A5025228078- Multimodal Machine Learning Applications
- Transportation and Mobility Innovations
- Blood groups and transfusion
- Machine Learning and Data Classification
- Ruminant Nutrition and Digestive Physiology
- Animal Nutrition and Physiology
- Mobile Crowdsensing and Crowdsourcing
- Domain Adaptation and Few-Shot Learning
- Electric Vehicles and Infrastructure
- Smart Parking Systems Research
- Infrastructure Maintenance and Monitoring
- Geophysical Methods and Applications
- Nitrogen and Sulfur Effects on Brassica
- Turfgrass Adaptation and Management
- Transportation Planning and Optimization
- Animal Behavior and Welfare Studies
- Radiomics and Machine Learning in Medical Imaging
- Single-cell and spatial transcriptomics
- Generative Adversarial Networks and Image Synthesis
- Color perception and design
- Rabbits: Nutrition, Reproduction, Health
- Aging and Gerontology Research
- Soft Robotics and Applications
- Advanced Materials and Mechanics
- Millimeter-Wave Propagation and Modeling
Heilongjiang Bayi Agricultural University
2025
Carnegie Mellon University
2021-2024
South China University of Technology
2024
Jilin University
2023
Shanghai Jiao Tong University
2021-2022
Beijing University of Posts and Telecommunications
2021
China Agricultural University
2021
Sichuan Agricultural University
2018
BGI Group (China)
2017
China National GeneBank
2017
Object models are gradually progressing from predicting just category labels to providing detailed descriptions of object instances. This motivates the need for large datasets which go beyond traditional masks and provide richer annotations such as part attributes. Hence, we introduce PACO: Parts Attributes Common Objects. It spans 75 categories, 456 object-part categories 55 attributes across image (LVIS) video (Eg04D) datasets. We 641K an-notated 260K boxes, with roughly half them...
Training text-to-image models with web scale image-text pairs enables the generation of a wide range visual concepts from text. However, these pre-trained often face challenges when it comes to generating highly aesthetic images. This creates need for alignment post pre-training. In this paper, we propose quality-tuning effectively guide model exclusively generate visually appealing images, while maintaining generality across concepts. Our key insight is that supervised fine-tuning set...
Two experiments were conducted to investigate the feeding value of wood vinegar in weanling pigs.In Experiment 1, pigs (n = 224; Landrace×Yorkshire×Duroc, 21±3 d-old, initial BW 6.12±0.10kg) assigned four dietary treatments.Different levels added diets as treatments (0, 0.1, 0.2 and 0.3%).Each treatment comprised 4 replicates with 14 piglets each.Experimental was for 28 d two phases (phase I, 0 phase II, 15 28).Feeding linearly (p<0.05)improved II overall ADG increased (linear, p<0.05)...
In order to investigate the effect of physical forms starter diets on performance, weaning age, nutrient digestibility and rumen biochemical factors, 24 female neonatal Brown Swiss calves (average body weight 39.5±1.2 kg) were randomly assigned three treatments. Dietary treatments mashed (MS), pelleted (PS), texturized (TS) using 8 from birth till 90 days age in each treatment. Diets formulated be iso-nitrogenous with 21% crude protein. Based experimental results, that received PS TS diets,...
This study aimed to investigate the relationship between functional limitation and happiness among Chinese older people examined multiple mediating effects of intergenerational support (instrumental financial support) relationship.Data was drawn from Family Panel Survey (CFPS) 2018 2020. Structural equation modeling adopted analyze association limitations, support, relationship, adults happiness.There a significant limitations lower levels adults. The instrumental adult children positively...
This paper proposes a new primitive that allows soft robots to be physically controlled in completely non-line-of-sight context using wireless energy - process we call actuation. Soft robots, which are composed entirely of materials and exclude any rigid components, highly flexible platforms can change their shape. considers specific class liquid-crystal elastomers (LCE) electronics-free engineered shape when heated 60 °C. Traditionally, such robotic systems must line-of-sight light source,...
Nowadays, crowdsourcing has become an increasingly popular paradigm for large-scale data collection, annotation, and classification. Today's rapid growth of platforms calls effective worker selection mechanisms, which oftentimes have to operate with a priori unknown reliability. We discover that the empirical entropy workers' results, measures uncertainty in final aggregated naturally becomes suitable metric evaluate outcome tasks. Therefore, this paper designs mechanism minimizes results...
The label noise transition matrix, characterizing the probabilities of a training instance being wrongly annotated, is crucial to designing popular solutions learning with noisy labels. Existing works heavily rely on finding "anchor points" or their approximates, defined as instances belonging particular class almost surely. Nonetheless, anchor points remains non-trivial task, and estimation accuracy also often throttled by number available points. In this paper, we propose an alternative...
Companies today increasingly rely on microservice architectures to deliver service for their large-scale mobile or web applications. However, not all developers working these applications are distributed systems engineers and therefore do anticipate partial failure: where one more of the dependencies might be unavailable once deployed into production. Therefore, it is paramount that issues raised early often, ideally in a testing environment before code ships
Background: Intention to smoke is an important predictor of future smoking among adolescents. The purpose our study was examine the interaction between academic performance and parents/peer tobacco use on adolescents’ intention smoke. Methods: A multi-stage stratified sampling used select participants, involving 9394 students aged 9–16 years in Changchun city, northeastern China. Multiple logistic regression analyses were conducted individual effect peer/parental behavior. Stratified...
Recent years have witnessed the increasing popularity of electric buses (e-buses) around globe due to their environment friendly nature. However, various factors, such as prohibitive purchasing costs and scarcity large-scale charging facilities, hinder wider adoption e-buses. Thus, effectively cut cost building maintaining urban e-bus systems, we optimize dispatching strategy for systems satisfy public transportation demands with minimum fleet. Specifically, propose systematically exploit at...
Nowadays, the rapid development of self-driving technology and its fusion with current vehicle electrification process has given rise to electric taxis (es-taxis). Foreseeably, es-taxis will become a major force that serves massive urban mobility demands not far into future. Though promising, it is still fundamental unsolved problem effectively deciding when where city-scale fleet should be charged, so enough available whenever wherever ride requests are submitted. Furthermore, charging...
The rapid development of information and communication technologies has given rise to mobility-on-demand (MoD) systems (e.g., Uber, Didi) that have fundamentally revolutionized urban transportation. One common feature today's MoD is the integration ridesharing due its cost-efficient environment-friendly natures. However, a fundamental unsolved problem for such how serve people's heterogeneous transportation demands with as few vehicles possible. Naturally, solving minimum fleet essential...
HLA-DRB3, DRB4 and DRB5 (DRB3/4/5) are paralogues of HLA-DRB1. They have important roles in transplantation been reported to be related many diseases. HLA typing methods for DRB3/4/5 based on NGS data limitations now, such as need polymerase chain reaction (PCR) or low accuracy.We present a method read mapping haplotype assembly from data. Also, copy number is determined by k-means clustering according ratio sequencing depth between DRB1.We achieved 100%, 100% accuracy simulated 95.88%,...
Channel estimation is critical to millimeter-wave capability. Unlike sub-6 GHz WiFi, commercial-off-the-shelf 60 WiFi devices adopt a single RF-chain and can only report the combined received signal strength (RSS) instead of antenna-wise channel state information (CSI). Therefore, recovering CSI using limited number RSS measurements important but faces following challenges: (i) solving non-convex objective hard computationally heavy, (ii) error high with insufficient measurements, (iii)...
To establish prediction models for human leukocyte antigen (HLA) haplotypes and HLA genotypes, verify the accuracy.
Early detection of dental disease is crucial to prevent adverse outcomes. Today, X-rays are currently the most accurate gold standard for detection. Unfortunately, regular X-ray exam still a privilege billions people around world. In this paper, we ask: "Can develop low-cost sensing system that enables self-examination in comfort one's home?" This paper presents ToMoBrush, health explores using off-the-shelf sonic toothbrushes condition Our solution leverages fact toothbrush produces rich...
Microwave ovens have become an essential cooking appliance owing to their convenience and efficiency. However, microwave suffer from uneven distribution of energy, which causes prolonged delays, unpleasant experiences, even safety concerns. Despite significant research efforts, current solutions remain inadequate. In this paper, we first conduct measurement studies understand the energy for 10 show in both 2D 3D is very skewed, with notably lower levels at center cavity, where food commonly...
Object models are gradually progressing from predicting just category labels to providing detailed descriptions of object instances. This motivates the need for large datasets which go beyond traditional masks and provide richer annotations such as part attributes. Hence, we introduce PACO: Parts Attributes Common Objects. It spans 75 categories, 456 object-part categories 55 attributes across image (LVIS) video (Ego4D) datasets. We 641K annotated 260K boxes, with roughly half them...