Explore the challenges and advancements in automatic summarization techniques for diverse information sources in this comprehensive lecture by Jessica Ouyang from Columbia University. Delve into adapting summarization methods for online personal narratives and cross-lingual content in low-resource languages. Learn about data collection, content selection tasks, sentence-level metrics, and heuristics used in summarization. Examine experiments, classification tasks, and real-world evaluations for Arabic summarization. Gain insights into ongoing work in multimodal summarization, ethical considerations in NLP, and interannotator agreement. Discover how these techniques can improve information accessibility across various genres and languages.
Adapting Automatic Summarization to New Sources of Information - 2019
Center for Language & Speech Processing(CLSP), JHU via YouTube
Overview
Syllabus
Intro
Challenges
Outline
Define
Data Collection
Agreement
Content Selection Tas
Sentence Level Metrics
Heuristics
Experiments
Classification Task
More Data
Material
Crosslingual Summarization
Example from Material
Approach
Machine Translation
Real World Evaluation
Arabic Evaluation
Ongoing Work
multimodal summarization
branching out
ethical NLP
internotator agreement
agree vs disagree
other summary
Taught by
Center for Language & Speech Processing(CLSP), JHU
