18,211 results on '"CHENG Hao"'
Search Results
2. Detection of Brucella abortus using an electrochemical immunosensor modified with PB-MWCNTs-GNPs
- Author
-
CHI Yu, CAO Yu, and CHENG Hao
- Subjects
electrochemical immunosensor ,pb-mwcnts-gnps ,brucella abortus ,Medicine (General) ,R5-920 - Abstract
Objective To prepare a high performance electrochemical immunosensor for detecting Brucella abortus (B. abortus). Methods Prussian blue (PB), multi walled carbon nanotubes (MWCNTs) and gold nanoparticles (GNPs)(PB-MWCNTs-GNPs) nanocomposites were prepared, and appropriate antibody was used to construct the immunosensor for detecting B. abortus samples. The optimal conditions were clarified by examining the key factors in sensor construction, and then the performance of the sensor was evaluated. Results The optimal construction conditions were determined as follows: the ratio of MWCNTs-PB was 1 ∶5, the drying temperature was 37 ℃, the pH value of buffer system was 7.5, and the incubation time of antibody and sample was 1 h and 30 min, respectively. B. abortus exhibited a good linear relationship, when ranging from 10 to 1×105 CFU/mL. The sensor had good anti-interference ability, repeatability, stability and high accuracy. Conclusion Our prepared PB-MWCNTs GNPs nanomaterials modified electrochemical immunosensor for detecting B. abortus is easy to prepare, has good performance, and can provide reference for the early clinical diagnosis of brucellosis.
- Published
- 2024
- Full Text
- View/download PDF
3. Grain and chaff separation detection method based on machine vision
- Author
-
LI Xin, QI Jiamin, CHENG Hao, and WANG Yanchun
- Subjects
grain and chaff separation ,machine vision ,image processing ,feature extraction ,Food processing and manufacture ,TP368-456 - Abstract
[Objective] To solve the problem of poor manual detection accuracy of traditional grain and chaff separator and improve production efficiency. [Methods] An image detection method based on machine vision was proposed, which realized the feature recognition and separation of grain rough through multi-stage progressive fusion of different image algorithms. The acquired images were selected in the ROI region and enhanced by Retinex algorithm. The Otsu algorithm was used to segment the image, and then the median filtering wwas combined with morphology to remove the image noise. The improved Canny algorithm was used to detect edge features of binary images, and the position information of the contour of the valley rough image was extracted by combining the Hough transform. Finally, the state estimation of the position information was performed by using the Kalman filter, and the best predicted value of the separated position was obtained, while the position offset error was reduced. [Results] The average detection error of the system was 3.14 mm, a decrease of 1.82 mm compared to before filtering, and the average standard deviation of filtering error was 0.8 mm. [Conclusion] This method can effectively detect the grain rough feature information and improve the separation accuracy.
- Published
- 2024
- Full Text
- View/download PDF
4. Establishment of Typical Operating Conditions for Train Traction System Based on Short Trip Analysis Method
- Author
-
LU Xiulong, GUO Yidan, CHENG Hao, DENG Ming, and ZOU Xiaoyang
- Subjects
traction system ,energy efficiency ,typical operating conditions ,short trip analysis method ,nonparametric test ,Control engineering systems. Automatic machinery (General) ,TJ212-225 ,Technology - Abstract
This paper presents a method to establish typical operating conditions for train traction systems by combining site-based short trip analysis with nonparametric tests, in order to overcome various obstacles, such as inadequate efficiency testing results based on rated points of train traction systems, which often fail to accurately reflect the actual energy efficiency level of systems during long-distance operation under multiple conditions, and the existing absence of typical operating conditions essential for supporting the evaluation of traction systems' energy efficiency levels and guiding the energy-saving design of products. First, short trips were categorized, screened and randomly combined. Then, the optimal short-trip combination was identified through nonparametric tests. Subsequently, the typical operating conditions were developed, taking the traction system of an electric locomotive model as a case study. Furthermore, the proposed method was verified by analyzing errors in operating characteristic parameters and assessing the similarities in velocity-force (V-F) distribution probability two-dimensional matrix. The verification results revealed a deviation between the constructed typical operating conditions and the characteristic values derived from sampled data at 4.54%, and the consistency in V-F distribution probabilities at 98.06%. These findings indicate that the method proposed in this paper accurately reflects the actual operating characteristics of train traction systems, with the established typical operating conditions closely aligning with the actual operating conditions of train traction systems, underscoring their potential for aiding in energy-saving design and energy efficiency evaluation for train traction systems.
- Published
- 2024
- Full Text
- View/download PDF
5. Comparison of plan quality and robustness using VMAT and IMRT for breast cancer
- Author
-
Yin Chuou, Deng Juan, Mei Guojian, Cheng Hao, He Yingying, and Liu Jiang
- Subjects
robustness ,delivery time ,dosimetric metrics ,breast cancer ,volumetric modulated arc therapy ,halcyon ,Physics ,QC1-999 - Abstract
To evaluate the plan quality and robustness of volumetric modulated arc therapy (VMAT) and intensity modulated radiation therapy (IMRT) for breast cancer, 50 patients, including 25 patients who received radiotherapy after breast-conserving surgery (BCR) and 25 patients who received postmastectomy radiotherapy (PRT), were selected for this study. Nominal VMAT and IMRT plans were generated for each patient on Eclipse treatment planning system (version 15.6). The dosimetric metrics, dose distribution, gamma passing rate, and delivery time were compared. In addition, 12 uncertainty plans with plan isocenter uncertainty and CT density uncertainty were recalculated based on the nominal plans for each patient. The dose volume histogram (DVH) band width (DVHBW) was adopted to quantify the plan robustness of the nominal plans for the perturbed scenarios in this study. For BCR, the dosimetric metrics except planning target volume (PTV) conformal index (CI) and ipsilateral lung V 5 were not statistically different for IMRT and VMAT plans. PTV CI of VMAT plans was better than that of IMRT plans (VMAT: 0.923 ± 0.024, IMRT: 0.855 ± 0.032, p = 0.003). The ipsilateral lung V 5 of VMAT plan was higher than that of IMRT plan (VMAT: 42.4% ± 2.8%, IMRT: 40.5% ± 4.0%, p = 0.045). The VMAT plans save more than 1.20 min compared to the IMRT plans (VMAT: 0.87 min, IMRT: 2.08 min, p < 0.001). The gamma passing rates of VMAT plans were better than those of IMRT plans (3 mm/3%, VMAT: 99.7% ± 0.2%, IMRT: 99.4% ± 0.4%, p < 0.001; 2 mm/2%, VMAT: 97.2% ± 1.0%, IMRT: 96.9% ± 0.6%, p = 0.108). For PRT, the dosimetric metrics of VMAT plans, including PTV D mean, homogeneity index (HI), CI, and D max of spinal cord, were significantly better than those of IMRT plans. The VMAT plans save more than 45% time compared with IMRT plans (VMAT: 1.54 min, IMRT: 2.81 min, p < 0.001). The difference in gamma passing rates between VMAT plans and IMRT plans was not statistically significant. For the plan robustness, the DVHBW of VMAT plans and IMRT plans for BCR were 2.09% ± 0.23% and 2.98% ± 0.40%, respectively (p < 0.05). For PRT, the DVHBW of VMAT plans was significantly better than those of IMRT plans (VMAT: 3.05% ± 0.26%, IMRT: 3.57% ± 0.27%, p < 0.05). The results show that the dosimetric metrics of VMAT plans were comparable to those of IMRT plans. More importantly, the VMAT plans had excited dose distribution and fast execution efficiency. The plan robustness of VMAT plans were superior.
- Published
- 2024
- Full Text
- View/download PDF
6. Protective effect of TLR2/TLR9 agonists on pulmonary Acinetobacter baumannii infection in mice
- Author
-
CHENG Hao, YANG Yun, and SUN Hongwu
- Subjects
acinetobacter baumannii ,tlr2 ,tlr9 ,intranasal immunization ,pulmonary infection ,Medicine (General) ,R5-920 - Abstract
Objective To investigate the protective effect of Toll-like receptor (TLR) 2 /TLR9 agonists, Pam2CSK4 (Pam) and CpG ODN (CpG) on mice infected with Acinetobacter baumannii (Ab) in the lungs. Methods Female C57 mice (6~8 weeks old) were randomly divided into PBS, Pam, CpG and Pam+CpG groups. In 24 h after intranasal immunization with different doses of the corresponding agonists, the mice were given a lethal dose of Ab infection in the lungs, and the survival rates of the mice were observed. A sublethal dose lung infection model of Ab was then established, and the bacterial colonization in the blood, lungs, liver, kidneys and spleen was measured respectively in the mice after infection. HE staining was used to observe the pathological damages in the lungs and kidneys. The protective effect of the agonists in the immunized mice against Ab was examined at 1, 3 and 7 d after immunization to explore the protective time window. Pam+CpG was used to stimulate A549 cells and RAW264.7 cells to investigate the killing or phagocytic effects on Ab. Results Compared to PBS, Pam+CpG treatment significantly improved the survival rate of the mice after a lethal dose of Ab lung infection (P
- Published
- 2024
- Full Text
- View/download PDF
7. Uncertainty Analysis and Application of Temperature Rise Measurement for Traction Motor Winding in Rail Transit
- Author
-
DENG Min, LU Xiulong, CHENG Hao, CHEN Mingyang, ZOU Xiaoyang, and WU Shuangyi
- Subjects
traction motor ,temperature rise ,least squares method ,resistance method ,evaluation of uncertainty ,Control engineering systems. Automatic machinery (General) ,TJ212-225 ,Technology - Abstract
Temperature rise measurement is crucial for evaluating the performance of motors and ensuring their safe operation. In order to more effectively, accurately, and scientifically evaluate temperature rise measurement results for traction motors, the incorporation of the measurement uncertainty concept holds practical significance. This paper proposes a method for evaluating uncertainties in temperature rise measurement of traction motors based on technical specifications outlined in JJF 1059.1-2012 Evaluation and expression of uncertainty in measurement. The current study started with a comprehensive analysis of uncertainty sources in temperature rise measurement results of traction motors using the standard IEC 60349-2:2010 Electric traction-Rotating electrical machines for rail and road vehicles-Part 2: Electronic converter-fed alternating current motors. Subsequently, an uncertainty evaluation model was constructed, and practical examples were used to evaluate measurement uncertainties, specifically involving standard uncertainty evaluation and calculations of uncertainties in synthesis and extension. The results showed that, with a confidence probability of 95%, the motor temperature rise measurement result stood at 79.22 K, containing an extension uncertainty of 1.30 K, which quantitatively characterized the quality of the test results. The proposed method provides a technical basis for determining the confidence level of temperature rise measurement results for rolling stock traction motors, while offering guidance for evaluating uncertainties in temperature rise measurement for motor winding, especially concerning curve fitting using the least squares method.
- Published
- 2024
- Full Text
- View/download PDF
8. A Study of Color Fixation Agents in Secondary Alcohol Ethoxylate-Based Reverse Micellar Cotton Dyeing System with Reactive Dyes
- Author
-
Yiu Lun Alan Tang, Ho Shing Law, Shixin Jin, Cheng Hao Lee, Jiali Yu, Yanming Wang, and Chi-Wai Kan
- Subjects
Cotton ,reactive dye ,alkali ,secondary alcohol ethoxylate ,reverse micelle ,dyeing ,Science ,Textile bleaching, dyeing, printing, etc. ,TP890-933 - Abstract
Influence of color fixation agent (CFA) and dyebath pH in secondary alcohol ethoxylate (SAE)-based reverse micellar dyeing system of cotton with reactive dyes was investigated and compared with water-based dyeing system using different alkali (i.e. CFA) such as: (i) sodium bicarbonate (NaHCO3), (ii) sodium carbonate (Na2CO3) and (iii) sodium hydroxide (NaOH). The color, tensile strength, fastness and surface morphological properties of dyed samples were examined. Experimental results showed that samples dyed with Na2CO3 can achieve the highest color yield, followed by NaHCO3 and NaOH in both water-based and reverse micellar dyeing system. The color yield and reflectance percentage of the dyed cotton samples were found to be closely related to the dyebath pH value. Relative unlevelness indices (RUI) also reflected that cotton samples dyed with NaOH in reverse micellar dyeing system have a higher chance of color unlevelness when compared with NaHCO3 and Na2CO3. Tensile strength results affirm that higher alkalinity of dyebath could cause higher strength loss to the colored cotton samples. Both cotton samples dyed by water and SAE approach showed good to excellent color fastness properties while scanning electron microscopic (SEM) images exhibit that the use of different alkalis may cause some damage to the cotton fiber.
- Published
- 2024
- Full Text
- View/download PDF
9. Cerebral cortex functional reorganization in preschool children with congenital sensorineural hearing loss: a resting-state fMRI study
- Author
-
Yi Yin, Xinyue Lyu, Jian Zhou, Kunlin Yu, Mingming Huang, Guiquan Shen, Cheng Hao, Zhengfu Wang, Hui Yu, and Bo Gao
- Subjects
congenital sensorineural hearing loss ,cortex functional reorganization ,functional MRI ,functional connectivity ,brain networks ,Neurology. Diseases of the nervous system ,RC346-429 - Abstract
PurposeHow cortical functional reorganization occurs after hearing loss in preschool children with congenital sensorineural hearing loss (CSNHL) is poorly understood. Therefore, we used resting-state functional MRI (rs-fMRI) to explore the characteristics of cortical reorganization in these patents.MethodsSixty-three preschool children with CSNHL and 32 healthy controls (HCs) were recruited, and the Categories of Auditory Performance (CAP) scores were determined at the 6-month follow-up after cochlear implantation (CI). First, rs-fMRI data were preprocessed, and amplitude of low-frequency fluctuation (ALFF) and regional homogeneity (ReHo) were calculated. Second, whole-brain functional connectivity (FC) analysis was performed using bilateral primary auditory cortex as seed points. Finally, Spearman correlation analysis was performed between the differential ALFF, ReHo and FC values and the CAP score.ResultsALFF analysis showed that preschool children with CSNHL had lower ALFF values in the bilateral prefrontal cortex and superior temporal gyrus than HCs, but higher ALFF values in the bilateral thalamus and calcarine gyrus. And correlation analysis showed that some abnormal brain regions were weak negatively correlated with CAP score (p
- Published
- 2024
- Full Text
- View/download PDF
10. Comparison of the main components, volatile components, and antioxidant activity of two types of fermented bamboo shoots
- Author
-
LIAO An, DU Haiping, CHENG Hao, LIU Tianyi, and TIAN Yan
- Subjects
fermented bamboo shoots ,volatile components ,antioxidant activity ,Food processing and manufacture ,TP368-456 - Abstract
Objective: The primary objective of this study is to investigate the chemical composition, volatile components, and biological activity of pickled bamboo shoots in the Guangxi and Guangdong regions, as well as the regional characteristics and flavor formation mechanisms of these pickled bamboo shoots. Methods: Physical and chemical experiments were used to analyze the major components in fermented bamboo shoots and their extracts. The volatile flavor components in fermented bamboo shoots were identified and analyzed using headspace solid-phase microextraction gas chromatography-mass spectrometry (HS-SPME-GC-MS). The antioxidant activity of fermented bamboo shoot extracts was evaluated using in vitro antioxidant activity methods (ABTS free radical scavenging method, DPPH free radical scavenging method, and FRAP method). Results: There were no significant differences in moisture, pH, total soluble sugars, fats, amino acid nitrogen, and nitrite content between the pickled bamboo shoots from Guangxi Liuzhou Tianyi Agricultural Technology Co Ltd (hereinafter referred to as GXLZTY) and those from the Shaoguan Nanxiong Agricultural Trade Market in Northern Guangdong (hereinafter referred to as GDYBNX). However, there were significant differences between the two in terms of chloride, total acids, proteins, dietary fibers, total flavonoids, total polyphenols, and mineral element content. Analysis of volatile components indicated that the main constituents in GXLZTY fermented bamboo shoots were P-cresol, hexanoic acid, octanoic acid, 2-ethylhexanol, and sulfur dioxide, while the primary components in fermented bamboo shoots from GDYBNX were p-cresol, acetic acid, and serinol. Antioxidant activity test results showed that the GXLZTY sour bamboo shoot extracts were superior in terms of clearing ABTS free radicals, DPPH free radicals, and FRAP activity compared to the extracts from GDYBNX. Conclusion: GXLZTY fermented bamboo shoots have higher nutritional value and antioxidant activity. These research results provide a theoretical basis for the fermentation process and industrial production of fermented bamboo shoots.
- Published
- 2024
- Full Text
- View/download PDF
11. Risk Factors Analysis and Risk Prediction Model Establishment for Rebleeding Events within 3 Years after Endoscopic Treatment of Gastroesophageal Varices in Liver Cirrhosis Patients
- Author
-
CHENG Hao, ZHOU Jinchi, LIU Xi, KANG Lin, FAN Ahui, DOU Weijia, LIU Zhenxiong
- Subjects
liver cirrhosis ,esophageal and gastric varices ,endoscopy ,rebleeding ,risk factors ,proportional hazards models ,Medicine - Abstract
Background Patients with liver cirrhosis complicated by gastroesophageal variceal and rupture hemorrhage have a certain probability of rebleeding events after endoscopic treatment, and the bleeding volume of rebleeding events is greater with higher risk, which seriously affects the survival rates of patients. Objective To investigate the independent risk factors of rebleeding events within 3 years after endoscopic treatment in cirrhotic patients with gastroesophageal variceal hemorrhage, construct a nomogram risk prediction model and validate it internally. Methods Four hundred and three patients who underwent endoscopic treatment for liver cirrhosis associated gastroesophageal varices at the Tangdu Hospital and Xijing Hospital, Air Force Medical University from 2011-2022 were retrospectively collected and divided into the rebleeding group (n=252) and control group (n=151) based on the presence of rebleeding within 3 years. The general data and auxiliary examination results of the patients between both groups were compared, and the statistically significant factors were included in the multivariate Logistic regression analysis to analyze the independent risk factors. These data were then input into the R language software to construct a nomogram risk prediction model by using a specific program package. Results Multivariate Logistic regression analysis showed that smoking〔OR=2.499, 95%CI (1.232, 5.066), P=0.011〕, portal vein internal diameter〔OR=1.047, 95%CI (1.028, 1.066), P
- Published
- 2023
- Full Text
- View/download PDF
12. Terahertz spin-to-charge conversion in ferromagnetic Ni nanofilms
- Author
-
Cheng Hao, Wang Yangkai, Liu Zheng, Jia Xiangyu, Huang Qiuping, and Lu Yalin
- Subjects
ferromagnetic films ,inverse spin hall effect ,spin-to-charge conversion ,spintronic terahertz emitters ,Physics ,QC1-999 - Abstract
Spintronic terahertz (THz) emission via spin-to-charge conversion (SCC) has been widely studied in ferromagnets (FM)/nonmagnets (NM) structures, in which various mechanisms of SCC have been confirmed in different NM materials. However, it is rare to find a material having multiple SCC mechanisms at the same time. Here, we report a ferromagnetic metal Ni film with diverse functions in the SCC process, by performing THz emission experiments in single Ni layer, FM/Ni, Ni/NM bilayers and FM/Ni/NM trilayers. It is demonstrated that in Ni monolayer, THz emission is radiated by the anomalous Hall effect and ultrafast demagnetization of Ni film. In FM/Ni, the Ni film acts as an SCC implementer and THz emission is mainly generated by the inverse spin Hall effect (ISHE) of Ni. In Ni/NM, the Ni film acts as a spin injector and provides spin currents to be converted to charge current via ISHE of heavy metal NM, inducing THz emission. In FM/Ni/NM, THz emission mainly comes from ISHE of FM/Ni, Ni/NM, and FM/NM, and their domination is relative to Ni thickness. Our findings show a ferromagnetic film not only acts as a spin injector but also as an SCC implementer, providing a new concept to design spintronic THz emitters.
- Published
- 2023
- Full Text
- View/download PDF
13. Load forecasting method based on CEEMDAN and TCN-LSTM.
- Author
-
Luo Heng, Cheng Hao, and Liu Chen Nan
- Subjects
Medicine ,Science - Abstract
Aiming at the problems of high stochasticity and volatility of power loads as well as the difficulty of accurate load forecasting, this paper proposes a power load forecasting method based on CEEMDAN (Completely Integrated Empirical Modal Decomposition) and TCN-LSTM (Temporal Convolutional Networks and Long-Short-Term Memory Networks). The method combines the decomposition of raw load data by CEEMDAN and the spatio-temporal modeling capability of TCN-LSTM model, aiming to improve the accuracy and stability of forecasting. First, the raw load data are decomposed into multiple linearly stable subsequences by CEEMDAN, and then the sample entropy is introduced to reorganize each subsequence. Then the reorganized sequences are used as inputs to the TCN-LSTM model to extract sequence features and perform training and prediction. The modeling prediction is carried out by selecting the electricity compliance data of New South Wales, Australia, and compared with the traditional prediction methods. The experimental results show that the algorithm proposed in this paper has higher accuracy and better prediction effect on load forecasting, which can provide a partial reference for electricity load forecasting methods.
- Published
- 2024
- Full Text
- View/download PDF
14. Effect of heat input on microstructure and mechanical properties of laser welded joint of Inconel 617 nickel-based superalloy
- Author
-
CHENG Hao, ZHOU Liangang, LIU Jian, WANG Yuning, TONG Lingyun, and DU Dong
- Subjects
nickel-based superalloy ,laser welding ,heat input ,microstructure ,mechanical property ,Materials of engineering and construction. Mechanics of materials ,TA401-492 - Abstract
Inconel 617 nickel-based superalloy with 3 mm wall thickness was welded by laser beam using two different heat input welding parameters. The microstructure of the welded joint was observed by optical microscope and scanning electron microscope, and the mechanical properties of the welded joint at room temperature (25℃) and high temperature (900℃) were tested. The results show that the laser welding heat input has a significant effect on the microstructure and mechanical properties of Inconel 617 welded joints. The front width of the laser weld obtained by high heat input (200 J/mm) process parameters is 3.88 mm. The grain size in the middle of the weld fusion zone is coarse, and the grain orientation is disordered. The secondary dendrite arm spacing in the middle of the weld is large (6.71 μm). The carbide particle size between the dendrites is coarse, and the solidification microsegregation of Mo, Cr alloy elements is serious. The width of heat affected zone is about 0.29 mm. The eutectic structure of γ+carbide is formed in the grain boundary and grain interior. This is because during the heating process of the welding, the spherical carbide particles and the surrounding austenite in the heat-affected zone liquefy and the eutectic structure is formed during the solidification process after welding. The front width of the laser weld obtained by low heat input (90 J/mm) process parameters is 2.28 mm, the grains inside the weld are columnar which is formed by epitaxial growth along the fusion line and directional solidification along the heat flow direction. The secondary dendrite arm spacing in the middle of the weld is small (2.26 μm), the carbide particles between dendrites are small, and the width of heat affected zone is about 0.15 mm. The tensile strength test at room temperature (25℃) shows that the welded joints obtained under high heat input fracture from the middle of the weld, and the tensile strength and elongation decrease, which is caused by the segregation of solid solution elements in the weld. The welded joints obtained under low heat input fracture from the base metal. At high temperature (900℃), all samples fracture from the base metal, which is due to the weakening of the grain boundary of the base metal at high temperatures.
- Published
- 2023
- Full Text
- View/download PDF
15. Current status and prospect for development of COVID-19 vaccine adjuvants
- Author
-
LI Haibo and CHENG Hao
- Subjects
sars-cov-2 ,vaccine ,adjuvant ,current status ,prospect ,Medicine (General) ,R5-920 - Abstract
Vaccines are the most economical and effective means to protect against SARS-CoV-2 infection. Adjuvants are able to enhance antigen-specific immune responses, reduce antigen doses and vaccination times, and reshape adaptive immune responses, which are crucial to improving the protective efficacy of vaccines. In this article, we review the mechanisms and advantages/disadvantages of the adjuvants used in COVID-19 vaccines that have been authorized for emergency use or are undergoing clinical trials around the world, and analyze the issues that must be considered in the application of adjuvants to COVID-19 vaccines. In addition, we put forward suggestions for future research strategies for COVID-19 vaccine adjuvants: ① adjuvants for immunocompromised people, ② adjuvants for inducing T cell responses, ③ novel mucosal immune adjuvants, and ④ the rational design of combination adjuvant.
- Published
- 2022
- Full Text
- View/download PDF
16. Fusion of Visual and Audio Signals for Wildlife Surveillance
- Author
-
Cheng Hao Ng, Tee Connie, Kan Yeep Choo, and Michael Kah Ong Goh
- Subjects
deep learning ,fusion ,machine learning ,wildlife surveillance ,Technology ,Technology (General) ,T1-995 - Abstract
Wildlife-vehicle collision (WVC) has been a significant threat to endangered species in Malaysia. Due to excessive development, tropical rainforests and their inhabitants have been edged towards extinction. Road buildings and other linear infrastructures, for instance, have caused forest destruction and forced wild animals to come out from their natural habitats to compete for resources with the human-beings. In Malaysia, much precious wildlife have been lost due to road accidents. Road signs and warning lights have been set up near wildlife crossing, but these do not help much. In this paper, we aim to propose a wildlife surveillance mechanism to detect the existence of wildlife near roadways using visual and audio input. Machine learning classifiers, including Convolution Neural Network (CNN), Support Vector Machine (SVM), K-nearest neighbors (KNN), and Naive Bayes, are adopted in the study. We focus on five types of the most frequently occurring wildlife on the roads: elephants, tapirs, Malayan bears, tigers, and wild boars. Experimental results demonstrate that a good accuracy as high as 99% can be achieved using the proposed approach. On the other hand, the Naïve Bayes classifier ranks the lowest in performance with an accuracy value only up to 86%.
- Published
- 2022
- Full Text
- View/download PDF
17. Antimicrobial potential of Chlorella sorokiniana on MRSA – An in vitro study and an in silico analysis on ClpP protease
- Author
-
Charmaine Lloyd, Malcolm Wai Kit Wong, Li Jiao Sin, Punitha Pandurangan Manickavasagam, Shoba Gunasekaran, Sim Ray Yue, Felicia Min En Goh, Rhea Thulasi Manoharan, Hao Yuin Kong, Jayme Zhen Yi Ang, Hui Ping Kang, Cheng Hao Tan, Ernest Jun Ming Teo, Xiu Qun Cui, Saraniya Subramaniam, Jasmine Hui Min Low, Chloe Jia Ye Oon, Isaac Pang Yi Khor, Grace Zhi Qi Lim, Nur Carmellia Bte Mia Kiong, Jeanette Teo, Jen Yan New, and A.S. Smiline Girija
- Subjects
Chlorella sorokiniana ,Microalgae ,MRSA ,Staphylococci ,ClpP1 ,Zone of inhibition of growth ,Science (General) ,Q1-390 - Abstract
Objective: Methicillin-resistant Staphylococcus aureus (MRSA) strains are a leading cause of communicable disease in community and nosocomial settings. They are responsible for high morbidity and mortality. Researchers currently pursue novel antimicrobials from natural sources against non-traditional drug targets of staphylococci to ensure a pipeline of potent drugs, in the face of rising drug resistance. The focus of this study was to screen compounds from a freshwater isolate of Chlorella sorokiniana for anti-staphylococcal activity, using traditional microbiology, phytochemical analysis and bioinformatics approaches. Methods: Chlorella sorokiniana methanol extract was investigated for its antimicrobial potential on Staphylococcus aureus strains (ATCC and MRSA isolates) by Kirby Bauer disc diffusion, broth microdilution, cell cytotoxicity and thin layer chromatography-bioautography (TLC-BA). Two antimicrobial TLC-BA antimicrobial fractions (A and B) were subject to gas chromatography mass spectrometry (GCMS). The structures of 9 compounds representing GCMS peaks were tested in silico, for their pharmacokinetic properties and binding energy efficiency with the target, using Molinspiration tool and Autodock 4.2. Results: Mean zone diameter of inhibition of growth by CSME (20 mg) was 21 mm, MIC/MBC was 0.31/2.5 mg/L. GCMS analysis of TLC fraction-A revealed 31 phytochemicals, of which 2-pentanone,4-hydroxy-4-methyl- had the highest area % (65.61) and TLC fraction-B revealed 4 peaks of which pentadecanoic acid and 1-(+)-ascorbic acid 2,6-dihexadecanoate had the highest area % (45.57, 48.09).In silico analysis of 9 peak compounds on the target of interest showed that compound 2: 2-pentanone,4-hydroxy-4-methyl- and compound 7: 1,2 – benzene dicarboxylic acid, mono (2- ethylhexyl) ester, satisfied Lipinski’s rule of 5, and displayed the least binding energies −6.93 and −5.74 with ClpP protease, thus holding pharmaceutical potential, and supporting further investment into in vitro and in vivo studies. Conclusions: C. sorokiniana, a less studied microalga thus offers a promising natural resource for anti-MRSA phytochemicals, capable of targeting ClpP1 protease.(290 words)
- Published
- 2023
- Full Text
- View/download PDF
18. Immunotherapeutic potential of blinatumomab-secreting γ9δ2 T Cells
- Author
-
Shang-Ju Wu, Chien-Ting Lin, Cheng Hao Liao, and Chun-Ming Lin
- Subjects
γ9δ2 T cells ,Bispecific antibody ,Blinatumomab ,CD19-targeting ,Neoplasms. Tumors. Oncology. Including cancer and carcinogens ,RC254-282 - Abstract
Previous studies have explored the use of engineered blinatumomab-secreting autologous αβ T cells for CD19-targeted cancer therapy. To create a more flexible allogeneic delivery system, we utilized γ9δ2 T cells rather than αβ T cells in a similar application. First, we showed that γ9δ2 T cells could serve as effector cells for blinatumomab, and these effector memory cells could survive for at least 7 days after infusion. The genetically modified blinatumomab-secreting γ9δ2 T cells induced significant cytotoxicity in CD19+ tumor cell lines and primary cells from chronic lymphocytic leukemia patients. Of note, blinatumomab-secreting γ9δ2 T cells might also exhibit dual-targeting of CD19 and isopentenyl pyrophosphate, a universal tumor-associated antigen. Furthermore, blinatumomab-secreting γ9δ2 T cells killed CD19-transfected adherent cells, suggesting that the γ9δ2 T cells might be effective for treating solid tumors with appropriate cancer antigens. Together, these results demonstrate the promise of blinatumomab-secreting γ9δ2 T cells as a cancer therapy.
- Published
- 2023
- Full Text
- View/download PDF
19. Polyethylene Glycol (PEG) Non-Ionic Surfactant-Based Reverse Micellar Dyeing of Cotton Fabric with Hot Type Trichloropyrimidine (TCP)-Based Reactive Dyes
- Author
-
Alan Yiu Lun Tang, Cheng Hao Lee, Yanming Wang, and Chi-Wai Kan
- Subjects
cotton fabric ,reactive dyes ,hot type ,reverse micelle ,non-aqueous dyeing ,polyethylene glycol ,Science ,Textile bleaching, dyeing, printing, etc. ,TP890-933 - Abstract
This work aims to examine the feasibility of using poly(ethylene glycol) (PEG)-based reverse micellar dyeing system for dyeing cotton fabric with the use of hot type trichloropyrimidine (TCP)-based reactive dyes and to evaluate the possibility of saving the dyeing energy used during the dyeing process. Experimental results show that fabrics dyed in reverse micellar system at 90°C can achieve higher color yield than the conventional water-based system at the same temperature. Further experiments using reverse micellar system for dyeing cotton at 80°C (a reduction of working temperature of 10°C) provided color yield comparable to that of conventional water-based system except the use of yellow dye. Both water-dyed and octane-dyed samples had good to excellent levelness, washing and rubbing fastness and 98.5% of octane could be recycled after distillation. Reflectance curves were identical in shape and the SEM images showed neither of the dyeing systems caused any significant damage to cotton fibers. These findings validated the possibility of using reverse micellar dyeing system for hot type TCP-based reactive dyeing of cotton fabrics at lower coloration temperature and energy consumption.
- Published
- 2023
- Full Text
- View/download PDF
20. Alkyl Polyglucoside (APG) Nonionic Surfactant-Based Reverse Micellar Dyeing of Cotton Fabric – A Study of Reactive Dyes with Different Functional Groups
- Author
-
Yiu Lun Alan Tang, Cheng Hao Lee, Chiu Yuk Chan, Yanming Wang, and Chi-Wai Kan
- Subjects
cotton fabric ,reactive dyes ,functional groups ,reverse micelle ,non-aqueous dyeing ,alkyl polyglucoside ,decamethylcyclopentasiloxane ,Science ,Textile bleaching, dyeing, printing, etc. ,TP890-933 - Abstract
Reverse micellar dyeing of cotton woven fabrics with natural and biodegradable alkyl polyglucoside-based (APG-based) nonionic surfactant in decamethylcyclopentasiloxane (D5) non-aqueous medium with reactive dyes of different functional groups was investigated and compared with conventional aqueous water-based dyeing in terms of reflectance, color yield, levelness, CIE L*a*b* values, washing fastness and color fading properties. Experimental results have revealed that APG reverse micellar method can provide a higher color yield (K/Ssum value) and lower color reflectance (better dye uptake) than conventional water-based method, along with comparable washing fastness, leveling and color fading properties. Hetero-bifunctional reactive dyes can attain the highest color yield, followed by homo-bifunctional reactive dyes and mono-functional reactive dyes. Mono-functional reactive dyes and homo-bifunctional reactive dyes produce the best levelling properties with smallest levelness variation when compared with that of hetero-bifunctional reactive dyes in non-aqueous D5 dyeing medium.
- Published
- 2023
- Full Text
- View/download PDF
21. Reverse Micellar Dyeing of Cotton Fabric with Reactive Dye Using Biodegradable Non-Ionic Surfactant as Nanoscale Carrier: An Optimisation Study by One-Factor-at-One-Time Approach
- Author
-
Yiu Lun Alan Tang, Shixin Jin, Cheng Hao Lee, Ho Shing Law, Jiali Yu, Yanming Wang, and Chi-wai Kan
- Subjects
Tergitol surfactant ,secondary alcohol ethoxylate ,cotton fabric ,reactive dyes ,non-aqueous dyeing ,reverse micelle ,Organic chemistry ,QD241-441 - Abstract
This study investigates the feasibility of using biodegradable secondary alcohol ethoxylate (SAE) non-ionic surfactant as a building block for the formation of reverse micelles, functioning as reactive dye carriers for the dyeing of cotton fabric in non-aqueous octane medium. Ten dyeing parameters were optimised, by a one-factor-at-a-time approach, namely: (i) effect of colour fixation agent; (ii) surfactant-to-water mole ratio; (iii) surfactant-to-co-surfactant mole ratio; (iv) volume of soda ash; (v) volume of dye; (vi) solvent-to-cotton ratio; (vii) dyeing temperature; (viii) dyeing time; (ix) fixation time; (x) soda-ash-to-cotton ratio. The colour properties, fastness properties and physical properties of SAE-dyed samples were experimentally compared with the conventional water-dyed samples. The optimised condition was found when SAE samples were dyed as follows: (a) 1:20 surfactant-to-water ratio; (b) 1:8 surfactant-to-co-surfactant ratio; (c) 10:1 solvent ratio; (d) 40 min dyeing time; (e) 60 min fixation time; and (f) 70 °C dyeing and fixation temperature. The results showed that SAE-dyed samples have better colour strength, lower reflectance percentage and comparable levelness, fastness and physical properties than that of water-dyed samples. SEM images revealed that the dyed cotton fibres had no severe surface damage caused by an SAE-based reverse micellar dyeing system. The TEM image depicts that the reverse micelle was of nanoscale, spherical-shaped and had a core–shell structure, validating the presence of reverse micelle as a reactive dye carrier and the potential of an SAE-based reverse micellar system for dyeing of cotton fabrics.
- Published
- 2023
- Full Text
- View/download PDF
22. Study on stability of underlying room and pillar old goaf in close coal seam and mining of the upper coal seam
- Author
-
Hongtao Liu, Cheng Hao, Zhiwen Wang, Chong Li, Linfeng Guo, Jialu Liang, and Haozhu Wang
- Subjects
close coal seams ,upper coal seam mining ,stability of old goaf ,stress distribution ,evolution of plastic zone ,Science - Abstract
Possible issues during mining of the upper coal seam in old goaf of nearby coal seams, including step subsidence, gas overflow in goaf, and roadway around rock fragmentation. Using the Hanjiawa Coal Mine’s upper coal seam mining, which takes place 28 m above the working face of the lower coal seam, as the research’s focal point. The paper focuses on the self-stability of the coal pillar in the old goaf, the failure form of the upper coal seam mining floor, the roof caving rule of the old goaf in the lower coal seam mining of the upper coal seam, and the bearing capacity of the interlayer rock strata using the pillar goaf stability evaluation system, field geological borehole electrical logging and borehole peeping, finite element difference numerical calculation, and other methods. The conclusion that the old goaf’s coal pillar can be completely stable and that the interlayer rock strata can bear the stress of upper coal seam mining is reached. The results show that the failure depth of the coal pillar in the lower coal seam old goaf is 1–3 m, the maximum failure depth accounting for 15% of the width of the coal pillar, and the failure depth of the roof in the old goaf is 0–3 m; After the mining of the upper coal seam, the floor above the coal pillar of the lower coal seam is plastic failure, and the failure depth is 1–10 m, and the failure depth of the roof of the old goaf of the lower coal seam is 3–15 m, which is 4 times greater than that before the mining. The maximum failure depth of the interlayer rock strata is 22 m, accounting for 78.6% of the rock strata spacing. The interlayer rock strata can bear the mining disturbance of the upper coal seam. The plastic zone of the floor of the upper coal seam is not connected with the plastic zone of the roof of the lower coal seam.
- Published
- 2023
- Full Text
- View/download PDF
23. Impact of human capital and social capital on employability of Chinese college students under COVID-19 epidemic—Joint moderating effects of perception reduction of employment opportunities and future career clarity
- Author
-
Yang Shiyuan, Yang Jinxiu, Xu Jingfei, Zhao Yuling, Yue Longhua, Li Houjian, Li Wei, Cheng Hao, He Guorong, and Chen Juan
- Subjects
human capital ,social capital ,employability ,perception reduction of employment opportunities ,future career clarity ,COVID-19 epidemic ,Psychology ,BF1-990 - Abstract
This research constructed a relationship model between human capital, social capital, and the employability of college students. With two moderating variables introduced, the perception reduction of employment opportunities under the COVID-19 epidemic and future career clarity, this research studied the direct impact of human capital and social capital on the employability of college students and boundary conditions. Research data from 810 employed Chinese college graduates shows that both human capital and social capital have a positive and significant impact on the employability; the perception reduction of employment opportunities under the COVID-19 epidemic negatively regulates the relationship between human capital and the employability of college students; the future career clarity positively regulates the relationship between human capital and the employability of college students; the perception reduction of employment opportunities under COVID-19 epidemic and the future career clarity jointly regulate the relationship between human capital, social capital and the employability of college students. These conclusions enrich the relevant theoretical and practical research on the employability of college students under the COVID-19 epidemic.
- Published
- 2022
- Full Text
- View/download PDF
24. Iterative Self-Tuning LLMs for Enhanced Jailbreaking Capabilities
- Author
-
Sun, Chung-En, Liu, Xiaodong, Yang, Weiwei, Weng, Tsui-Wei, Cheng, Hao, San, Aidan, Galley, Michel, and Gao, Jianfeng
- Subjects
Computer Science - Computation and Language ,Computer Science - Machine Learning - Abstract
Recent research has shown that Large Language Models (LLMs) are vulnerable to automated jailbreak attacks, where adversarial suffixes crafted by algorithms appended to harmful queries bypass safety alignment and trigger unintended responses. Current methods for generating these suffixes are computationally expensive and have low Attack Success Rates (ASR), especially against well-aligned models like Llama2 and Llama3. To overcome these limitations, we introduce ADV-LLM, an iterative self-tuning process that crafts adversarial LLMs with enhanced jailbreak ability. Our framework significantly reduces the computational cost of generating adversarial suffixes while achieving nearly 100\% ASR on various open-source LLMs. Moreover, it exhibits strong attack transferability to closed-source models, achieving 99% ASR on GPT-3.5 and 49% ASR on GPT-4, despite being optimized solely on Llama3. Beyond improving jailbreak ability, ADV-LLM provides valuable insights for future safety alignment research through its ability to generate large datasets for studying LLM safety. Our code is available at: https://github.com/SunChungEn/ADV-LLM, Comment: 18 pages
- Published
- 2024
25. VisionCoder: Empowering Multi-Agent Auto-Programming for Image Processing with Hybrid LLMs
- Author
-
Zhao, Zixiao, Sun, Jing, Wei, Zhiyuan, Cai, Cheng-Hao, Hou, Zhe, and Dong, Jin Song
- Subjects
Computer Science - Software Engineering ,Computer Science - Computer Vision and Pattern Recognition ,Computer Science - Multiagent Systems - Abstract
In the field of automated programming, large language models (LLMs) have demonstrated foundational generative capabilities when given detailed task descriptions. However, their current functionalities are primarily limited to function-level development, restricting their effectiveness in complex project environments and specific application scenarios, such as complicated image-processing tasks. This paper presents a multi-agent framework that utilises a hybrid set of LLMs, including GPT-4o and locally deployed open-source models, which collaboratively complete auto-programming tasks. Each agent plays a distinct role in the software development cycle, collectively forming a virtual organisation that works together to produce software products. By establishing a tree-structured thought distribution and development mechanism across project, module, and function levels, this framework offers a cost-effective and efficient solution for code generation. We evaluated our approach using benchmark datasets, and the experimental results demonstrate that VisionCoder significantly outperforms existing methods in image processing auto-programming tasks.
- Published
- 2024
26. ReasonAgain: Using Extractable Symbolic Programs to Evaluate Mathematical Reasoning
- Author
-
Yu, Xiaodong, Zhou, Ben, Cheng, Hao, and Roth, Dan
- Subjects
Computer Science - Artificial Intelligence - Abstract
Existing math datasets evaluate the reasoning abilities of large language models (LLMs) by either using the final answer or the intermediate reasoning steps derived from static examples. However, the former approach fails to surface model's uses of shortcuts and wrong reasoning while the later poses challenges in accommodating alternative solutions. In this work, we seek to use symbolic programs as a means for automated evaluation if a model can consistently produce correct final answers across various inputs to the program. We begin by extracting programs for popular math datasets (GSM8K and MATH) using GPT4-o. For those executable programs verified using the original input-output pairs, they are found to encapsulate the proper reasoning required to solve the original text questions. We then prompt GPT4-o to generate new questions using alternative input-output pairs based the extracted program. We apply the resulting datasets to evaluate a collection of LLMs. In our experiments, we observe significant accuracy drops using our proposed evaluation compared with original static examples, suggesting the fragility of math reasoning in state-of-the-art LLMs.
- Published
- 2024
27. CorrectionLM: Self-Corrections with SLM for Dialogue State Tracking
- Author
-
Lee, Chia-Hsuan, Cheng, Hao, and Ostendorf, Mari
- Subjects
Computer Science - Computation and Language - Abstract
Large language models (LLMs) have demonstrated self-improvement capabilities via feedback and refinement, but current small language models (SLMs) have had limited success in this area. Existing correction approaches often rely on distilling knowledge from LLMs, which imposes significant computation demands. In this work, we introduce CORRECTIONLM, a novel correction framework that enables SLMs to self-correct using in-context exemplars without LLM involvement. Applied to two dialogue state tracking (DST) tasks in low-resource settings, CORRECTIONLM achieves results similar to a state-of-the-art LLM at a small fraction of the computation costs.
- Published
- 2024
28. Resolvability of classical-quantum channels
- Author
-
Hayashi, Masahito, Cheng, Hao-Chung, and Gao, Li
- Subjects
Quantum Physics ,Computer Science - Emerging Technologies ,Computer Science - Information Theory - Abstract
Channel resolvability concerns the minimum resolution for approximating the channel output. We study the resolvability of classical-quantum channels in two settings, for the channel output generated from the worst input, and form the fixed independent and identically distributed (i.i.d.) input. The direct part of the worst-input setting is derived from sequential hypothesis testing as it involves of non-i.i.d.~inputs. The strong converse of the worst-input setting is obtained via the connection to identification codes. For the fixed-input setting, while the direct part follows from the known quantum soft covering result, we exploit the recent alternative quantum Sanov theorem to solve the strong converse., Comment: 20 pages, 3 figures. Comments are welcome!
- Published
- 2024
29. Configurable Embodied Data Generation for Class-Agnostic RGB-D Video Segmentation
- Author
-
Opipari, Anthony, Krishnan, Aravindhan K, Gayaka, Shreekant, Sun, Min, Kuo, Cheng-Hao, Sen, Arnie, and Jenkins, Odest Chadwicke
- Subjects
Computer Science - Robotics ,Computer Science - Computer Vision and Pattern Recognition - Abstract
This paper presents a method for generating large-scale datasets to improve class-agnostic video segmentation across robots with different form factors. Specifically, we consider the question of whether video segmentation models trained on generic segmentation data could be more effective for particular robot platforms if robot embodiment is factored into the data generation process. To answer this question, a pipeline is formulated for using 3D reconstructions (e.g. from HM3DSem) to generate segmented videos that are configurable based on a robot's embodiment (e.g. sensor type, sensor placement, and illumination source). A resulting massive RGB-D video panoptic segmentation dataset (MVPd) is introduced for extensive benchmarking with foundation and video segmentation models, as well as to support embodiment-focused research in video segmentation. Our experimental findings demonstrate that using MVPd for finetuning can lead to performance improvements when transferring foundation models to certain robot embodiments, such as specific camera placements. These experiments also show that using 3D modalities (depth images and camera pose) can lead to improvements in video segmentation accuracy and consistency. The project webpage is available at https://topipari.com/projects/MVPd, Comment: Accepted in IEEE Robotics and Automation Letters October 2024
- Published
- 2024
30. Tunable Einstein-Bohr recoiling-slit gedankenexperiment at the quantum limit
- Author
-
Zhang, Yu-Chen, Cheng, Hao-Wen, Zengxu, Zhao-Qiu, Wu, Zhan, Lin, Rui, Duan, Yu-Cheng, Rui, Jun, Chen, Ming-Cheng, Lu, Chao-Yang, and Pan, Jian-Wei
- Subjects
Quantum Physics ,Physics - Atomic Physics ,Physics - Optics ,Physics - Popular Physics - Abstract
In 1927, during the fifth Solvay Conference, Einstein and Bohr described a double-slit interferometer with a "movable slit" that can detect the momentum recoil of one photon. Here, we report a faithful realization of the Einstein-Bohr interferometer using a single atom in an optical tweezer, cooled to the motional ground state in three dimensions. The single atom has an intrinsic momentum uncertainty comparable to a single photon, which serves as a movable slit obeying the minimum Heisenberg uncertainty principle. The atom's momentum wavefunction is dynamically tunable by the tweezer laser power, which enables observation of an interferometric visibility reduction at a shallower trap, demonstrating the quantum nature of this interferometer. We further identify classical noise due to atom heating and precession, illustrating a quantum-to-classical transition., Comment: 18 pages, 4 figures
- Published
- 2024
31. Enhancing Single Image to 3D Generation using Gaussian Splatting and Hybrid Diffusion Priors
- Author
-
Basak, Hritam, Tabatabaee, Hadi, Gayaka, Shreekant, Li, Ming-Feng, Yang, Xin, Kuo, Cheng-Hao, Sen, Arnie, Sun, Min, and Yin, Zhaozheng
- Subjects
Computer Science - Computer Vision and Pattern Recognition - Abstract
3D object generation from a single image involves estimating the full 3D geometry and texture of unseen views from an unposed RGB image captured in the wild. Accurately reconstructing an object's complete 3D structure and texture has numerous applications in real-world scenarios, including robotic manipulation, grasping, 3D scene understanding, and AR/VR. Recent advancements in 3D object generation have introduced techniques that reconstruct an object's 3D shape and texture by optimizing the efficient representation of Gaussian Splatting, guided by pre-trained 2D or 3D diffusion models. However, a notable disparity exists between the training datasets of these models, leading to distinct differences in their outputs. While 2D models generate highly detailed visuals, they lack cross-view consistency in geometry and texture. In contrast, 3D models ensure consistency across different views but often result in overly smooth textures. We propose bridging the gap between 2D and 3D diffusion models to address this limitation by integrating a two-stage frequency-based distillation loss with Gaussian Splatting. Specifically, we leverage geometric priors in the low-frequency spectrum from a 3D diffusion model to maintain consistent geometry and use a 2D diffusion model to refine the fidelity and texture in the high-frequency spectrum of the generated 3D structure, resulting in more detailed and fine-grained outcomes. Our approach enhances geometric consistency and visual quality, outperforming the current SOTA. Additionally, we demonstrate the easy adaptability of our method for efficient object pose estimation and tracking.
- Published
- 2024
32. Distributed Quantum Hypothesis Testing under Zero-rate Communication Constraints
- Author
-
Sreekumar, Sreejith, Hirche, Christoph, Cheng, Hao-Chung, and Berta, Mario
- Subjects
Quantum Physics ,Computer Science - Information Theory - Abstract
The trade-offs between error probabilities in quantum hypothesis testing are by now well-understood in the centralized setting, but much less is known for distributed settings. Here, we study a distributed binary hypothesis testing problem to infer a bipartite quantum state shared between two remote parties, where one of these parties communicates classical information to the tester at zero-rate (while the other party communicates classical or quantum information to the tester at zero-rate or higher). As our main contribution, we derive an efficiently computable single-letter formula for the Stein's exponent of this problem, when the state under the alternative is product. For the general case, we show that the Stein's exponent is given by a multi-letter expression involving max-min optimization of regularized measured relative entropy. While this becomes single-letter for the fully classical case, we further prove that this already does not happen in the same way for classical-quantum states in general. As a key tool for proving the converse direction of our results, we develop a quantum version of the blowing-up lemma which may be of independent interest.
- Published
- 2024
33. Exponents for Shared Randomness-Assisted Channel Simulation
- Author
-
Oufkir, Aadil, Cao, Michael X., Cheng, Hao-Chung, and Berta, Mario
- Subjects
Computer Science - Information Theory ,Quantum Physics - Abstract
We determine the exact error and strong converse exponents of shared randomness-assisted channel simulation in worst case total-variation distance. Namely, we find that these exponents can be written as simple optimizations over the R\'enyi channel mutual information. Strikingly, and in stark contrast to channel coding, there are no critical rates, allowing a tight characterization for arbitrary rates below and above the simulation capacity. We derive our results by asymptotically expanding the meta-converse for channel simulation [Cao {\it et al.}, IEEE Trans.~Inf.~Theory (2024)], which corresponds to non-signaling assisted codes. We prove this to be asymptotically tight by employing the approximation algorithms from [Berta {\it et al.}, Proc.~IEEE ISIT (2024)], which show how to round any non-signaling assisted strategy to a strategy that only uses shared randomness. Notably, this implies that any additional quantum entanglement-assistance does not change the error or the strong converse exponents., Comment: 27+6 pages
- Published
- 2024
34. VEC-Sim: A Simulation Platform for Evaluating Service Caching and Computation Offloading Policies in Vehicular Edge Networks
- Author
-
Wu, Fan, Xu, Xiaolong, Bilal, Muhammad, Wang, Xiangwei, Cheng, Hao, and Wu, Siyu
- Subjects
Computer Science - Networking and Internet Architecture - Abstract
Computer simulation platforms offer an alternative solution by emulating complex systems in a controlled manner. However, existing Edge Computing (EC) simulators, as well as general-purpose vehicular network simulators, are not tailored for VEC and lack dedicated support for modeling the distinct access pattern, entity mobility trajectory and other unique characteristics of VEC networks. To fill this gap, this paper proposes VEC-Sim, a versatile simulation platform for in-depth evaluation and analysis of various service caching and computation offloading policies in VEC networks. VEC-Sim incorporates realistic mechanisms to replicate real-world access patterns, including service feature vector, vehicle mobility modeling, evolving service popularity, new service upload and user preference shifts, etc. Moreover, its modular architecture and extensive Application Programming Interfaces (APIs) allow seamless integration of customized scheduling policies and user-defined metrics. A comprehensive evaluation of VEC-Sim's capabilities is undertaken in comparison to real-world ground truths. Results prove it to be accurate in reproducing classical scheduling algorithms and extremely effective in conducting case studies.
- Published
- 2024
35. The aspect of bipartite coherence in quantum discord to semi-device-independent nonlocality and its implication for quantum information processing
- Author
-
Jebarathinam, Chellasamy, Ku, Huan-Yu, Cheng, Hao-Chung, and Goan, Hsi-Sheng
- Subjects
Quantum Physics - Abstract
\textit{Quantum discord} can demonstrate \textit{quantum nonlocality} in the context of a \textit{semi-device-independent} Bell or steering scenario, i.e., by assuming only the Hilbert-space dimension. This work addresses which aspect of \textit{bipartite coherence} is essential to such semi-device-independent quantum information tasks going beyond standard Bell nonlocality or quantum steering. It has been shown that the \textit{global coherence} of a single system can be transformed into \textit{bipartite entanglement}. However, global coherence can also be present in quantum discord. At the same time, discord can display bipartite coherence locally, i.e., only in a subsystem or both subsystems. Thus, global coherence of bipartite separable states is defined here as a form of bipartite coherence that is not reducible to local coherence in any of the subsystems or both subsystems. To answer the above-mentioned question, we demonstrate that global coherence is necessary to demonstrate semi-device-independent nonlocality of quantum discord in Bell or steering scenarios. From this result, it follows that any \textit{local operations} of the form $\Phi_A \otimes \Phi_B$ that may create \textit{coherence locally} are \textit{free operations} in the resource theory of semi-device-independent nonlocality of discord. As a byproduct, we identify the precise quantum resource for the quantum communication task of \textit{remote state preparation} using two-qubit separable states., Comment: 12 pages, 4 figures
- Published
- 2024
36. LoRC: Low-Rank Compression for LLMs KV Cache with a Progressive Compression Strategy
- Author
-
Zhang, Rongzhi, Wang, Kuang, Liu, Liyuan, Wang, Shuohang, Cheng, Hao, Zhang, Chao, and Shen, Yelong
- Subjects
Computer Science - Machine Learning ,Computer Science - Artificial Intelligence ,Computer Science - Computation and Language ,I.2 - Abstract
The Key-Value (KV) cache is a crucial component in serving transformer-based autoregressive large language models (LLMs), enabling faster inference by storing previously computed KV vectors. However, its memory consumption scales linearly with sequence length and batch size, posing a significant bottleneck in LLM deployment. Existing approaches to mitigate this issue include: (1) efficient attention variants integrated in upcycling stages, which requires extensive parameter tuning thus unsuitable for pre-trained LLMs; (2) KV cache compression at test time, primarily through token eviction policies, which often overlook inter-layer dependencies and can be task-specific. This paper introduces an orthogonal approach to KV cache compression. We propose a low-rank approximation of KV weight matrices, allowing for plug-in integration with existing transformer-based LLMs without model retraining. To effectively compress KV cache at the weight level, we adjust for layerwise sensitivity and introduce a progressive compression strategy, which is supported by our theoretical analysis on how compression errors accumulate in deep networks. Our method is designed to function without model tuning in upcycling stages or task-specific profiling in test stages. Extensive experiments with LLaMA models ranging from 8B to 70B parameters across various tasks show that our approach significantly reduces the GPU memory footprint while maintaining performance., Comment: 15 pages, 4 figures
- Published
- 2024
37. Spiking Neural Network as Adaptive Event Stream Slicer
- Author
-
Cao, Jiahang, Sun, Mingyuan, Wang, Ziqing, Cheng, Hao, Zhang, Qiang, Zhou, Shibo, and Xu, Renjing
- Subjects
Computer Science - Computer Vision and Pattern Recognition ,Computer Science - Neural and Evolutionary Computing - Abstract
Event-based cameras are attracting significant interest as they provide rich edge information, high dynamic range, and high temporal resolution. Many state-of-the-art event-based algorithms rely on splitting the events into fixed groups, resulting in the omission of crucial temporal information, particularly when dealing with diverse motion scenarios (e.g., high/low speed). In this work, we propose SpikeSlicer, a novel-designed plug-and-play event processing method capable of splitting events stream adaptively. SpikeSlicer utilizes a lightweight (0.41M) and low-energy spiking neural network (SNN) to trigger event slicing. To guide the SNN to fire spikes at optimal time steps, we propose the Spiking Position-aware Loss (SPA-Loss) to modulate the neuron's state. Additionally, we develop a Feedback-Update training strategy that refines the slicing decisions using feedback from the downstream artificial neural network (ANN). Extensive experiments demonstrate that our method yields significant performance improvements in event-based object tracking and recognition. Notably, SpikeSlicer provides a brand-new SNN-ANN cooperation paradigm, where the SNN acts as an efficient, low-energy data processor to assist the ANN in improving downstream performance, injecting new perspectives and potential avenues of exploration., Comment: Accepted to NeurIPS 2024
- Published
- 2024
38. ExACT: Teaching AI Agents to Explore with Reflective-MCTS and Exploratory Learning
- Author
-
Yu, Xiao, Peng, Baolin, Vajipey, Vineeth, Cheng, Hao, Galley, Michel, Gao, Jianfeng, and Yu, Zhou
- Subjects
Computer Science - Computation and Language ,Computer Science - Computer Vision and Pattern Recognition - Abstract
Autonomous agents have demonstrated significant potential in automating complex multistep decision-making tasks. However, even state-of-the-art vision-language models (VLMs), such as GPT-4o, still fall short of human-level performance, particularly in intricate web environments and long-horizon tasks. To address these limitations, we present ExACT, an approach to combine test-time search and self-learning to build o1-like models for agentic applications. We first introduce Reflective Monte Carlo Tree Search (R-MCTS), a novel test time algorithm designed to enhance AI agents' ability to explore decision space on the fly. R-MCTS extends traditional MCTS by 1) incorporating contrastive reflection, allowing agents to learn from past interactions and dynamically improve their search efficiency; and 2) using multi-agent debate for reliable state evaluation. Next, we introduce Exploratory Learning, a novel learning strategy to teach agents to search at inference time without relying on any external search algorithms. On the challenging VisualWebArena benchmark, our GPT-4o based R-MCTS agent achieves a 6% to 30% relative improvement across various tasks compared to the previous state-of-the-art. Additionally, we show that the knowledge and experience gained from test-time search can be effectively transferred back to GPT-4o via fine-tuning. After Exploratory Learning, GPT-4o 1) demonstrates the ability to explore the environment, evaluate a state, and backtrack to viable ones when it detects that the current state cannot lead to success, and 2) matches 87% of R-MCTS's performance while using significantly less compute. Notably, our work demonstrates the compute scaling properties in both training - data collection with R-MCTS - and testing time. These results suggest a promising research direction to enhance VLMs' capabilities for agentic applications via test-time search and self-learning.
- Published
- 2024
39. Lessons Learned from a Unifying Empirical Study of Parameter-Efficient Transfer Learning (PETL) in Visual Recognition
- Author
-
Mai, Zheda, Zhang, Ping, Tu, Cheng-Hao, Chen, Hong-You, Zhang, Li, and Chao, Wei-Lun
- Subjects
Computer Science - Machine Learning ,Computer Science - Artificial Intelligence ,Computer Science - Computer Vision and Pattern Recognition - Abstract
Parameter-efficient transfer learning (PETL) has attracted significant attention lately, due to the increasing size of pre-trained models and the need to fine-tune (FT) them for superior downstream performance. This community-wide enthusiasm has sparked a plethora of approaches. Nevertheless, a systematic study to understand their performance and suitable application scenarios is lacking, leaving questions like when to apply PETL and which approach to use largely unanswered. In this paper, we conduct a unifying empirical study of representative PETL methods in the context of Vision Transformers. We systematically tune their hyper-parameters to fairly compare their accuracy on downstream tasks. Our study not only offers a valuable user guide but also unveils several new insights. First, if tuned carefully, different PETL methods can obtain similar accuracy in the low-shot benchmark VTAB-1K. This includes simple methods like FT the bias terms that were reported inferior. Second, though with similar accuracy, we find that PETL methods make different mistakes and high-confidence predictions, likely due to their different inductive biases. Such an inconsistency (or complementariness) opens up the opportunity for ensemble methods, and we make preliminary attempts at this. Third, going beyond the commonly used low-shot tasks, we find that PETL is also useful in many-shot regimes -- it achieves comparable and sometimes better accuracy than full FT, using much fewer learnable parameters. Last but not least, we investigate PETL's ability to preserve a pre-trained model's robustness to distribution shifts (e.g., a CLIP backbone). Perhaps not surprisingly, PETL methods outperform full FT alone. However, with weight-space ensembles, the fully fine-tuned model can better balance target (i.e., downstream) distribution and distribution shift performance, suggesting a future research direction for PETL., Comment: Code is available at https://github.com/OSU-MLB/PETL_Vision
- Published
- 2024
40. Fine-Tuning is Fine, if Calibrated
- Author
-
Mai, Zheda, Chowdhury, Arpita, Zhang, Ping, Tu, Cheng-Hao, Chen, Hong-You, Pahuja, Vardaan, Berger-Wolf, Tanya, Gao, Song, Stewart, Charles, Su, Yu, and Chao, Wei-Lun
- Subjects
Computer Science - Machine Learning ,Computer Science - Artificial Intelligence ,Computer Science - Computer Vision and Pattern Recognition - Abstract
Fine-tuning is arguably the most straightforward way to tailor a pre-trained model (e.g., a foundation model) to downstream applications, but it also comes with the risk of losing valuable knowledge the model had learned in pre-training. For example, fine-tuning a pre-trained classifier capable of recognizing a large number of classes to master a subset of classes at hand is shown to drastically degrade the model's accuracy in the other classes it had previously learned. As such, it is hard to further use the fine-tuned model when it encounters classes beyond the fine-tuning data. In this paper, we systematically dissect the issue, aiming to answer the fundamental question, "What has been damaged in the fine-tuned model?" To our surprise, we find that the fine-tuned model neither forgets the relationship among the other classes nor degrades the features to recognize these classes. Instead, the fine-tuned model often produces more discriminative features for these other classes, even if they were missing during fine-tuning! {What really hurts the accuracy is the discrepant logit scales between the fine-tuning classes and the other classes}, implying that a simple post-processing calibration would bring back the pre-trained model's capability and at the same time unveil the feature improvement over all classes. We conduct an extensive empirical study to demonstrate the robustness of our findings and provide preliminary explanations underlying them, suggesting new directions for future theoretical analysis. Our code is available at https://github.com/OSU-MLB/Fine-Tuning-Is-Fine-If-Calibrated., Comment: The paper has been accepted to NeurIPS 2024. The first three authors contribute equally
- Published
- 2024
41. Joint State-Channel Decoupling and One-Shot Quantum Coding Theorem
- Author
-
Cheng, Hao-Chung, Dupuis, Frédéric, and Gao, Li
- Subjects
Quantum Physics - Abstract
In this work, we consider decoupling a bipartite quantum state via a general quantum channel. We propose a joint state-channel decoupling approach to obtain a one-shot error exponent bound without smoothing, in which trace distance is used to measure how good the decoupling is. The established exponent is expressed in terms of a sum of two sandwiched R{\'e}nyi entropies, one quantifying the amount of initial correlation between the state and environment, while the other characterizing the effectiveness of the quantum channel. This gives an explicit exponential decay of the decoupling error in the whole achievable region, which was missing in the previous results [Commun. Math. Phys. 328, 2014]. Moreover, it strengthens the error exponent bound obtained in a recent work [IEEE Trans. Inf. Theory, 69(12), 2023], for exponent from the channel part. As an application, we establish a one-shot error exponent bound for quantum channel coding given by a sandwiched R\'enyi coherent information., Comment: 25 pages, 2 figures. Presented in QIP 2023. Comments are very welcome
- Published
- 2024
42. Morphological Detection and Classification of Microplastics and Nanoplastics Emerged from Consumer Products by Deep Learning
- Author
-
Rezvani, Hadi, Zarrabi, Navid, Mehta, Ishaan, Kolios, Christopher, Jaafar, Hussein Ali, Kao, Cheng-Hao, Saeedi, Sajad, and Yousefi, Nariman
- Subjects
Computer Science - Computer Vision and Pattern Recognition ,Computer Science - Artificial Intelligence ,Statistics - Applications ,Statistics - Methodology - Abstract
Plastic pollution presents an escalating global issue, impacting health and environmental systems, with micro- and nanoplastics found across mediums from potable water to air. Traditional methods for studying these contaminants are labor-intensive and time-consuming, necessitating a shift towards more efficient technologies. In response, this paper introduces micro- and nanoplastics (MiNa), a novel and open-source dataset engineered for the automatic detection and classification of micro and nanoplastics using object detection algorithms. The dataset, comprising scanning electron microscopy images simulated under realistic aquatic conditions, categorizes plastics by polymer type across a broad size spectrum. We demonstrate the application of state-of-the-art detection algorithms on MiNa, assessing their effectiveness and identifying the unique challenges and potential of each method. The dataset not only fills a critical gap in available resources for microplastic research but also provides a robust foundation for future advancements in the field.
- Published
- 2024
43. Manipulation Facing Threats: Evaluating Physical Vulnerabilities in End-to-End Vision Language Action Models
- Author
-
Cheng, Hao, Xiao, Erjia, Yu, Chengyuan, Yao, Zhao, Cao, Jiahang, Zhang, Qiang, Wang, Jiaxu, Sun, Mengshu, Xu, Kaidi, Gu, Jindong, and Xu, Renjing
- Subjects
Computer Science - Computer Vision and Pattern Recognition - Abstract
Recently, driven by advancements in Multimodal Large Language Models (MLLMs), Vision Language Action Models (VLAMs) are being proposed to achieve better performance in open-vocabulary scenarios for robotic manipulation tasks. Since manipulation tasks involve direct interaction with the physical world, ensuring robustness and safety during the execution of this task is always a very critical issue. In this paper, by synthesizing current safety research on MLLMs and the specific application scenarios of the manipulation task in the physical world, we comprehensively evaluate VLAMs in the face of potential physical threats. Specifically, we propose the Physical Vulnerability Evaluating Pipeline (PVEP) that can incorporate as many visual modal physical threats as possible for evaluating the physical robustness of VLAMs. The physical threats in PVEP specifically include Out-of-Distribution, Typography-based Visual Prompt, and Adversarial Patch Attacks. By comparing the performance fluctuations of VLAMs before and after being attacked, we provide generalizable \textbf{\textit{Analyses}} of how VLAMs respond to different physical security threats.
- Published
- 2024
44. GRIN: GRadient-INformed MoE
- Author
-
Liu, Liyuan, Kim, Young Jin, Wang, Shuohang, Liang, Chen, Shen, Yelong, Cheng, Hao, Liu, Xiaodong, Tanaka, Masahiro, Wu, Xiaoxia, Hu, Wenxiang, Chaudhary, Vishrav, Lin, Zeqi, Zhang, Chenruidong, Xue, Jilong, Awadalla, Hany, Gao, Jianfeng, and Chen, Weizhu
- Subjects
Computer Science - Computation and Language ,Computer Science - Artificial Intelligence ,Computer Science - Machine Learning - Abstract
Mixture-of-Experts (MoE) models scale more effectively than dense models due to sparse computation through expert routing, selectively activating only a small subset of expert modules. However, sparse computation challenges traditional training practices, as discrete expert routing hinders standard backpropagation and thus gradient-based optimization, which are the cornerstone of deep learning. To better pursue the scaling power of MoE, we introduce GRIN (GRadient-INformed MoE training), which incorporates sparse gradient estimation for expert routing and configures model parallelism to avoid token dropping. Applying GRIN to autoregressive language modeling, we develop a top-2 16$\times$3.8B MoE model. Our model, with only 6.6B activated parameters, outperforms a 7B dense model and matches the performance of a 14B dense model trained on the same data. Extensive evaluations across diverse tasks demonstrate the potential of GRIN to significantly enhance MoE efficacy, achieving 79.4 on MMLU, 83.7 on HellaSwag, 74.4 on HumanEval, and 58.9 on MATH., Comment: 58 pages
- Published
- 2024
45. Photo-nuclear reaction rates of $^{157,159}$Ho and $^{163,165}$Tm and their impact in the $\gamma$--process
- Author
-
Cheng, Hao, Sun, Bao-Hua, Zhu, Li-Hua, Kusakabe, Motohiko, Luo, Yudong, Kajino, Toshitaka, Wang, Chang-Jian, Yao, Xing-Qun, He, Chuang-Ye, Liu, Fu-Long, and Guo, Bing
- Subjects
Astrophysics - Solar and Stellar Astrophysics ,Nuclear Experiment ,Nuclear Theory - Abstract
Reliable photo-nuclear reaction rates at the stellar conditions are essential to understand the origin of the heavy stable neutron-deficient isotopes between $^{74}$Se and $^{196}$Hg-p-nuclei, however, many reaction rates of relevance still have to rely on the Hauser-Feshbach model due to rare experimental progress. One such case is in the mass range of 160 for Dy, Er, Ho and Tm isotopes. In this work we attempt to constrain the Hauser-Feshbach model in the TALYS package by reproducing the available experimental data of $^{160}$Dy($p,\gamma$)$^{161}$Ho and $^{162}$Er($p,\gamma$)$^{163}$Tm in the $A\sim 160$ mass region, and examine the effects of level density, gamma strength function and the optical model potential. The constrained model then allows us to calculate the reaction rates of $^{157, 159}$Ho($\gamma$, $p$) and $^{163,165}$Tm($\gamma$, $p$) for the $\gamma$-process nucleosynthesis in carbon-deflagration SNe Ia model. Our recommended rates differ from the JINA REACLIB by more than 1 order of magnitude in the temperature range of 2-3 GK. This results in the changes of final abundance of $p$-nuclei in the $A\sim 160$ mass range by -5.5-3\% from those with JINA, which means that the ($\gamma$, $p$) reactions uncertainty is not predominant for the synthesis of these nuclei., Comment: 12 pages,7 figures
- Published
- 2024
- Full Text
- View/download PDF
46. Model Tells Itself Where to Attend: Faithfulness Meets Automatic Attention Steering
- Author
-
Zhang, Qingru, Yu, Xiaodong, Singh, Chandan, Liu, Xiaodong, Liu, Liyuan, Gao, Jianfeng, Zhao, Tuo, Roth, Dan, and Cheng, Hao
- Subjects
Computer Science - Computation and Language ,Computer Science - Artificial Intelligence - Abstract
Large language models (LLMs) have demonstrated remarkable performance across various real-world tasks. However, they often struggle to fully comprehend and effectively utilize their input contexts, resulting in responses that are unfaithful or hallucinated. This difficulty increases for contexts that are long or contain distracting information, which can divert LLMs from fully capturing essential evidence. To address this issue, many works use prompting to help LLMs utilize contextual information more faithfully. For instance, iterative prompting highlights key information in two steps that first ask the LLM to identify important pieces of context and then derive answers accordingly. However, prompting methods are constrained to highlighting key information implicitly in token space, which is often insufficient to fully steer the model's attention. To improve model faithfulness more reliably, we propose AutoPASTA, a method that automatically identifies key contextual information and explicitly highlights it by steering an LLM's attention scores. Like prompting, AutoPASTA is applied at inference time and does not require changing any model parameters. Our experiments on open-book QA demonstrate that AutoPASTA effectively enables models to grasp essential contextual information, leading to substantially improved model faithfulness and performance, e.g., an average improvement of 7.95% for LLAMA3-70B-Instruct. Code will be publicly available at https://github.com/QingruZhang/AutoPASTA ., Comment: 12 pages, 4 figures
- Published
- 2024
47. Mamba Policy: Towards Efficient 3D Diffusion Policy with Hybrid Selective State Models
- Author
-
Cao, Jiahang, Zhang, Qiang, Sun, Jingkai, Wang, Jiaxu, Cheng, Hao, Li, Yulin, Ma, Jun, Shao, Yecheng, Zhao, Wen, Han, Gang, Guo, Yijie, and Xu, Renjing
- Subjects
Computer Science - Robotics ,Computer Science - Computer Vision and Pattern Recognition - Abstract
Diffusion models have been widely employed in the field of 3D manipulation due to their efficient capability to learn distributions, allowing for precise prediction of action trajectories. However, diffusion models typically rely on large parameter UNet backbones as policy networks, which can be challenging to deploy on resource-constrained devices. Recently, the Mamba model has emerged as a promising solution for efficient modeling, offering low computational complexity and strong performance in sequence modeling. In this work, we propose the Mamba Policy, a lighter but stronger policy that reduces the parameter count by over 80% compared to the original policy network while achieving superior performance. Specifically, we introduce the XMamba Block, which effectively integrates input information with conditional features and leverages a combination of Mamba and Attention mechanisms for deep feature extraction. Extensive experiments demonstrate that the Mamba Policy excels on the Adroit, Dexart, and MetaWorld datasets, requiring significantly fewer computational resources. Additionally, we highlight the Mamba Policy's enhanced robustness in long-horizon scenarios compared to baseline methods and explore the performance of various Mamba variants within the Mamba Policy framework. Our project page is in https://andycao1125.github.io/mamba_policy/., Comment: 7 pages, 5 figures
- Published
- 2024
48. Linear Convergence in Hilbert's Projective Metric for Computing Augustin Information and a R\'{e}nyi Information Measure
- Author
-
Tsai, Chung-En, Wang, Guan-Ren, Cheng, Hao-Chung, and Li, Yen-Huan
- Subjects
Mathematics - Optimization and Control ,Computer Science - Information Theory - Abstract
Consider the problems of computing the Augustin information and a R\'{e}nyi information measure of statistical independence, previously explored by Lapidoth and Pfister (IEEE Information Theory Workshop, 2018) and Tomamichel and Hayashi (IEEE Trans. Inf. Theory, 64(2):1064--1082, 2018). Both quantities are defined as solutions to optimization problems and lack closed-form expressions. This paper analyzes two iterative algorithms: Augustin's fixed-point iteration for computing the Augustin information, and the algorithm by Kamatsuka et al. (arXiv:2404.10950) for the R\'{e}nyi information measure. Previously, it was only known that these algorithms converge asymptotically. We establish the linear convergence of Augustin's algorithm for the Augustin information of order $\alpha \in (1/2, 1) \cup (1, 3/2)$ and Kamatsuka et al.'s algorithm for the R\'{e}nyi information measure of order $\alpha \in [1/2, 1) \cup (1, \infty)$, using Hilbert's projective metric., Comment: 15 pages, last sentence of the first paragraph and Eq. (2) corrected
- Published
- 2024
49. Spiking Diffusion Models
- Author
-
Cao, Jiahang, Guo, Hanzhong, Wang, Ziqing, Zhou, Deming, Cheng, Hao, Zhang, Qiang, and Xu, Renjing
- Subjects
Computer Science - Neural and Evolutionary Computing ,Computer Science - Computer Vision and Pattern Recognition - Abstract
Recent years have witnessed Spiking Neural Networks (SNNs) gaining attention for their ultra-low energy consumption and high biological plausibility compared with traditional Artificial Neural Networks (ANNs). Despite their distinguished properties, the application of SNNs in the computationally intensive field of image generation is still under exploration. In this paper, we propose the Spiking Diffusion Models (SDMs), an innovative family of SNN-based generative models that excel in producing high-quality samples with significantly reduced energy consumption. In particular, we propose a Temporal-wise Spiking Mechanism (TSM) that allows SNNs to capture more temporal features from a bio-plasticity perspective. In addition, we propose a threshold-guided strategy that can further improve the performances by up to 16.7% without any additional training. We also make the first attempt to use the ANN-SNN approach for SNN-based generation tasks. Extensive experimental results reveal that our approach not only exhibits comparable performance to its ANN counterpart with few spiking time steps, but also outperforms previous SNN-based generative models by a large margin. Moreover, we also demonstrate the high-quality generation ability of SDM on large-scale datasets, e.g., LSUN bedroom. This development marks a pivotal advancement in the capabilities of SNN-based generation, paving the way for future research avenues to realize low-energy and low-latency generative applications. Our code is available at https://github.com/AndyCao1125/SDM., Comment: Accepted by IEEE Transactions on Artificial Intelligence
- Published
- 2024
50. Prediction of Safety Risk Levels of Benzopyrene Residues in Edible Oils in China Based on the Variable-Weight Combined LSTM-XGBoost Prediction Model
- Author
-
Cheng Hao, Qingchuan Zhang, Shimin Wang, Tongqiang Jiang, and Wei Dong
- Subjects
risk assessment ,LSTM ,XGBoost ,risk prediction ,edible oil ,BaP ,Chemical technology ,TP1-1185 - Abstract
To assess and predict the food safety risk of benzopyrene (BaP) in edible oils in China, this study collected national sampling data of edible oils from 20 Chinese provinces and their prefectures in 2019, and constructed a risk assessment model of BaP in edible oils with consumption data. Initially, the k-means algorithm was used for risk classification; then the data were pre-processed and trained to predict the data using the Long Short-Term Memory (LSTM) and the eXtreme Gradient Boosting (XGBoost) models, respectively, and finally, the two models were combined using the inverse error method. To test the effectiveness of the prediction model, this study experimentally validated the model according to five evaluation metrics: root mean square error (RMSE), mean absolute error (MAE), precision, recall, and F1 score. The variable-weight combined LSTM-XGBoost prediction model proposed in this paper achieved a precision of 94.62%, and the F1 score value reached 95.16%, which is significantly better than other neural network models; the results demonstrate that the prediction model has certain stability and feasibility. Overall, the combined model used in this study not only improves the accuracy but also enhances the practicality, real-time capabilities, and expandability of the model.
- Published
- 2023
- Full Text
- View/download PDF
Catalog
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.