Haifeng Chen
A5100456776- Anomaly Detection Techniques and Applications
- Network Security and Intrusion Detection
- Software System Performance and Reliability
- Particle physics theoretical and experimental studies
- Quantum Chromodynamics and Particle Interactions
- High-Energy Particle Collisions Research
- Time Series Analysis and Forecasting
- Advanced Graph Neural Networks
- Complex Network Analysis Techniques
- Topic Modeling
- Cloud Computing and Resource Management
- EEG and Brain-Computer Interfaces
- Data Stream Mining Techniques
- Advanced Malware Detection Techniques
- Fault Detection and Control Systems
- Recommender Systems and Techniques
- Natural Language Processing Techniques
- Domain Adaptation and Few-Shot Learning
- Software Engineering Research
- Network Traffic and Congestion Control
- Image Processing Techniques and Applications
- Semantic Web and Ontologies
- Human Mobility and Location-Based Analysis
- Machine Learning and Data Classification
- Software Reliability and Analysis Research
Eye & ENT Hospital of Fudan University
2025
Chinese Academy of Medical Sciences & Peking Union Medical College
2025
University of Chinese Academy of Sciences
2021-2025
Institute of High Energy Physics
2014-2025
NEC (United States)
2009-2024
Shaanxi University of Science and Technology
2015-2024
Xiamen University
2024
Xi’an University of Posts and Telecommunications
2024
Taizhou University
2024
Halmstad University
2024
Nowadays, multivariate time series data are increasingly collected in various real world systems, e.g., power plants, wearable devices, etc. Anomaly detection and diagnosis refer to identifying abnormal status certain steps pinpointing the root causes. Building such a system, however, is challenging since it not only requires capture temporal dependency each series, but also need encode inter-correlations between different pairs of series. In addition, system should be robust noise provide...
Massive and dynamic networks arise in many practical applications such as social media, security public health. Given an evolutionary network, it is crucial to detect structural anomalies, vertices edges whose "behaviors'' deviate from underlying majority of the a real-time fashion. Recently, network embedding has proven powerful tool learning low-dimensional representations that can capture preserve structure. However, most existing approaches are designed for static networks, thus may not...
The problem of network representation learning, also known as embedding, arises in many machine learning tasks assuming that there exist a small number variabilities the vertex representations which can capture "semantics" original structure. Most existing embedding models, with shallow or deep architectures, learn from sampled sequences such low-dimensional embeddings preserve locality property and/or global reconstruction capability. resultant representations, however, are difficult for...
Detecting anomalies in dynamic graphs is a vital task, with numerous practical applications areas such as security, finance, and social media. Existing network embedding based methods have mostly focused on learning good node representations, whereas largely ignoring the subgraph structural changes related to target nodes given time window. In this paper, we propose StrGNN, an end-to-end temporal Graph Neural Network model for detecting anomalous edges graphs. particular, first extract h-hop...
Detecting abnormal activities in real-world surveillance videos is an important yet challenging task as the prior knowledge about video anomalies usually limited or unavailable. Despite that many approaches have been developed to resolve this problem, few of them can capture normal spatio-temporal patterns effectively and efficiently. Moreover, existing works seldom explicitly consider local consistency at frame level global coherence temporal dynamics sequences. To end, we propose...
The Born cross sections of the process $e^{+}e^{-}\to D^{*0}D^{*-}\pi^{+}$ at center-of-mass energies from 4.189 to 4.951 GeV are measured for first time. data samples used correspond an integrated luminosity $17.9\,{\rm fb}^{-1}$ and were collected by BESIII detector operating BEPCII storage ring. Three enhancements around 4.20, 4.47 4.67 visible. resonances have masses $4209.6\pm4.7\pm5.9\,{\rm MeV}/c^{2}$, $4469.1\pm26.2\pm3.6\,{\rm MeV}/c^{2}$ $4675.3\pm29.5\pm3.5\,{\rm widths...
Temporal data, notably time series and spatio-temporal are prevalent in real-world applications. They capture dynamic system measurements produced vast quantities by both physical virtual sensors. Analyzing these data types is vital to harnessing the rich information they encompass thus benefits a wide range of downstream tasks. Recent advances large language other foundational models have spurred increased use mining. Such methodologies not only enable enhanced pattern recognition reasoning...
The hindrances to the adoption of public cloud computing services include service reliability, data security and privacy, regulation compliant requirements, so on. To address those concerns, we propose a hybrid model which users may adopt as viable cost-saving methodology make best use along with their privately-owned (legacy) centers. As core this model, an intelligent workload factoring is designed for proactive management. It enables federation between on- off-premise infrastructures...
Modern world has witnessed a dramatic increase in our ability to collect, transmit and distribute real-time monitoring surveillance data from large-scale information systems cyber-physical systems. Detecting system anomalies thus attracts significant amount of interest many fields such as security, fault management, industrial optimization. Recently, invariant network shown be powerful way characterizing complex behaviours. In the network, node represents component an edge indicates stable,...
Dynamic networks are ubiquitous. Link prediction in dynamic has attracted tremendous research interests. Many models have been developed to predict links that may emerge the immediate future from past evolution of networks. There two key factors: 1) a node is more likely form link near with another within its close proximity, rather than random node; 2) network usually evolves smoothly. Existing approaches seldom unify these factors strive for spatial and temporal consistency network. To...
Network embedding aims to learn a low-dimensional vector representation for each node in the social and information networks, with constraint preserve network structures. Most existing methods focus on single embedding, ignoring relationship between multiple networks. In many real-world applications, however, networks may contain complementary information, which can lead further refined embeddings. Thus, this paper, we propose novel multi-network method, DMNE. DMNE is flexible. It allows...
The J/ψ→Ξ0¯Ξ0 process and subsequent decays are investigated using (10087±44)×106 J/ψ events collected at the BESIII experiment. decay parameters of Ξ0 ¯Ξ0 simultaneously measured to be αΞ=−0.3750±0.0034±0.0016, ¯αΞ=0.3790±0.0034±0.0021, ϕΞ=0.0051±0.0096±0.0018 rad, ¯ϕΞ=−0.0053±0.0097±0.0019 rad with unprecedented accuracies, where first second uncertainties statistical systematic, respectively. most precise values for CP asymmetry observables obtained AΞCP=(−5.4±6.5±3.1)×10−3...
With the prevalence of Internet services and increase their complexity, there is a growing need to improve operational reliability availability. While large amount monitoring data can be collected from systems for fault analysis, it hard correlate this effectively across distributed observation time. In paper, we analyze mass characteristics user requests propose novel approach model track transaction flow dynamics detection in complex information systems. We measure intensity at multiple...
Due to their growing complexity, it becomes extremely difficult detect and isolate faults in complex systems. While large amount of monitoring data can be collected from such systems for fault analysis, one challenge is how correlate the effectively across distributed observation time. Much internal reacts volume user requests accordingly when flow through In this paper, we use Gaussian mixture models characterize probabilistic correlation between flow-intensities measured at multiple...
Large amount of monitoring data can be collected from distributed systems as the observables to analyze system behaviors. However, without reasonable models characterize systems, we hardly interpret such effectively for management. In this paper, a new concept named flow intensity is introduced measure with which internal reacts volume user requests in transaction systems. We propose novel approach automatically model and search relationships between intensities measured at various points...
Recently, recommender systems have been able to emit substantially improved recommendations by leveraging user-provided reviews. Existing methods typically merge all reviews of a given user (item) into long document, and then process item documents in the same manner. In practice, however, these two sets are notably different: users' reflect variety items that they bought hence very heterogeneous their topics, while an item's pertain only single thus topically homogeneous. this work, we...
Discrete event sequences are ubiquitous, such as an ordered series of process interactions in Information and Communication Technology systems. Recent years have witnessed increasing efforts detecting anomalies with discrete sequences. However, it remains extremely difficult task due to several intrinsic challenges including data imbalance issues, property the events, sequential nature data. To address these challenges, this paper, we propose OC4Seq, a multi-scale one-class recurrent neural...
The ${\mathrm{\ensuremath{\Xi}}}^{0}$ asymmetry parameters are measured using entangled quantum ${\mathrm{\ensuremath{\Xi}}}^{0}\ensuremath{-}{\overline{\mathrm{\ensuremath{\Xi}}}}^{0}$ pairs from a sample of $(448.1\ifmmode\pm\else\textpm\fi{}2.9)\ifmmode\times\else\texttimes\fi{}{10}^{6}\text{ }\text{ }\ensuremath{\psi}(3686)$ events collected with the BESIII detector at BEPCII. relative phase between transition amplitudes...
Using data samples with an integrated luminosity of 5.85 fb−1 collected at center-of-mass energies from 4.61 to 4.95 GeV the BESIII detector operating BEPCII storage ring, we measure cross section for process e+e−→K+K−J/ψ. A new resonance a mass M=4708−15+17±21 MeV/c2 and width Γ=126−23+27±30 MeV is observed in energy-dependent line shape e+e−→K+K−J/ψ significance over 5σ. The K+J/ψ system also investigated search charged charmoniumlike states, but no significant Zcs+ states are observed....
One of the main goals studying semileptonic decays in flavor physics is to gain a better understanding hadronic transitions nonperturbative region Quantum Chromodynamics. This involves measuring Cabibbo-Kobayashi-Maskawa matrix elements, form factors, and comparing them with theoretical predictions. We report first study decay <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" display="inline"><a:mrow><a:msup><a:mrow><a:mi>D</a:mi></a:mrow><a:mrow><a:mn>0</a:mn></a:mrow></a:msup><a:mo...
The <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" display="inline"><a:mi>C</a:mi><a:mi>P</a:mi></a:math>-even fractions (<c:math xmlns:c="http://www.w3.org/1998/Math/MathML" display="inline"><c:msub><c:mi>F</c:mi><c:mo>+</c:mo></c:msub></c:math>) of the decays <e:math xmlns:e="http://www.w3.org/1998/Math/MathML" display="inline"><e:mrow><e:msup><e:mrow><e:mi>D</e:mi></e:mrow><e:mrow><e:mn>0</e:mn></e:mrow></e:msup><e:mo...
Chatbot-based multimodal AI holds promise for collecting medical histories and diagnosing ophthalmic diseases using textual imaging data. This study developed evaluated the ChatGPT-powered Intelligent Ophthalmic Multimodal Interactive Diagnostic System (IOMIDS) to enable patient self-diagnosis self-triage. IOMIDS included a text model three models (text + slit-lamp, smartphone, slit-lamp smartphone). The performance was through two-stage cross-sectional across centers involving 10...
Using <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" display="inline"><a:mrow><a:mo stretchy="false">(</a:mo><a:mn>2712.4</a:mn><a:mo>±</a:mo><a:mn>14.3</a:mn><a:mo stretchy="false">)</a:mo><a:mo>×</a:mo><a:msup><a:mrow><a:mn>10</a:mn></a:mrow><a:mrow><a:mn>6</a:mn></a:mrow></a:msup><a:mi>ψ</a:mi><a:mo stretchy="false">(</a:mo><a:mn>3686</a:mn><a:mo stretchy="false">)</a:mo></a:mrow></a:math> events collected by the BESIII detector operating at BEPCII collider, we present first...