Zhen Li

Understanding Hidden Memories of Recurrent Neural Networks

2017

Recurrent neural networks (RNNs) have been successfully applied to various natural language processing (NLP) tasks and achieved better results than conventional methods. However, the lack of understanding of the mechanisms behind their effectiveness limits further …

Open Entity Semantic Relation Extraction in Big Data Environment

2020

Aiming at the problem that traditional entity relationship extraction needs to define a comprehensive entity relationship type system, poor portability, and the inability to process massive Internet texts, an open Chinese entity relationship extraction method …

Inverse Problem of Nonlinear Schrödinger Equation as Learning of Convolutional Neural Network

2021 · arXiv (Cornell University)

In this work, we use an explainable convolutional neural network (NLS-Net) to solve an inverse problem of the nonlinear Schrödinger equation, which is widely used in fiber-optic communications. The landscape and minimizers of the non-convex …

CO2Sum:Contrastive Learning for Factual-Consistent Abstractive Summarization

2021 · arXiv (Cornell University)

Generating factual-consistent summaries is a challenging task for abstractive summarization. Previous works mainly encode factual information or perform post-correct/rank after decoding. In this paper, we provide a factual-consistent solution from the perspective of contrastive learning, …

Old Wine in A New Bottle: A Homogeneous Fraud Sites Discovery Framework

2021 · 2021 7th International Conference on Computer and Communications (ICCC)

Internet fraud is a serious threat to the normal order of the Internet, endangering the personal and property safety of its users. Nowadays in China, the most rampant forms of Internet fraud include false winning, …

Feature Augmentation Reconstruction Network for Few-Shot Image Classification

2023

In recent years, deep learning has achieved significant success in various computer vision tasks, notably in image classification. Nonetheless, labeled data is frequently insufficient in real-world applications, presenting significant impediments to the efficacy of deep …

RankMatch: Fostering Confidence and Consistency in Learning with Noisy Labels

2023

Learning with noisy labels (LNL) is one of the most important and challenging problems in weakly-supervised learning. Recent advances adopt the sample selection strategy to mitigate the interference of noisy labels and use small-loss criteria …

Research and Application of Forecasting Based on ARIMA and ANOVA Models

2023

In this paper, a prediction model was built by ARIMA model and ANOVA based on the WordLE data. An ARIMA (1,1,0) prediction model was constructed using the historical data, and the number of results reported …

OSN Bots Traffic Transformer : MAE-Based Multimodal Social Bots Behavior Pattern Mining

2024

In recent years, online social networks (OSN) have rapidly gained popularity worldwide, becoming important platforms for information dissemination. Cyber manipulators use OSN bots to disseminate harmful information and manipulate public opinion, which can engage in …

HDFG: Ethereum Smart Contract Honeypot Detection Based on Pre-Training Techniques

2025

In recent years, a new fraud method, namely smart contract honeypots, has emerged on the famous blockchain platform Ethereum. The difference from smart contract vulnerabilities is that the contract honeypot essentially has no vulnerabilities, luring …

Incomplete Multi-View Drug Recommendation via Multi-Level Representation Learning and Curriculum Learning

2025

The drug recommendation task aims to provide effective and safe prescription decision support for clinical treatment based on patients' past Electronic Health Records (EHR). However, the prevalent phenomenon of missing views in multi-source heterogeneous EHR …

Modeling and Controlling Cyber-Physical Systems Based on Hybrid Turing Machine and Reinforcement Learning

2025 · International Journal of Software Engineering and Knowledge Engineering

This paper proposes a Hybrid Turing Machine (HTM) framework for modeling and controlling cyber-physical systems (CPS), which integrates discrete symbolic computation with continuous physical evolution. In HTM, each symbolic configuration is associated with real-valued dynamics …