A5100375342- Natural Language Processing Techniques
- Topic Modeling
- Multimodal Machine Learning Applications
- Translation Studies and Practices
- Speech Recognition and Synthesis
- Speech and dialogue systems
- Domain Adaptation and Few-Shot Learning
- Technology Adoption and User Behaviour
- Advanced Text Analysis Techniques
- Information Technology Governance and Strategy
- Web Data Mining and Analysis
- Distributed systems and fault tolerance
- Advanced Data Storage Technologies
- Big Data and Business Intelligence
- Advanced Computational Techniques and Applications
- Advanced Wireless Communication Techniques
- Advanced Neural Network Applications
- Distributed and Parallel Computing Systems
- Folklore, Mythology, and Literature Studies
- Wireless Communication Networks Research
- Cooperative Communication and Network Coding
- Adversarial Robustness in Machine Learning
- Text Readability and Simplification
- Digital Marketing and Social Media
- Advanced MIMO Systems Optimization
Chinese Academy of Sciences
2014-2025
Suzhou Institute of Biomedical Engineering and Technology
2025
Beijing University of Posts and Telecommunications
2022-2024
Huawei Technologies (China)
2023-2024
Nanjing University
2014-2023
Chinese University of Hong Kong, Shenzhen
2023
University of Chinese Academy of Sciences
2018-2023
University of Michigan
2023
Roskilde University
2023
Nanjing Medical University
2023
Yiming Cui, Ting Liu, Wanxiang Che, Li Xiao, Zhipeng Chen, Wentao Ma, Shijin Wang, Guoping Hu. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint (EMNLP-IJCNLP). 2019.
Multi-task learning (MTL) has been widely applied in Natural Language Processing. A major task and its associated auxiliary tasks share the same encoder; hence, an MTL encoder can learn sharing abstract information between tasks. Task-specific towers are then employed upon to task-specific information. Previous works demonstrated that exchanging yielded extra gains. This is known as soft-parameter MTL. In this paper, we propose a novel gating mechanism for bridging of towers. Our method...
Human faces are used extensively in print advertisements. In prior literature, researchers have studied spokespersons general, but few explicitly. This paper aims to answer three questions that important both and practitioners: (1) Do affect how a viewer reacts an advertisement on the metrics advertisers care about? (2) If do effect, is it large enough warrant careful selection of when constructing advertisements? (3) effect large, what facial features elicit such differential reactions...
The aggregation-induced enhanced photosensitization and emission (AIEPE) properties of spiropyrans in nanoassemblies enable reversibly controlled singlet oxygen generation.
We propose a conceptual framework that includes the antecedents and consequences of firms’ adopting integrating robotics into their customer service operations. Drawing insights from literature...
The fluorescence quantum yield of side-chain AIE polymers was remarkably promoted just by shortening the linking spacer.
Using reinforcement learning to learn control policies is a challenge when the task complex with potentially long horizons. Ensuring adequate but safe exploration also crucial for controlling physical systems. In this paper, we use temporal logic facilitate specification and of tasks. We combine Lyapunov functions improve exploration. incorporate barrier safeguard deployment process. develop flexible learnable system that allows users specify objectives constraints in different forms at...
Abstract. Organic aerosol (OA) has been considered as one of the most important uncertainties in climate modeling due to complexity presenting its chemical production and depletion mechanisms. To better understand capability models probe into associated simulating OA, we evaluate Community Earth System Model version 2.1 (CESM2.1) configured with Atmosphere 6 (CAM6) comprehensive tropospheric stratospheric chemistry representation (CAM6-Chem) through a long-term simulation (1988–2019)...
Recent machine reading comprehension datasets such as ReClor and LogiQA require performing logical reasoning over text. Conventional neural models are insufficient for reasoning, while symbolic reasoners cannot directly apply to To meet the challenge, we present a neural-symbolic approach which, predict an answer, passes messages graph representing relations between text units. It incorporates adaptive logic network (AdaLoGN) which adaptively infers extend and, essentially, realizes mutual...
Zixian Huang, Yulin Shen, Xiao Li, Yu'ang Wei, Gong Cheng, Lin Zhou, Xinyu Dai, Yuzhong Qu. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint (EMNLP-IJCNLP). 2019.
One important issue that affects the performance of neural machine translation is scale available parallel data. For low-resource languages, amount data not sufficient, which results in poor quality. In this paper, we propose a diversity augmentation method does use extra monolingual We expand training by generating pseudo on source and target sides. To generate data, restricted sampling strategy employed at decoding steps. Finally, filter merge origin synthetic corpus to train final model....
Numerical reasoning over hybrid data containing tables and long texts has recently received research attention from the AI community. To generate an executable program consisting of math table operations to answer a question, state-of-the-art methods use retriever-generator pipeline. However, their retrieval results are static, while different generation steps may rely on sentences. attend retrieved information that is relevant each step, in this paper, we propose DyRRen, extended...
We propose DGST, a novel and simple Dual-Generator network architecture for text Style Transfer. Our model employs two generators only, does not rely on any discriminators or parallel corpus training. Both quantitative qualitative experiments the Yelp IMDb datasets show that our gives competitive performance compared to several strong baselines with more complicated designs.
Scenario-based question answering (SQA) has attracted an increasing research interest. Compared with the well-studied machine reading comprehension (MRC), SQA is a more challenging task: scenario may contain not only textual passage to read but also structured data like tables, i.e., tabular based (TSQA). AI applications of TSQA such as multiple-choice questions in high-school exams require synthesizing multiple cells and combining tables texts domain knowledge infer answers. To support...
This paper addresses the problem of learning multiple spoken language understanding (SLU) tasks that have overlapping sets slots. In such a scenario, it is possible to achieve better slot filling performance by simultaneously, as opposed them independently. We focus on presenting number simple multi-task algorithms for systems based semi-Markov CRFs, assuming knowledge shared Furthermore, we discuss an intradomain clustering method automatically discovers slots from training data. The...
Current treatment of recurrent glioblastoma multiforme (GBM) demands dose-intense temozolomide (TMZ), a prodrug 5-(3-methyltriazen-1-yl) imidazole-4-carboxamide (MTIC), based on the spontaneous hydrolysis TMZ at basic pH. However, how to control activity MTIC remains unknown, which poses particular challenge search reliable receptor. We reported that copper, for first time, is found recognize and bind in process degradation, means copper can play an important role enhancing bioavailability...
In this paper, we consider distributed optimization problems where <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$n$</tex-math></inline-formula> agents, each possessing a local cost function, collaboratively minimize the average of functions over connected network. To solve problem, propose random reshuffling (D-RR) algorithm that invokes (RR) update in agent. We show D-RR inherits favorable...