular sugars in food and their effects on the main physical and chemical properties of starch (gel texture, transparency, coagulation, gelatinization, freeze-thaw stability, aging, digestion, etc.), with emphasis on the impact of different types and amounts of small molecular sugars on the properties of starch from different sources. This paper aimed to provide reference for the application of small molecular sugars in starch based foods.
- Published
- 2023
- Full Text
- View/download PDF
9. Evaluation of color and stability of ethyl-linked anthocyanin-flavanol pigments in model wine solutions using combined chemical analysis and 3D molecular simulations
- Author
-
Zhao Jian, Guo Min, Wang Ruoyao, Li Lingxi, and Sun Baoshan
- Subjects
red wine ,ethyl-linked anthocyanin-flavanol pigments ,anthocyanins ,color ,stability ,Plant culture ,SB1-1110 - Abstract
Ethyl-linked anthocyanin-flavanol pigments are one of the most important condensation products formed during the red winemaking and ageing period. They have great contribution to the color characteristics and stability of aged red wines. In this study, the color characteristics and stability of ethyl-linked anthocyanin-flavanol pigments and their precursor anthocyanins were evaluated by combined spectrophotometry and 3D molecular simulations. In model wine solutions, the condensation reactions between three anthocyanins and (-)-epicatechin, mediated by acetaldehyde, were conducted to produce ethyl-linked anthocyanin-flavanol pigments. The color was assessed by the CIELab method, and the concentration changes were analyzed by HPLC-DAD. On the other hand, the stability of these pigmented compounds was also calculated by the three 3D molecular simulation methods, that is molecular mechanics, molecular dynamics, and quantum chemistry simulation. The results obtained from CIELab analysis indicated that the formation of ethyl-linked anthocyanin-flavanol pigments resulted in a decrease of L*, a*, b* and C* values, and conversely, a rising of h* value. The 3D molecular simulations revealed that the stability of anthocyanins was as follows: Mv-3-O-glu > Pn-3-O-glu > Cy-3-O-glu. The cis or trans ethyl-linked anthocyanin-flavanol pigments were much more stable than their precursor anthocyanins. Among the pigments, ethyl-linked malvidin-3-O-glucoside-flavanol was more stable than ethyl-linked cyanidin-3-O-glucoside-flavanol and ethyl-linked peonidin-3-O-glucoside-flavanol.
- Published
- 2023
- Full Text
- View/download PDF
10. Comparison of the profiles of first-line PD-1/PD-L1 inhibitors for advanced NSCLC lacking driver gene mutations: a systematic review and Bayesian network meta-analysis
- Author
-
Fu Wenfan, Xu Manman, Shi Xingyuan, Jiang Zeyong, Zhao Jian, and Dai Lu
- Subjects
Therapeutics. Pharmacology ,RM1-950 - Abstract
Background: Numerous first-line immune checkpoint inhibitors (ICI) were developed for patients with advanced non-small cell lung cancer (NSCLC) lacking driver gene mutations. However, this group consists of a heterogeneous patient population, for whom the optimal therapeutic choice is yet to be confirmed. Objective: To identify the best first-line immunotherapy regimen for overall advanced NSCLC patients and different subgroups. Design: Systematic review and Bayesian network meta-analysis (NMA). Methods: We searched several databases to retrieve relevant literature. We performed Bayesian NMA for the overall survival (OS), progression-free survival (PFS), objective response rate (ORR), and treatment-related adverse events (tr-AEs) with a grade equal or more than 3 (grade ⩾ 3 tr-AEs). Subgroup analysis was conducted according to programed death ligand 1 (PD-L1) levels, histologic type, central nervous system (CNS) metastases and tobacco use history. Results: For the PD-L1 non-selective patients, sintilimab plus chemotherapy (sinti-chemo) provided the best OS [hazard ratio (HR) = 0.59, 95% confidence interval (CI):0.42–0.83]. Nivolumab plus bevacizumab plus chemotherapy (nivo-bev-chemo) was comparable to atezolizumab plus bevacizumab plus chemotherapy (atezo-bev-chemo) in prolonging PFS (HR = 0.99, 95% CI: 0.51–1.91). Atezo-bev-chemo remarkably elevated the ORR than chemotherapy (OR = 3.13, 95% CI: 1.51–6.59). Subgroup analysis showed pembrolizumab plus chemotherapy (pembro-chemo) ranked first in OS in subgroups of PD-L1
- Published
- 2023
- Full Text
- View/download PDF
11. Effects of Cholesterol-lowering Agents on Proliferation, Invasion and Neutrophil Extracellular Trap Formation in Liver Cancer Cells
- Author
-
TANG Qiqi, LI Yan, SUN Guowei, LIANG Beibei, and ZHAO Jian
- Subjects
liver cancer ,cholesterol biosynthesis ,proliferation ,invasion ,stemness character ,neutrophil extracellular traps ,Neoplasms. Tumors. Oncology. Including cancer and carcinogens ,RC254-282 - Abstract
Objective To investigate the effects of cholesterol-lowering agents on the proliferation, stemness characters, migration, invasion, and neutrophil extracellular traps formation (NETs) formation in liver cancer cells. Methods ASPP2 or HMGCR gene was knocked down in mouse liver cancer cell Hepa1-6 to establish cells with high or low cholesterol, respectively. Simvastatin and berberine were used to reduce cholesterol synthesis. CCK-8 and plate cloning assays were conducted to detect the proliferation ability of liver cancer cells. Sphere formation assay and qRT-PCR were used to analyze the stemness character and expression of related genes. Wound-healing assay and Transwell assay were used to analyze the ability of cell migration and invasion. Immunofluorescence staining was carried out to analyze the effect of lipid-lowering agent on NETs formation. Results Cholesterol-lowering agents significantly inhibited the proliferation and stemness-related gene expression of Hepa1-6 cells (P < 0.001), significantly inhibited the migration and invasion of Hepa1-6 cells (P < 0.001), and significantly inhibited the neutrophil-induced invasion and formation of NETs (P < 0.001). Conclusion Cholesterol-lowering agents suppress the proliferation and invasion via inhibiting the stemness characters and NETs formation in liver cancer cells. It is a potential strategy for the treatment of liver cancer metastasis.
- Published
- 2022
- Full Text
- View/download PDF
12. Analysis of risk factors for epileptic seizure after combined revascularization in adult patients with moyamoya disease
- Author
-
YANG Shuai-feng, LI Xue-hua, JIN Li-de, TAO Wei-hua, QIAN Xi-ying, and ZHAO Jian-hua
- Subjects
moyamoya disease ,cerebral revascularization ,epilepsy ,postoperative complications ,risk factors ,logistic models ,Neurology. Diseases of the nervous system ,RC346-429 - Abstract
Objective To summarize the clinical features of postoperative epileptic seizure in adults with moyamoya disease (MMD) treated with combined revascularization, and screen the related influencing factors. Methods Total 189 patients with MMD who underwent combined revascularization in Department of Neurosurgery of the First People's Hospital of Yunnan Province from January 2015 to June 2021 were included. Univariate and multivariate forward Logistic regression analysis was used to screen the influencing factors of epilepsy after combined revascularization in MMD patients. Results Logistic regression analysis showed preoperative seizure (OR = 5.523, 95%CI: 1.486-20.524; P = 0.011), revascularization bypass vessel located in frontal lobe (OR = 2.793, 95%CI: 1.027-7.597; P = 0.044), postoperative cerebral infarction (OR = 9.025, 95%CI: 2.121-38.404; P = 0.003), cerebral hyperperfusion syndrome (OR = 3.624, 95%CI: 1.281-10.253; P = 0.015) and intracranial hemorrhage (OR = 5.646, 95%CI: 1.219-26.142; P = 0.027) were the risk factors of postoperative epileptic seizure in patients with MMD. Conclusions Preoperative seizure, revascularization bypass vessel located in frontal lobe, postoperative cerebral infarction, cerebral hyperperfusion syndrome and intracranial hemorrhage are the clinical risk factors of epileptic seizure in adult MMD patients, timely intervention of risk factors may reduce postoperative epileptic seizure in patients with MMD.
- Published
- 2022
- Full Text
- View/download PDF
13. Design and application of an information interaction device for household photovoltaic inverters
- Author
-
Wei Xiaozhao, Zhang Weijian, Liu Hao, Qiao Sen, Zhang Xiangcong, Zhao Jian, and Hu Yurong
- Subjects
Environmental sciences ,GE1-350 - Abstract
The rapid growth of household photovoltaics poses a significant challenge to the safe operation of distribution networks. To enable the unified monitoring of household photovoltaic inverters by power grid companies, this paper introduces an information interaction device for household photovoltaic inverters based on high-speed power line carrier communication and high-speed radio frequency communication. The study begins by analyzing the information access architecture of the household photovoltaic inverter, followed by a presentation of the hardware and software design of the information interaction device. Subsequent experiments conducted at various photovoltaic stations demonstrate an impressive 98% success rate for information interaction, highlighting the device’s robust reliability.
- Published
- 2024
- Full Text
- View/download PDF
14. Interlaminar fracture toughness of carbon/flax fiber hybrid composite
- Author
-
YUAN Weike, LI Yan, and ZHAO Jian
- Subjects
hybrid composite ,flax fiber ,interlaminar fracture toughness ,scanning electron microscopy (sem) ,Motor vehicles. Aeronautics. Astronautics ,TL1-4050 - Abstract
The application of carbon fiber reinforced composites (CFRP) was limited by their poor delamination resistance. However, the hierarchical microstructure of flax fibers could help to improve the interlaminar properties of the composite. Therefore, the interlaminar fracture toughness of CFRP composites can be improved by hybridized flax with carbon fibers. In this study, the carbon/flax fiber hybrid composites (CFFRP) were manufactured by moulding process. The mode Ⅰand mode Ⅱ interlaminar fracture toughnesses of CFFRP composites were studied by double cantilever beam (DCB) and end-notched flexure (ENF) tests, and were compared with those of CFRP composites. The results show that the Mode Ⅰinterlaminar fracture toughness of CFFRP composite is 1.29 kJ/m2, which is about 3.5 times higher than that of CFRP composites (0.37 kJ/m2). The Mode Ⅱinterlaminar fracture toughness is 1.09 kJ/m2 and is about 23.86% higher than that of CFRP composites (0.88 kJ/m2). The interlaminar fractured surfaces of CFRP and CFFRP composites are observed with the aid of scanning electron microscopy (SEM). From the microscopies of the fractured CFRP specimens it can be seen that pure delamination by the peeling of carbon fibers from epoxy resin is obtained for CFRP composites. The surfaces of carbon fibers were relatively clean and few epoxy resin fragments attached. The weak interfacial properties between carbon fiber and epoxy resin cause a lower GⅠc for CFRP composites. On the contrary, from the observation of SEM photographs of flax fiber layers in the interlaminar fractured CFFRP composites, fiber breakage, fiber peeling and fiber entanglement are founded. On the fractured carbon fiber layer, there are some flax fibers tangling with carbon fibers. These multi-scale failure modes due to the unique microstructure of the flax fibers may make the crack propagation become difficult and thus lead to the GⅠc and GⅡc increased for CFFRP composites than those of CFRP composites.
- Published
- 2022
- Full Text
- View/download PDF
15. Single‐cell RNA sequencing reveals a mechanism underlying the susceptibility of the left atrial appendage to intracardiac thrombogenesis during atrial fibrillation
- Author
-
Jie Yang, Hu Tan, Mengjia Sun, Renzheng Chen, Zhao Jian, Yuanbin Song, Jihang Zhang, Shizhu Bian, Bo Zhang, Yi Zhang, Xubin Gao, Zhen Chen, Boji Wu, Xiaowei Ye, Hailin Lv, Zhen Liu, and Lan Huang
- Subjects
ADAMTS1 ,atrial fibrillation ,endocardial endothelial cells ,single‐cell RNA‐sequencing ,TFPI family ,thrombogenesis ,Medicine (General) ,R5-920 - Abstract
Abstract Background Atrial fibrillation (AF) is associated with an increased risk of thrombosis of the left atrial appendage (LAA). However, the molecular mechanisms underlying this site‐specificity remain poorly understood. Here, we present a comparative single‐cell transcriptional profile of paired atrial appendages from patients with AF and illustrate the chamber‐specific properties of the main cell types. Methods Single‐cell RNA sequencing analysis of matched atrial appendage samples from three patients with persistent AF was evaluated by 10× genomics. The AF mice model was created using Tbx5 knockout mice. Validation experiments were performed by glutathione S‐transferase pull‐down assays, coimmunoprecipitation (Co‐IP), cleavage assays and shear stress experiments in vitro. Results In LAA, phenotype switching from endothelial cells to fibroblasts and inflammation associated with proinflammatory macrophage infiltration were observed. Importantly, the coagulation cascade is highly enriched in LAA endocardial endothelial cells (EECs), accompanying the up‐regulation of a disintegrin and metalloproteinase with thrombospondin motifs 1 (ADAMTS1) and the down‐regulation of the tissue factor pathway inhibitor (TFPI) and TFPI2. Similar alterations were verified in an AF mouse model (Tbx5+/−) and EECs treated with simulated AF shear stress in vitro. Furthermore, we revealed that the cleavage of both TFPI and TFPI2 based on their interaction with ADAMTS1 would lead to loss of anticoagulant activities of EECs. Conclusions This study highlights the decrease in the anticoagulant status of EECs in LAA as a potential mechanism underlying the propensity for thrombosis, which may aid the development of anticoagulation therapeutic approaches targeting functionally distinct cell subsets or molecules during AF.
- Published
- 2023
- Full Text
- View/download PDF
16. Optimal Antithrombotic Therapy after Implantation of a Transcatheter Aortic Valve: Warfarin, Aspirin, or Non-Vitamin K Antagonist Oral Anticoagulants? A Systematic Review and Meta-Analysis
- Author
-
Wenjuan Yang, Xiaoyu Fang, Yu Zhu, Fuqin Tang, Zhao Jian, Tianyu Xu, and Xiaoxia Fu
- Subjects
Diseases of the circulatory (Cardiovascular) system ,RC666-701 - Abstract
Abstract. Objective:. Diverse antithrombotic strategies were applied to patients undergoing aortic valve replacement. However, the optimal therapeutic regimen for patients undergoing transcatheter aortic valve implantation/replacement (TAVI/TAVR) remains unclear. The purpose of this study was to compare the efficacy and safety of various antithrombotic therapies following TAVI/TAVR. Methods:. Relevant clinical trials evaluating the effect of anticoagulation or antiplatelet regimens on patients after TAVI/TAVR from inception to September 2020 were identified using the PubMed, EMBASE, and the Cochrane Library databases. The inclusion criteria including (1) all patients underwent TAVI/TAVR; (2) the interventions were antithrombotic strategies that prevent the occurrence of thrombotic events in patients; (3) randomized controlled trials or prospective observational studies; and (4) investigation of at least 1 outcome with a follow-up period of ≥3 months. The exclusion criteria including (1) research content was identical or irrelevant to the purpose of the present study; (2) lack of the required outcome index or availability of fragmentary original information; and (3) the full text is not available. The major outcomes were all-cause mortality, thromboembolic complications, and bleeding events. The Cochrane Collaboration's tool and the Newcastle-Ottawa Scale were used for assessing the risk of bias in included studies. Results:. Thirteen studies (3 randomized controlled trials and 10 non-randomized studies) were identified, with a total of 23,497 patients. Four studies compared direct oral anticoagulants (DOACs) with warfarin, 1 study compared aspirin with warfarin, 6 studies compared aspirin plus clopidogrel (dual antiplatelet therapy (DAPT)) with aspirin monotherapy, and 2 studies compared DAPT and aspirin monotherapy with warfarin concurrently. There were no significant differences found between the DOAC and warfarin groups regarding all-cause mortality (risk ratio (RR): 1.03; 95% confidence interval (CI): 0.65–1.64; P = 0.909; Phet = 0.105), clinical adverse events (RR: 1.59; 95% CI: 0.99–2.58; P = 0.057; Phet = 0.738), or bleeding events (RR: 0.93; 95% CI: 0.78–1.11; P = 0.437; Phet = 0.338). The rates of all-cause mortality (RR: 0.71; 95% CI: 0.54–0.93; P = 0.012; Phet = 0.845) and bleeding events (RR: 0.43; 95% CI: 0.22–0.83; P = 0.012; Phet = 0.569) were lower in the aspirin group versus the warfarin group; however, there was no difference in the rate of clinical adverse events (RR: 0.38; 95% CI: 0.14–1.07; P = 0.068; Phet = 0.593). The DAPT group had an advantage versus the aspirin group in all-cause mortality (RR: 0.89; 95% CI: 0.82–0.98; P = 0.013; Phet = 0.299); however, the incidence of bleeding events (RR: 2.06; 95% CI: 1.39–3.07; P
- Published
- 2022
- Full Text
- View/download PDF
17. Assisting Role of Pulmonary Hypostasis Phenomenon in Diagnosis of Drowning
- Author
-
WU Jian, LI Zeng-qiang, DAI Wen-dao, ZHAO Jian, ZHOU Ya-ping, QUAN Guo-lin, ZHAO Qian-hao, MA Yan-bing, and CHENG Jian-ding
- Subjects
forensic pathology ,drowning ,entering the water after death ,pulmonary hypostasis ,cadaveric signs ,Medicine - Abstract
ObjectiveTo study the phenomenon of pulmonary hypostasis in corpses of various causes of death, and to explore the potential value of this phenomenon in assisting forensic pathological diagnosis of drowning.MethodsA total of 235 cases with clear cause of death through systematic autopsy were collected from January 2011 to June 2021 in Guangzhou. According to the location of body discovery, the cases were divided into the water body group (97 cases) and the non-water body group (138 cases), and the water body group was further divided into the water drowning group (90 cases) and the water non-drowning group (7 cases). Non-water body group was further divided into the non-water drowning group (1 case) and the non-water non-drowning group (137 cases). Three senior forensic pathologists independently reviewed autopsy photos to determine whether there was hypostasis in the lungs. The detection rate of pulmonary hypostasis was calculated.ResultsThe detection rate of pulmonary hypostasis in the water drowning group (90 cases) was 0, and the negative rate was 100%. The detection rate of pulmonary hypostasis in the water non-drowning group (7 cases) was 100% and the negative rate was 0. The detection rate of pulmonary hypostasis in the water body group and in the non-water body group (after excluding 2 cases, 136 cases were calculated) was 7.22% and 87.50%, respectively. There were statistically significant differences in the detection rate of pulmonary hypostasis between water body group and non-water body group, and between water drowning group and water non-drowning group (P
- Published
- 2022
- Full Text
- View/download PDF
18. Evaluation of Inspection Efficiency of Diatom Artificial Intelligence Search System Based on Scanning Electron Microscope
- Author
-
YU Dan-yuan, LIU Jing-jian, LIU Chao, DU Yu-kun, HUANG Ping, ZHANG Ji, YU Wei-min, HU Ying-chao, ZHAO Jian, and CHENG Jian-ding
- Subjects
forensic pathology ,drowning ,diatom test ,artificial intelligence ,automatic searching ,scanning electron microscope ,manual identification ,Medicine - Abstract
ObjectiveTo explore the application values of diatom artificial intelligence (AI) search system in the diagnosis of drowning.MethodsThe liver and kidney tissues of 12 drowned corpses were taken and were performed with the diatom test, the view images were obtained by scanning electron microscopy (SEM). Diatom detection and forensic expert manual identification were carried out under the thresholds of 0.5, 0.7 and 0.9 of the diatom AI search system, respectively. Diatom recall rate, precision rate and image exclusion rate were used to detect and compare the efficiency of diatom AI search system.ResultsThere was no statistical difference between the number of diatoms detected in the target marked by the diatom AI search system and the number of diatoms identified manually (P>0.05); the recall rates of the diatom AI search system were statistically different under different thresholds (P
- Published
- 2022
- Full Text
- View/download PDF
19. Comparison among Four Deep Learning Image Classification Algorithms in AI-based Diatom Test
- Author
-
ZHU Yong-zheng, ZHANG Ji, CHENG Qi, YU Hui-xiao, DENG Kai-fei, ZHANG Jian-hua, QIN Zhi-qiang, ZHAO Jian, SUN Jun-hong, and HUANG Ping
- Subjects
forensic pathology ,artificial intelligence ,drowning ,deep learning ,diatom test ,convolutional neural network ,Medicine - Abstract
ObjectiveTo select four algorithms with relatively balanced complexity and accuracy among deep learning image classification algorithms for automatic diatom recognition, and to explore the most suitable classification algorithm for diatom recognition to provide data reference for automatic diatom testing research in forensic medicine.MethodsThe “diatom” and “background” small sample size data set (20 000 images) of digestive fluid smear of corpse lung tissue in water were built to train, validate and test four convolutional neural network (CNN) models, including VGG16, ResNet50, InceptionV3 and Inception-ResNet-V2. The receiver operating characteristic curve (ROC) of subjects and confusion matrixes were drawn, recall rate, precision rate, specificity, accuracy rate and F1 score were calculated, and the performance of each model was systematically evaluated.ResultsThe InceptionV3 model achieved much better results than the other three models with a balanced recall rate of 89.80%, a precision rate of 92.58%. The VGG16 and Inception-ResNet-V2 had similar diatom recognition performance. Although the performance of diatom recall and precision detection could not be balanced, the recognition ability was acceptable. ResNet50 had the lowest diatom recognition performance, with a recall rate of 55.35%. In terms of feature extraction, the four models all extracted the features of diatom and background and mainly focused on diatom region as the main identification basis.ConclusionIncluding the Inception-dependent model, which has stronger directivity and targeting in feature extraction of diatom. The InceptionV3 achieved the best performance on diatom identification and feature extraction compared to the other three models. The InceptionV3 is more suitable for daily forensic diatom examination.
- Published
- 2022
- Full Text
- View/download PDF
20. Application of Diatoms Quantitative Analysis in the Diagnosis of Drowning
- Author
-
DU Yu-kun, ZHANG Tian-ye, LIU Jing-jian, LIU Chao, KANG Xiao-dong, ZHENG Dong-yun, SHI He, XU Qu-yi, MA Kai-jun, and ZHAO Jian
- Subjects
forensic pathology ,drowning ,diatom test ,quantitative analysis ,microwave digestion-vacuum filtration-automated scanning electron microscopy method ,Medicine - Abstract
ObjectiveTo retrospectively analyze diatom test cases of corpses in water and discuss the value of quantitative analysis of diatoms in the diagnosis of drowning.MethodsA total of 490 cases of water-related death were collected. They were divided into drowning group and postmortem immersion group according to the cause of death. Diatoms in lung, liver, kidney tissue and water sample were analyzed quantitatively by microwave digestion-vacuum filtration-automated scanning electron microscopy (MD-VF-Auto SEM) method. The ratios of content of diatoms in lung tissue and water sample (CL/CD) were calculated.ResultsThe results of diatom test for three organs (lung, liver and kidney) were all positive in 400 cases (85.5%); the content of diatom in lung, liver, kidney tissues, and water samples of drowning group were (113 235.9±317 868.1), (26.7±75.6), (23.3±52.2) and (12 113.3±21 760.0) cells/10 g, respectively; the species of diatom were (7.5±2.8), (2.6±1.9), (2.9±2.1) and (8.9±3.0) types, respectively; the CL/CD of drowning group and postmortem immersion group were (100.6±830.7) and (0.3±0.4), respectively.ConclusionQuantitative analysis of diatoms can provide supportive evidence for the diagnosis of drowning, and the parameter CL/CD can be introduced into the analysis to make a more accurate diagnosis of drowning.
- Published
- 2022
- Full Text
- View/download PDF
21. Effects of Digestive Temperature and Time on Diatom Test
- Author
-
LIU Jing-jian, DU Yu-kun, ZHAO Jian, KANG Xiao-dong, YU Zhong-hao, ZHENG Dong-yun, SHI He, XU Qu-yi, CHEN Li-fang, and LIU Chao
- Subjects
forensic pathology ,diatom test ,digestive temperature ,digestive time ,Medicine - Abstract
ObjectiveTo study the effects of temperature and time for diatoms digestion and find out suitable digestive temperature and time.MethodsEighty pieces of liver tissues were collected, each piece of tissue was 2 g, and 2 mL Pearl River water was added to each piece of tissue. The digestion temperature was set at 100 ℃, 120 ℃, 140 ℃, 160 ℃, 180 ℃ and the digestion time was set at 40, 50, 60, 70, 80 min. The liver tissue and water mixture were divided into 8 portions in each group. All the samples were tested by microwave digestive - vacuum filtration - automated scanning electron microscopy method. The quantity of diatom recovered and the quality of residue on the membrane were recorded.ResultsWhen the digestion time was set to 60 min, there were statistically significant differences in the number of diatoms recovered at different temperatures (P0.05).ConclusionThe effect of diatom digestion is related to temperature and time. When the digestion temperature was 140 ℃ and the digestion time was 40, 50 and 60 min, it is favorable for diatom test.
- Published
- 2022
- Full Text
- View/download PDF
22. Research Progress of Automatic Diatom Test by Artificial Intelligence
- Author
-
ZHU Yong-zheng, ZHANG Ji, CHENG Qi, DENG Kai-fei, MA Kai-jun, ZHANG Jian-hua, ZHAO Jian, SUN Jun-hong, HUANG Ping, and QIN Zhi-qiang
- Subjects
forensic pathology ,drowning ,diatom test ,artificial intelligence ,deep learning ,review ,Medicine - Abstract
Diatom test is the main laboratory test method in the diagnosis of drowning in forensic medicine. It plays an important role in differentiating the antemortem drowning from the postmortem drowning and inferring drowning site. Artificial intelligence (AI) automatic diatom test is a technological innovation in forensic drowning diagnosis which is based on morphological characteristics of diatom, the application of AI algorithm to automatic identification and classification of diatom in tissues and organs. This paper discusses the morphological diatom test methods and reviews the research progress of automatic diatom recognition and classification involving AI algorithms. AI deep learning algorithm can assist diatom testing to obtain objective, accurate, and efficient qualitative and quantitative analysis results, which is expected to become a new direction of diatom testing research in the drowning of forensic medicine in the future.
- Published
- 2022
- Full Text
- View/download PDF
23. Pathway of Diatoms Enter Experimental Rabbits through the Lymphatic System of the Digestive Tract
- Author
-
DU Yu-kun, LIU Jing-jian, KANG Xiao-dong, YU Zhong-hao, ZHENG Dong-yun, SHI He, XU Qu-yi, REN Jian-jun, LIU Chao, and ZHAO Jian
- Subjects
forensic pathology ,drowning ,diatom ,digestive tract ,lymph ,gavage model ,Medicine - Abstract
ObjectiveTo study whether diatoms can enter the body through the lymphatic system of the digestive tract.MethodsTwenty experimental rabbits were divided into the test group and the control group randomly, and intragastric administration was performed with 20 mL water sample from the Pearl River and 20 mL ultrapure water, respectively. After 30 min, lymph, lungs, livers and kidneys were extracted for the diatom test. The concentration, size and type of diatoms were recorded.ResultsThe concentration of diatoms of the test group was higher than that of the control group (P
- Published
- 2022
- Full Text
- View/download PDF
24. Comparison of Application of MD-VF-Auto SEM Method and Plankton Gene Multiplex PCR System in the Diagnosis of Drowning
- Author
-
ZHANG Jian-miao, LIN Tian-chun, LIAO Zhen-yu, DU Yu-kun, YU Zhong-hao, LIU Jing-jian, WU Sai-qun, KANG Xiao-dong, XU Qu-yi, SHI He, ZHAO Jian, LIU Chao, and GU Dang-en
- Subjects
forensic pathology ,drowning ,diatom test ,plankton ,multiplex pcr system ,microwave digestion-vacuum filtration-automated scanning electron microscopy method ,Medicine - Abstract
ObjectiveTo compare the application effect of microwave digestion - vacuum filtration - automated scanning electron microscopy (MD-VF-Auto SEM) method and plankton gene multiplex PCR system in the diagnosis of drowning.MethodsLung, liver and kidney tissue of 10 non-drowning cases and 50 drowning cases were prepared for further MD-VF-Auto SEM method analysis and plankton gene multiplex PCR system analysis. The positive detection rate of the two methods in each tissue was calculated.ResultsThe positive rate of the MD-VF-Auto SEM method detecting diatoms in drowning cases was 100%, and few diatoms were detected in the liver and kidney tissues of 6 non-drowning cases. By using the plankton gene multiplex PCR system, the diatom positive rate of drowning cases was 84%, and all the non-drowning cases were negative. There were significant differences in the positive rate of the liver, kidney tissues between MD-VF-Auto SEM method and plankton gene multiplex PCR system (P0.05).ConclusionMD-VF-Auto SEM method is more sensitive than plankton gene multiplex PCR system in diatom test. But the plankton gene multiplex PCR system can also detect plankton other than diatoms. Combination of the two methods can provide a more reliable basis for the diagnosis of drowning.
- Published
- 2022
- Full Text
- View/download PDF
25. Brickify: Enabling Expressive Design Intent Specification through Direct Manipulation on Design Tokens
- Author
-
Shi, Xinyu, Wang, Yinghou, Rossi, Ryan, and Zhao, Jian
- Subjects
Computer Science - Human-Computer Interaction - Abstract
Expressing design intent using natural language prompts requires designers to verbalize the ambiguous visual details concisely, which can be challenging or even impossible. To address this, we introduce Brickify, a visual-centric interaction paradigm -- expressing design intent through direct manipulation on design tokens. Brickify extracts visual elements (e.g., subject, style, and color) from reference images and converts them into interactive and reusable design tokens that can be directly manipulated (e.g., resize, group, link, etc.) to form the visual lexicon. The lexicon reflects users' intent for both what visual elements are desired and how to construct them into a whole. We developed Brickify to demonstrate how AI models can interpret and execute the visual lexicon through an end-to-end pipeline. In a user study, experienced designers found Brickify more efficient and intuitive than text-based prompts, allowing them to describe visual details, explore alternatives, and refine complex designs with greater ease and control., Comment: CHI 2025
- Published
- 2025
26. DiffBrush:Just Painting the Art by Your Hands
- Author
-
Chu, Jiaming, Jin, Lei, Wang, Tao, Xing, Junliang, and Zhao, Jian
- Subjects
Computer Science - Computer Vision and Pattern Recognition ,Computer Science - Multimedia - Abstract
The rapid development of image generation and editing algorithms in recent years has enabled ordinary user to produce realistic images. However, the current AI painting ecosystem predominantly relies on text-driven diffusion models (T2I), which pose challenges in accurately capturing user requirements. Furthermore, achieving compatibility with other modalities incurs substantial training costs. To this end, we introduce DiffBrush, which is compatible with T2I models and allows users to draw and edit images. By manipulating and adapting the internal representation of the diffusion model, DiffBrush guides the model-generated images to converge towards the user's hand-drawn sketches for user's specific needs without additional training. DiffBrush achieves control over the color, semantic, and instance of objects in images by continuously guiding the latent and instance-level attention map during the denoising process of the diffusion model. Besides, we propose a latent regeneration, which refines the randomly sampled noise in the diffusion model, obtaining a better image generation layout. Finally, users only need to roughly draw the mask of the instance (acceptable colors) on the canvas, DiffBrush can naturally generate the corresponding instance at the corresponding location.
- Published
- 2025
27. AccessFixer: Enhancing GUI Accessibility for Low Vision Users With R-GCN Model
- Author
-
Zhang, Mengxi, Liu, Huaxiao, Chen, Chunyang, Gao, Guangyong, Li, Han, and Zhao, Jian
- Subjects
Computer Science - Software Engineering - Abstract
The Graphical User Interface (GUI) plays a critical role in the interaction between users and mobile applications (apps), aiming at facilitating the operation process. However, due to the variety of functions and non-standardized design, GUIs might have many accessibility issues, like the size of components being too small or their intervals being narrow. These issues would hinder the operation of low vision users, preventing them from obtaining information accurately and conveniently. Although several technologies and methods have been proposed to address these issues, they are typically confined to issue identification, leaving the resolution in the hands of developers. Moreover, it can be challenging to ensure that the color, size, and interval of the fixed GUIs are appropriately compared to the original ones. In this work, we propose a novel approach named AccessFixer, which utilizes the Relational-Graph Convolutional Neural Network (R-GCN) to simultaneously fix three kinds of accessibility issues, including small sizes, narrow intervals, and low color contrast in GUIs. With AccessFixer, the fixed GUIs would have a consistent color palette, uniform intervals, and adequate size changes achieved through coordinated adjustments to the attributes of related components. Our experiments demonstrate the effectiveness and usefulness of AccessFixer in fixing GUI accessibility issues. After fixing 30 real-world apps, our approach solves an average of 81.2% of their accessibility issues. Also, we apply AccessFixer to 10 open-source apps by submitting the fixed results with pull requests (PRs) on GitHub. The results demonstrate that developers approve of our submitted fixed GUIs, with 8 PRs being merged or under fixing. A user study examines that low vision users host a positive attitude toward the GUIs fixed by our method., Comment: 18 pages, 14 figures, has been published in IEEE Transactions on Software Engineering
- Published
- 2025
- Full Text
- View/download PDF
28. Don't Confuse! Redrawing GUI Navigation Flow in Mobile Apps for Visually Impaired Users
- Author
-
Zhang, Mengxi, Liu, Huaxiao, Zhou, Yuheng, Chen, Chunyang, Huang, Pei, and Zhao, Jian
- Subjects
Computer Science - Software Engineering - Abstract
Mobile applications (apps) are integral to our daily lives, offering diverse services and functionalities. They enable sighted users to access information coherently in an extremely convenient manner. However, it remains unclear if visually impaired users, who rely solely on the screen readers (e.g., Talkback) to navigate and access app information, can do so in the correct and reasonable order. This may result in significant information bias and operational errors. Considering these issues, in this work, we proposed a method named RGNF (Re-draw GUI Navigation Flow). It aimed to enhance the understandability and coherence of accessing the content of each component within the Graphical User Interface (GUI), together with assisting developers in creating well-designed GUI navigation flow (GNF). This method was inspired by the characteristics identified in our preliminary study, where visually impaired users expected navigation to be associated with close position and similar shape of GUI components that were read consecutively. Thus, our method relied on the principles derived from the Gestalt psychological model, aiming to group GUI components into different regions according to the laws of proximity and similarity, thereby redrawing the GNFs. To evaluate the effectiveness of our method, we calculated sequence similarity values before and after redrawing the GNF, and further employed the tools proposed by Alotaibi et al. to measure the reachability of GUI components. Our results demonstrated a substantial improvement in similarity (0.921) compared to the baseline (0.624), together with the reachability (90.31%) compared to the baseline GNF (74.35%). Furthermore, a qualitative user study revealed that our method had a positive effect on providing visually impaired users with an improved user experience., Comment: 18 pages, 13 figures, has be published in IEEE Transactions on Software Engineering
- Published
- 2025
- Full Text
- View/download PDF
29. Are your apps accessible? A GCN-based accessibility checker for low vision users
- Author
-
Zhang, Mengxi, Liu, Huaxiao, Song, Shenning, Chen, Chunyang, Huang, Pei, and Zhao, Jian
- Subjects
Computer Science - Software Engineering - Abstract
Context: Accessibility issues (e.g., small size and narrow interval) in mobile applications (apps) lead to obstacles for billions of low vision users in interacting with Graphical User Interfaces (GUIs). Although GUI accessibility scanning tools exist, most of them perform rule-based check relying on complex GUI hierarchies. This might make them detect invisible redundant information, cannot handle small deviations, omit similar components, and is hard to extend. Objective: In this paper, we propose a novel approach, named ALVIN (Accessibility Checker for Low Vision), which represents the GUI as a graph and adopts the Graph Convolutional Neural Networks (GCN) to label inaccessible components. Method: ALVIN removes invisible views to prevent detecting redundancy and uses annotations from low vision users to handle small deviations. Also, the GCN model could consider the relations between GUI components, connecting similar components and reducing the possibility of omission. ALVIN only requires users to annotate the relevant dataset when detecting new kinds of issues. Results: Our experiments on 48 apps demonstrate the effectiveness of ALVIN, with precision of 83.5%, recall of 78.9%, and F1-score of 81.2%, outperforming baseline methods. In RQ2, the usefulness is verified through 20 issues submitted to open-source apps. The RQ3 also illustrates the GCN model is better than other models. Conclusion: To summarize, our proposed approach can effectively detect accessibility issues in GUIs for low vision users, thereby guiding developers in fixing them efficiently., Comment: 27 pages, 9 figures, has been published in Information and Software Technology
- Published
- 2025
- Full Text
- View/download PDF
30. BIG-AOME: Designing Bodily Interaction Gamification towards Anti-sedentary Online Meeting Environments
- Author
-
Jiang, Jiaqi, Li, Shanghao, Li, Xian, Xu, Yingxin, Zhao, Jian, and An, Pengcheng
- Subjects
Computer Science - Human-Computer Interaction ,H.5.2 ,K.8.0 - Abstract
Online meetings have become an integral part of daily life, but prolonged screen time poses significant health risks. While various interventions address sedentary lifestyles, few focus on mitigating sedentary behavior during online meetings. Design opportunities in this context remain underexplored. This study investigates the design of gamified bodily interactions as anti-sedentary measures during online meetings using a research through design approach. In collaboration with 11 users, we co-designed and iterated three prototypes, resulting in the BIG-AOME (Bodily Interaction Gamification towards Anti-sedentary Online Meeting Environments) framework. User studies with 15 participants across three groups evaluated these prototypes through semi-structured interviews analyzed using Hsieh's qualitative content analysis. Findings show that gamified bodily interactions encourage physical movement while reducing awkwardness during online meetings. Participants valued the social engagement fostered by cooperative and competitive elements in these games, enhancing social dynamics while encouraging physical movement. Such games can also serve as online icebreakers or playful decision-making tools. This study offers a comprehensive analysis of design dimensions within the BIG-AOME framework, including body engagement, attention, bodily interplay, timeliness, and virtual/physical environments, highlighting the potential of anti-sedentary bodily interactions to mitigate sedentary behavior and enhance social connections in online meetings., Comment: Accepted for publication in Journal of Medical Internet Research (JMIR) Serious Games. This version is a preprint submitted to arXiv
- Published
- 2025
31. TherAIssist: Assisting Art Therapy Homework and Client-Practitioner Collaboration through Human-AI Interaction
- Author
-
Liu, Di, Bai, Jingwen, Zhang, Zhuoyi, Zhang, Yilin, Zhang, Zhenhao, Zhao, Jian, and An, Pengcheng
- Subjects
Computer Science - Human-Computer Interaction - Abstract
Art therapy homework is essential for fostering clients' reflection on daily experiences between sessions. However, current practices present challenges: clients often lack guidance for completing tasks that combine art-making and verbal expression, while therapists find it difficult to track and tailor homework. How HCI systems might support art therapy homework remains underexplored. To address this, we present TherAIssist, comprising a client-facing application leveraging human-AI co-creative art-making and conversational agents to facilitate homework, and a therapist-facing application enabling customization of homework agents and AI-compiled homework history. A 30-day field study with 24 clients and 5 therapists showed how TherAIssist supported clients' homework and reflection in their everyday settings. Results also revealed how therapists infused their practice principles and personal touch into the agents to offer tailored homework, and how AI-compiled homework history became a meaningful resource for in-session interactions. Implications for designing human-AI systems to facilitate asynchronous client-practitioner collaboration are discussed.
- Published
- 2025
32. Leader and Follower: Interactive Motion Generation under Trajectory Constraints
- Author
-
Wang, Runqi, Ma, Caoyuan, Zhao, Jian, Xu, Hanrui, Sun, Dongfang, Chen, Haoyang, Xiong, Lin, Wang, Zheng, and Li, Xuelong
- Subjects
Computer Science - Robotics ,Computer Science - Artificial Intelligence - Abstract
With the rapid advancement of game and film production, generating interactive motion from texts has garnered significant attention due to its potential to revolutionize content creation processes. In many practical applications, there is a need to impose strict constraints on the motion range or trajectory of virtual characters. However, existing methods that rely solely on textual input face substantial challenges in accurately capturing the user's intent, particularly in specifying the desired trajectory. As a result, the generated motions often lack plausibility and accuracy. Moreover, existing trajectory - based methods for customized motion generation rely on retraining for single - actor scenarios, which limits flexibility and adaptability to different datasets, as well as interactivity in two-actor motions. To generate interactive motion following specified trajectories, this paper decouples complex motion into a Leader - Follower dynamic, inspired by role allocation in partner dancing. Based on this framework, this paper explores the motion range refinement process in interactive motion generation and proposes a training-free approach, integrating a Pace Controller and a Kinematic Synchronization Adapter. The framework enhances the ability of existing models to generate motion that adheres to trajectory by controlling the leader's movement and correcting the follower's motion to align with the leader. Experimental results show that the proposed approach, by better leveraging trajectory information, outperforms existing methods in both realism and accuracy.
- Published
- 2025
33. 'You'll Be Alice Adventuring in Wonderland!' Processes, Challenges, and Opportunities of Creating Animated Virtual Reality Stories
- Author
-
Yuan, Lin-Ping, Han, Feilin, Xie, Liwenhan, Zhang, Junjie, Zhao, Jian, and Qu, Huamin
- Subjects
Computer Science - Human-Computer Interaction ,Computer Science - Graphics ,Computer Science - Multimedia - Abstract
Animated virtual reality (VR) stories, combining the presence of VR and the artistry of computer animation, offer a compelling way to deliver messages and evoke emotions. Motivated by the growing demand for immersive narrative experiences, more creators are creating animated VR stories. However, a holistic understanding of their creation processes and challenges involved in crafting these stories is still limited. Based on semi-structured interviews with 21 animated VR story creators, we identify ten common stages in their end-to-end creation processes, ranging from idea generation to evaluation, which form diverse workflows that are story-driven or visual-driven. Additionally, we highlight nine unique issues that arise during the creation process, such as a lack of reference material for multi-element plots, the absence of specific functionalities for story integration, and inadequate support for audience evaluation. We compare the creation of animated VR stories to general XR applications and distill several future research opportunities., Comment: Conditionally accepted to the ACM Conference on Human Factors in Computing Systems (CHI'25)
- Published
- 2025
- Full Text
- View/download PDF
34. Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling
- Author
-
Liu, Runze, Gao, Junqi, Zhao, Jian, Zhang, Kaiyan, Li, Xiu, Qi, Biqing, Ouyang, Wanli, and Zhou, Bowen
- Subjects
Computer Science - Computation and Language - Abstract
Test-Time Scaling (TTS) is an important method for improving the performance of Large Language Models (LLMs) by using additional computation during the inference phase. However, current studies do not systematically analyze how policy models, Process Reward Models (PRMs), and problem difficulty influence TTS. This lack of analysis limits the understanding and practical use of TTS methods. In this paper, we focus on two core questions: (1) What is the optimal approach to scale test-time computation across different policy models, PRMs, and problem difficulty levels? (2) To what extent can extended computation improve the performance of LLMs on complex tasks, and can smaller language models outperform larger ones through this approach? Through comprehensive experiments on MATH-500 and challenging AIME24 tasks, we have the following observations: (1) The compute-optimal TTS strategy is highly dependent on the choice of policy model, PRM, and problem difficulty. (2) With our compute-optimal TTS strategy, extremely small policy models can outperform larger models. For example, a 1B LLM can exceed a 405B LLM on MATH-500. Moreover, on both MATH-500 and AIME24, a 0.5B LLM outperforms GPT-4o, a 3B LLM surpasses a 405B LLM, and a 7B LLM beats o1 and DeepSeek-R1, while with higher inference efficiency. These findings show the significance of adapting TTS strategies to the specific characteristics of each task and model and indicate that TTS is a promising approach for enhancing the reasoning abilities of LLMs.
- Published
- 2025
35. Code Shaping: Iterative Code Editing with Free-form AI-Interpreted Sketching
- Author
-
Yen, Ryan, Zhao, Jian, and Vogel, Daniel
- Subjects
Computer Science - Human-Computer Interaction - Abstract
We introduce the concept of code shaping, an interaction paradigm for editing code using free-form sketch annotations directly on top of the code and console output. To evaluate this concept, we conducted a three-stage design study with 18 different programmers to investigate how sketches can communicate intended code edits to an AI model for interpretation and execution. The results show how different sketches are used, the strategies programmers employ during iterative interactions with AI interpretations, and interaction design principles that support the reconciliation between the code editor and sketches. Finally, we demonstrate the practical application of the code shaping concept with two use case scenarios, illustrating design implications from the study., Comment: 19 pages
- Published
- 2025
36. Multiple Queries with Multiple Keys: A Precise Prompt Matching Paradigm for Prompt-based Continual Learning
- Author
-
Tu, Dunwei, Yi, Huiyu, Wang, Yuchi, Xu, Baile, Zhao, Jian, and Shen, Furao
- Subjects
Computer Science - Computer Vision and Pattern Recognition - Abstract
Continual learning requires machine learning models to continuously acquire new knowledge in dynamic environments while avoiding the forgetting of previous knowledge. Prompt-based continual learning methods effectively address the issue of catastrophic forgetting through prompt expansion and selection. However, existing approaches often suffer from low accuracy in prompt selection, which can result in the model receiving biased knowledge and making biased predictions. To address this issue, we propose the Multiple Queries with Multiple Keys (MQMK) prompt matching paradigm for precise prompt selection. The goal of MQMK is to select the prompts whose training data distribution most closely matches that of the test sample. Specifically, Multiple Queries enable precise breadth search by introducing task-specific knowledge, while Multiple Keys perform deep search by representing the feature distribution of training samples at a fine-grained level. Experiments show that MQMK enhances the prompt matching rate by over 30% in challenging scenarios and achieves state-of-the-art performance on three widely adopted continual learning benchmarks. Once this paper is accepted, we will release the code.
- Published
- 2025
37. Research progress in high entropy alloys by additive manufacturing
- Author
-
WEI Shui-miao, MA Pan, JI Peng-cheng, MA Yong-chao, WANG Can, ZHAO Jian, and YU Zhi-shui
- Subjects
high-entropy alloys (heas) ,additive manufacturing ,microstructure ,strengthening mechanism ,Materials of engineering and construction. Mechanics of materials ,TA401-492 - Abstract
Based on different high-entropy alloys (HEAs) systems, the latest research progress in additive manufactured high-entropy alloys was reviewed. The rapid solidification microstructure, segregation and precipitation behaviors of high-entropy alloys fabricated by additive manufacturing with different compositions were described. Especially, the analysis was focused on the mechanical properties, deformation and strengthening mechanisms. It was pointed out that the appropriate additive manufacturing process should be selected for different high-entropy alloy systems, and the influencing factors of forming quality need to be further studied. Finally, it was proposed that high-entropy alloys with both excellent strength and high plasticity can be developed and prepared by additive manufacturing technology.
- Published
- 2021
- Full Text
- View/download PDF
38. Research Progress of Glucose Metabolism Reprogramming in Basic Research and Clinical Diagnosis and Treatment of Tumor
- Author
-
JIANG Yuan, CHEN Ya, XIE Yangyang, HUANG Xinyu, ZHAO Jian, and LIANG Beibei
- Subjects
malignant tumor ,glucose metabolism reprogramming ,key target ,Neoplasms. Tumors. Oncology. Including cancer and carcinogens ,RC254-282 - Abstract
Tumors provide materials and energy for themselves through glucose metabolic reprogramming to meet the needs of their rapid growth. Studies have shown that a variety of metabolic enzymes and signal molecules involved in glucose metabolism play an important role in tumorigenesis and development, and are considered to be important targets for the diagnosis and treatment of malignant tumors. The detection technology based on tumor glucose metabolism has been widely used in basic research and clinical diagnosis and treatment of tumor. This paper summarizes the characteristics, driving factors and key regulatory targets of tumor glucose metabolism, and the detection techniques and functions of tumor glucose metabolism from the aspects of basic medical research and clinical diagnosis and treatment.
- Published
- 2021
- Full Text
- View/download PDF
39. Hard and soft Lewis-base behavior for efficient and stable CsPbBr3 perovskite light-emitting diodes
- Author
-
Sun Changjiu, Wei Junli, Zhao Jian, Jiang Yuanzhi, Wang Yilong, Hu Haiqing, Wang Xin, Zhang Yongqin, and Yuan Mingjian
- Subjects
all-inorganic perovskite ,cspbbr3 ,lewis base ,perovskite light-emitting diodes ,Physics ,QC1-999 - Abstract
All-inorganic CsPbBr3 perovskite is an attractive emission material for high-stability perovskite light-emitting diodes (PeLEDs), due to the high thermal and chemical stability. However, the external quantum efficiencies (EQEs) of CsPbBr3 based PeLEDs are still far behind their organic–inorganic congeners. Massive defect states on the surface of CsPbBr3 perovskite grains should be the main reason. Lewis base additives have been widely used to passivate surface defects. However, systematic investigations which relate to improving the passivation effect via rational molecule design are still lacking. Here, we demonstrate that the CsPbBr3 film’s optical and electrical properties can be significantly boosted by tailoring the hardness–softness of the Lewis base additives. Three carboxylate Lewis bases with different tail groups are selected to in-situ passivate CsPbBr3 perovskite films. Our research indicates that 4-(trifluoromethyl) benzoate acid anion (TBA−) with the powerful electron-withdrawing group trifluoromethyl and benzene ring possesses the softest COO− bonding head. TBA− thus acts as a soft Lewis base and possesses a robust combination with unsaturated lead atoms caused by halogen vacancies. Based on this, the all-inorganic CsPbBr3 PeLEDs with a maximum EQE up to 16.75% and a half-lifetime over 129 h at an initial brightness of 100 cd m−2 is thus delivered.
- Published
- 2021
- Full Text
- View/download PDF
40. EraseAnything: Enabling Concept Erasure in Rectified Flow Transformers
- Author
-
Gao, Daiheng, Lu, Shilin, Walters, Shaw, Zhou, Wenbo, Chu, Jiaming, Zhang, Jie, Zhang, Bang, Jia, Mengxi, Zhao, Jian, Fan, Zhaoxin, and Zhang, Weiming
- Subjects
Computer Science - Computer Vision and Pattern Recognition - Abstract
Removing unwanted concepts from large-scale text-to-image (T2I) diffusion models while maintaining their overall generative quality remains an open challenge. This difficulty is especially pronounced in emerging paradigms, such as Stable Diffusion (SD) v3 and Flux, which incorporate flow matching and transformer-based architectures. These advancements limit the transferability of existing concept-erasure techniques that were originally designed for the previous T2I paradigm (e.g., SD v1.4). In this work, we introduce EraseAnything, the first method specifically developed to address concept erasure within the latest flow-based T2I framework. We formulate concept erasure as a bi-level optimization problem, employing LoRA-based parameter tuning and an attention map regularizer to selectively suppress undesirable activations. Furthermore, we propose a self-contrastive learning strategy to ensure that removing unwanted concepts does not inadvertently harm performance on unrelated ones. Experimental results demonstrate that EraseAnything successfully fills the research gap left by earlier methods in this new T2I paradigm, achieving state-of-the-art performance across a wide range of concept erasure tasks., Comment: 24 pages, 18 figures
- Published
- 2024
41. MambaVO: Deep Visual Odometry Based on Sequential Matching Refinement and Training Smoothing
- Author
-
Wang, Shuo, Li, Wanting, Wang, Yongcai, Fan, Zhaoxin, Huang, Zhe, Cai, Xudong, Zhao, Jian, and Li, Deying
- Subjects
Computer Science - Computer Vision and Pattern Recognition - Abstract
Deep visual odometry has demonstrated great advancements by learning-to-optimize technology. This approach heavily relies on the visual matching across frames. However, ambiguous matching in challenging scenarios leads to significant errors in geometric modeling and bundle adjustment optimization, which undermines the accuracy and robustness of pose estimation. To address this challenge, this paper proposes MambaVO, which conducts robust initialization, Mamba-based sequential matching refinement, and smoothed training to enhance the matching quality and improve the pose estimation in deep visual odometry. Specifically, when a new frame is received, it is matched with the closest keyframe in the maintained Point-Frame Graph (PFG) via the semi-dense based Geometric Initialization Module (GIM). Then the initialized PFG is processed by a proposed Geometric Mamba Module (GMM), which exploits the matching features to refine the overall inter-frame pixel-to-pixel matching. The refined PFG is finally processed by deep BA to optimize the poses and the map. To deal with the gradient variance, a Trending-Aware Penalty (TAP) is proposed to smooth training by balancing the pose loss and the matching loss to enhance convergence and stability. A loop closure module is finally applied to enable MambaVO++. On public benchmarks, MambaVO and MambaVO++ demonstrate SOTA accuracy performance, while ensuring real-time running performance with low GPU memory requirement. Codes will be publicly available.
- Published
- 2024
42. SMAC-Hard: Enabling Mixed Opponent Strategy Script and Self-play on SMAC
- Author
-
Deng, Yue, Yu, Yan, Ma, Weiyu, Wang, Zirui, Zhu, Wenhui, Zhao, Jian, and Zhang, Yin
- Subjects
Computer Science - Artificial Intelligence - Abstract
The availability of challenging simulation environments is pivotal for advancing the field of Multi-Agent Reinforcement Learning (MARL). In cooperative MARL settings, the StarCraft Multi-Agent Challenge (SMAC) has gained prominence as a benchmark for algorithms following centralized training with decentralized execution paradigm. However, with continual advancements in SMAC, many algorithms now exhibit near-optimal performance, complicating the evaluation of their true effectiveness. To alleviate this problem, in this work, we highlight a critical issue: the default opponent policy in these environments lacks sufficient diversity, leading MARL algorithms to overfit and exploit unintended vulnerabilities rather than learning robust strategies. To overcome these limitations, we propose SMAC-HARD, a novel benchmark designed to enhance training robustness and evaluation comprehensiveness. SMAC-HARD supports customizable opponent strategies, randomization of adversarial policies, and interfaces for MARL self-play, enabling agents to generalize to varying opponent behaviors and improve model stability. Furthermore, we introduce a black-box testing framework wherein agents are trained without exposure to the edited opponent scripts but are tested against these scripts to evaluate the policy coverage and adaptability of MARL algorithms. We conduct extensive evaluations of widely used and state-of-the-art algorithms on SMAC-HARD, revealing the substantial challenges posed by edited and mixed strategy opponents. Additionally, the black-box strategy tests illustrate the difficulty of transferring learned policies to unseen adversaries. We envision SMAC-HARD as a critical step toward benchmarking the next generation of MARL algorithms, fostering progress in self-play methods for multi-agent systems. Our code is available at https://github.com/devindeng94/smac-hard.
- Published
- 2024
43. Anger Speaks Louder? Exploring the Effects of AI Nonverbal Emotional Cues on Human Decision Certainty in Moral Dilemmas
- Author
-
Zhang, Chenyi, Zhang, Zhenhao, Zhang, Wei, Zeng, Tian, Sun, Black, Zhao, Jian, and An, Pengcheng
- Subjects
Computer Science - Human-Computer Interaction - Abstract
Exploring moral dilemmas allows individuals to navigate moral complexity, where a reversal in decision certainty, shifting toward the opposite of one's initial choice, could reflect open-mindedness and less rigidity. This study probes how nonverbal emotional cues from conversational agents could influence decision certainty in moral dilemmas. While existing research heavily focused on verbal aspects of human-agent interaction, we investigated the impact of agents expressing anger and sadness towards the moral situations through animated chat balloons. We compared these with a baseline where agents offered same responses without nonverbal cues. Results show that agents displaying anger significantly caused reversal shifts in decision certainty. The interaction between participant gender and agents' nonverbal emotional cues significantly affects participants' perception of AI's influence. These findings reveal that even subtly altering agents' nonverbal cues may impact human moral decisions, presenting both opportunities to leverage these effects for positive outcomes and ethical risks for future human-AI systems., Comment: 19 pages, 5 figures
- Published
- 2024
44. RL-LLM-DT: An Automatic Decision Tree Generation Method Based on RL Evaluation and LLM Enhancement
- Author
-
Lin, Junjie, Zhao, Jian, Liu, Lin, Deng, Yue, Zhao, Youpeng, Huang, Lanxiao, Lin, Xia, Zhou, Wengang, and Li, Houqiang
- Subjects
Computer Science - Artificial Intelligence ,Computer Science - Machine Learning ,68T05 ,I.2.6 ,I.2.11 - Abstract
Traditionally, AI development for two-player zero-sum games has relied on two primary techniques: decision trees and reinforcement learning (RL). A common approach involves using a fixed decision tree as one player's strategy while training an RL agent as the opponent to identify vulnerabilities in the decision tree, thereby improving its strategic strength iteratively. However, this process often requires significant human intervention to refine the decision tree after identifying its weaknesses, resulting in inefficiencies and hindering full automation of the strategy enhancement process. Fortunately, the advent of Large Language Models (LLMs) offers a transformative opportunity to automate the process. We propose RL-LLM-DT, an automatic decision tree generation method based on RL Evaluation and LLM Enhancement. Given an initial decision tree, the method involves two important iterative steps. Response Policy Search: RL is used to discover counter-strategies targeting the decision tree. Policy Improvement: LLMs analyze failure scenarios and generate improved decision tree code. In our method, RL focuses on finding the decision tree's flaws while LLM is prompted to generate an improved version of the decision tree. The iterative refinement process terminates when RL can't find any flaw of the tree or LLM fails to improve the tree. To evaluate the effectiveness of this integrated approach, we conducted experiments in a curling game. After iterative refinements, our curling AI based on the decision tree ranks first on the Jidi platform among 34 curling AIs in total, which demonstrates that LLMs can significantly enhance the robustness and adaptability of decision trees, representing a substantial advancement in the field of Game AI. Our code is available at https://github.com/Linjunjie99/RL-LLM-DT., Comment: Length:10 pages. Figures:10 figures. Additional Notes:In this paper, we have introduced a novel hybrid approach which leverages the strengths of both RL and LLMs to itera- tively refine decision tree tactics, enhancing their performance and adaptability
- Published
- 2024
45. Explaining Model Overfitting in CNNs via GMM Clustering
- Author
-
Dou, Hui, Mu, Xinyu, Yi, Mengjun, Han, Feng, Zhao, Jian, and Shen, Furao
- Subjects
Computer Science - Machine Learning ,Computer Science - Artificial Intelligence ,Computer Science - Computer Vision and Pattern Recognition - Abstract
Convolutional Neural Networks (CNNs) have demonstrated remarkable prowess in the field of computer vision. However, their opaque decision-making processes pose significant challenges for practical applications. In this study, we provide quantitative metrics for assessing CNN filters by clustering the feature maps corresponding to individual filters in the model via Gaussian Mixture Model (GMM). By analyzing the clustering results, we screen out some anomaly filters associated with outlier samples. We further analyze the relationship between the anomaly filters and model overfitting, proposing three hypotheses. This method is universally applicable across diverse CNN architectures without modifications, as evidenced by its successful application to models like AlexNet and LeNet-5. We present three meticulously designed experiments demonstrating our hypotheses from the perspectives of model behavior, dataset characteristics, and filter impacts. Through this work, we offer a novel perspective for evaluating the CNN performance and gain new insights into the operational behavior of model overfitting.
- Published
- 2024
46. Integrating Dual Prototypes for Task-Wise Adaption in Pre-Trained Model-Based Class-Incremental Learning
- Author
-
Xu, Zhiming, Yang, Suorong, Xu, Baile, Zhao, Jian, and Shen, Furao
- Subjects
Computer Science - Machine Learning ,Statistics - Machine Learning - Abstract
Class-incremental learning (CIL) aims to acquire new classes while conserving historical knowledge incrementally. Despite existing pre-trained model (PTM) based methods performing excellently in CIL, it is better to fine-tune them on downstream incremental tasks with massive patterns unknown to PTMs. However, using task streams for fine-tuning could lead to catastrophic forgetting that will erase the knowledge in PTMs. This paper proposes the Dual Prototype network for Task-wise Adaption (DPTA) of PTM-based CIL. For each incremental learning task, a task-wise adapter module is built to fine-tune the PTM, where the center-adapt loss forces the representation to be more centrally clustered and class separable. The dual prototype network improves the prediction process by enabling test-time adapter selection, where the raw prototypes deduce several possible task indexes of test samples to select suitable adapter modules for PTM, and the augmented prototypes that could separate highly correlated classes are utilized to determine the final result. Experiments on several benchmark datasets demonstrate the state-of-the-art performance of DPTA. The code will be open-sourced after the paper is published., Comment: 8 pages,6 figures,2 tables
- Published
- 2024
47. UnitedVLN: Generalizable Gaussian Splatting for Continuous Vision-Language Navigation
- Author
-
Dai, Guangzhao, Zhao, Jian, Chen, Yuantao, Qin, Yusen, Zhao, Hao, Xie, Guosen, Yao, Yazhou, Shu, Xiangbo, and Li, Xuelong
- Subjects
Computer Science - Computer Vision and Pattern Recognition ,Computer Science - Artificial Intelligence - Abstract
Vision-and-Language Navigation (VLN), where an agent follows instructions to reach a target destination, has recently seen significant advancements. In contrast to navigation in discrete environments with predefined trajectories, VLN in Continuous Environments (VLN-CE) presents greater challenges, as the agent is free to navigate any unobstructed location and is more vulnerable to visual occlusions or blind spots. Recent approaches have attempted to address this by imagining future environments, either through predicted future visual images or semantic features, rather than relying solely on current observations. However, these RGB-based and feature-based methods lack intuitive appearance-level information or high-level semantic complexity crucial for effective navigation. To overcome these limitations, we introduce a novel, generalizable 3DGS-based pre-training paradigm, called UnitedVLN, which enables agents to better explore future environments by unitedly rendering high-fidelity 360 visual images and semantic features. UnitedVLN employs two key schemes: search-then-query sampling and separate-then-united rendering, which facilitate efficient exploitation of neural primitives, helping to integrate both appearance and semantic information for more robust navigation. Extensive experiments demonstrate that UnitedVLN outperforms state-of-the-art methods on existing VLN-CE benchmarks.
- Published
- 2024
48. Embedding Space Allocation with Angle-Norm Joint Classifiers for Few-Shot Class-Incremental Learning
- Author
-
Tu, Dunwei, Yi, Huiyu, Zhang, Tieyi, Li, Ruotong, Shen, Furao, and Zhao, Jian
- Subjects
Computer Science - Computer Vision and Pattern Recognition - Abstract
Few-shot class-incremental learning (FSCIL) aims to continually learn new classes from only a few samples without forgetting previous ones, requiring intelligent agents to adapt to dynamic environments. FSCIL combines the characteristics and challenges of class-incremental learning and few-shot learning: (i) Current classes occupy the entire feature space, which is detrimental to learning new classes. (ii) The small number of samples in incremental rounds is insufficient for fully training. In existing mainstream virtual class methods, for addressing the challenge (i), they attempt to use virtual classes as placeholders. However, new classes may not necessarily align with the virtual classes. For the challenge (ii), they replace trainable fully connected layers with Nearest Class Mean (NCM) classifiers based on cosine similarity, but NCM classifiers do not account for sample imbalance issues. To address these issues in previous methods, we propose the class-center guided embedding Space Allocation with Angle-Norm joint classifiers (SAAN) learning framework, which provides balanced space for all classes and leverages norm differences caused by sample imbalance to enhance classification criteria. Specifically, for challenge (i), SAAN divides the feature space into multiple subspaces and allocates a dedicated subspace for each session by guiding samples with the pre-set category centers. For challenge (ii), SAAN establishes a norm distribution for each class and generates angle-norm joint logits. Experiments demonstrate that SAAN can achieve state-of-the-art performance and it can be directly embedded into other SOTA methods as a plug-in, further enhancing their performance.
- Published
- 2024
49. Region-Guided Attack on the Segment Anything Model (SAM)
- Author
-
Liu, Xiaoliang, Shen, Furao, and Zhao, Jian
- Subjects
Computer Science - Computer Vision and Pattern Recognition ,Computer Science - Artificial Intelligence ,Computer Science - Cryptography and Security - Abstract
The Segment Anything Model (SAM) is a cornerstone of image segmentation, demonstrating exceptional performance across various applications, particularly in autonomous driving and medical imaging, where precise segmentation is crucial. However, SAM is vulnerable to adversarial attacks that can significantly impair its functionality through minor input perturbations. Traditional techniques, such as FGSM and PGD, are often ineffective in segmentation tasks due to their reliance on global perturbations that overlook spatial nuances. Recent methods like Attack-SAM-K and UAD have begun to address these challenges, but they frequently depend on external cues and do not fully leverage the structural interdependencies within segmentation processes. This limitation underscores the need for a novel adversarial strategy that exploits the unique characteristics of segmentation tasks. In response, we introduce the Region-Guided Attack (RGA), designed specifically for SAM. RGA utilizes a Region-Guided Map (RGM) to manipulate segmented regions, enabling targeted perturbations that fragment large segments and expand smaller ones, resulting in erroneous outputs from SAM. Our experiments demonstrate that RGA achieves high success rates in both white-box and black-box scenarios, emphasizing the need for robust defenses against such sophisticated attacks. RGA not only reveals SAM's vulnerabilities but also lays the groundwork for developing more resilient defenses against adversarial threats in image segmentation.
- Published
- 2024
50. Approximate attention with MLP: a pruning strategy for attention-based model in multivariate time series forecasting
- Author
-
Guo, Suhan, Deng, Jiahong, Wei, Yi, Dou, Hui, Shen, Furao, and Zhao, Jian
- Subjects
Computer Science - Machine Learning - Abstract
Attention-based architectures have become ubiquitous in time series forecasting tasks, including spatio-temporal (STF) and long-term time series forecasting (LTSF). Yet, our understanding of the reasons for their effectiveness remains limited. This work proposes a new way to understand self-attention networks: we have shown empirically that the entire attention mechanism in the encoder can be reduced to an MLP formed by feedforward, skip-connection, and layer normalization operations for temporal and/or spatial modeling in multivariate time series forecasting. Specifically, the Q, K, and V projection, the attention score calculation, the dot-product between the attention score and the V, and the final projection can be removed from the attention-based networks without significantly degrading the performance that the given network remains the top-tier compared to other SOTA methods. For spatio-temporal networks, the MLP-replace-attention network achieves a reduction in FLOPS of $62.579\%$ with a loss in performance less than $2.5\%$; for LTSF, a reduction in FLOPs of $42.233\%$ with a loss in performance less than $2\%$.
- Published
- 2024
Catalog
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.