A5100379455- Advanced Combustion Engine Technologies
- Topic Modeling
- Recommender Systems and Techniques
- Domain Adaptation and Few-Shot Learning
- Natural Language Processing Techniques
- Reinforcement Learning in Robotics
- Combustion and flame dynamics
- Machine Learning and Data Classification
- Cryptography and Data Security
- Adversarial Robustness in Machine Learning
- Atmospheric chemistry and aerosols
- Advanced Bandit Algorithms Research
- Speech and Audio Processing
- Blockchain Technology Applications and Security
- Privacy-Preserving Technologies in Data
- Machine Learning and Algorithms
- Multimodal Machine Learning Applications
- Catalytic Processes in Materials Science
- Advanced Neural Network Applications
- Gaussian Processes and Bayesian Inference
- Data Management and Algorithms
- Thermochemical Biomass Conversion Processes
- Advanced Graph Neural Networks
- Direction-of-Arrival Estimation Techniques
- Model Reduction and Neural Networks
University of Electronic Science and Technology of China
2024-2025
Delft University of Technology
2022-2024
Jianghan University
2024
Shanghai Jiao Tong University
2011-2023
Kunming University of Science and Technology
2023
University of California, Los Angeles
2023
University of International Business and Economics
2023
Alibaba Group (China)
2023
Beihang University
2023
Georgia Institute of Technology
2023
Large language models (LLMs) have demonstrated remarkable performance on a variety of natural tasks based just few examples instructions, reducing the need for extensive feature engineering. However, most powerful LLMs are closed-source or limited in their capability languages other than English. In this technical report, we present Baichuan 2, series large-scale multilingual containing 7 billion and 13 parameters, trained from scratch, 2.6 trillion tokens. 2 matches outperforms open-source...
Agile skills learned with our proposed method enable the A1 robot to jump repeatedly (left) and walk a goal location
Semantic segmentation and height estimation tasks in remote sensing imagery exhibit distinctive characteristics, including scale sensitivity, category imbalance, insufficient fine details. Recent approaches have leveraged multi-task learning methods to jointly predict these along with auxiliary tasks, such as edge detection, improve the accuracy of fine-grained However, most only acquire knowledge from disregarding inter-task guidance across all tasks. To address challenges, we propose...
Recently, GAN-based AI painting tools, like Midjourney and Stable Diffusion, are making a huge impact among the designer community due to their rapidly advancing effects generalizability various design fields. However, attitude behavioral intention of designers towards tools vary greatly across This study focuses on exploring influential factors intentions for from Through semi-structured interviews with 9 5 different fields, we propose an extended UTAUT model, which highlights three...
To overcome the domain gap between synthetic and real-world datasets, unsupervised adaptation methods have been proposed for semantic segmentation. Majority of previous approaches attempted to reduce either at pixel or feature level, disregarding fact that two components interact positively. address this, we present CONtrastive FEaTure pIxel alignment (CON-FETI) bridging both levels using a unique contrastive formulation. We introduce well-estimated prototypes by including category-wise...
This paper leverages heterogeneous auxiliary information to address the data sparsity problem of recommender systems. We propose a model that learns shared feature space from data, such as item descriptions, product tags and online purchase history, obtain better predictions. Our consists autoencoders, not only for numerical categorical but also sequential which enables capturing user tastes, characteristics recent dynamics preference. learn autoencoder architecture each source independently...
In this study, we report new approaches for the fabrication of Gd–Ba–Cu–O (GdBCO) bulk with a large size and promising superconducting properties, applying cold‐seeding melt‐textured (MT) method. Firstly, NdBCO thin films were used as seeds because their superheating properties an additional mini pellet was inserted between seed precursor. This approach allows to reach higher maximum processing temperature ( T max ) wider window single grain growth, leading growth large‐sized bulk, 56 mm in...
Large language models (LLMs), including ChatGPT, Bard, and Llama, have achieved remarkable successes over the last two years in a range of different applications. In spite these successes, there exist concerns that limit wide application LLMs. A key problem is hallucination. Hallucination refers to fact addition correct responses, LLMs can also generate seemingly but factually incorrect responses. This report aims present comprehensive review current literature on both hallucination...
The aim of the study is to identify, and present CSR practices that have been, are also currently being implemented by football clubs in era Covid-19 pandemic. Moreover, it strives assess impact these on media coverage. results research show among posts published Zagłębie Sosnowiec social platforms, such as their official Facebook profile, about matches, tend achieve greatest reach. However, contrary, presenting ongoing initiatives for local community come take a backseat, considered be...
Modern automatic speech recognition (ASR) model is required to accurately transcribe diverse signals (from different domains, languages, accents, etc) given the specific contextual information in various application scenarios. Classic end-to-end models fused with extra language perform well, but mainly data matching scenarios and are gradually approaching a bottleneck. In this work, we introduce Seed-ASR, large (LLM) based model. Seed-ASR developed on framework of audio conditioned LLM...