Cross-Domain Multi-Event Tracking via CO-PMHT
Source
Evernote/IFTTT Feedly/Cross-Domain Multi-Event Tracking via CO-PMHT.md
Summary
이 문서는 인터넷 상의 다중 이벤트 효율적 모니터링을 위해 제안된 CO-PMHT 알고리즘을 소개합니다. Google News(텍스트)와 Flickr(이미지)라는 두 다른 도메인에서 수집된 대규모 데이터를 활용하여, 시맨틱 후验 확률(Semantic Posterior Probability)을 브릿지로 삼아 도메인 간 격차를 해소하고 다중 이벤트 추적 성능을 향상시키는 방법을 다룹니다.
Key Points
- CO-PMHT(CO-Probabilistic Multi-Hypothesis Tracking) 알고리즘 제안
- Google News와 Flickr를 활용한 크로스 도메인 다중 이벤트 추적
- 시맨틱 후验 확률을 통해 도메인 간 격차(Domain Gap) 해소
- 이벤트의 정보적 요약 및 시간적 진화 추이 파악 가능
Related
-
Social Event Classification via Boosted Multimodal Supervised Latent Dirichlet Allocation
-
Unsupervised Spatial Event Detection in Targeted Domains with Applications to Civil Unrest Modeling
-
Feature Ensemble Plus Sample Selection: Domain Adaptation for Sentiment Classification
-
Intel Research: Context Awareness - Social Proximity Detection
-
Moment-Based Spectral Analysis of Large-Scale Networks Using Local Structural Information
-
Improved Domain Adaptation for Statistical Machine Translation
-
Active Learning through Adaptive Heterogeneous Ensembling (AHE)
-
Enlisting the Ghost: Modeling Empty Categories for Machine Translation
-
An Unsupervised Feature Selection Framework for Social Media Data
-
Protecting Sensitive Labels in Social Network Data Anonymization
-
Efficient Multiview Maintenance under Insertion in Huge Social Networks
-
사회적·공간적 근접성을 활용한 공동 검색 (Joint Search by Social and Spatial Proximity)
-
A Hamming Embedding Kernel with Informative Bag-of-Visual Words for Video Semantic Indexing
-
Smooth Nonnegative Matrix Factorization for Unsupervised Audiovisual Document Structuring
-
Weakly Supervised Learning of Object Segmentations from Web-Scale Video
-
Information-Theoretic Outlier Detection for Large-Scale Categorical Data
-
Efficient Closed-Form Solution to Generalized Boundary Detection
-
Regularized Latent Semantic Indexing (RLSI): 대규모 토픽 모델링을 위한 새로운 접근법
-
Scalable Decipherment for Machine Translation via Hash Sampling
-
Efficient Inference and Structured Learning for Semantic Role Labeling
-
Continuous Birdsong Recognition Using Gaussian Mixture Modeling of Image Shape Features
-
Token and Type Constraints for Cross-Lingual Part-of-Speech Tagging
-
Similarity-based Clustering by Left-Stochastic Matrix Factorization