Towards Large-Scale Natural Language Inference with Distributional Semantics

Explore natural language inference and distributional semantics in this 50-minute conference talk by Jackie CK Cheung from the Center for Language & Speech Processing at Johns Hopkins University. Delve into automatic summarization techniques, patient record data analysis for depression, and the intersection of logical and distributional semantics. Learn about compositional distributional semantics and methods for evaluating these models using similarity judgments. Examine structure induction in distributional semantics, including Hidden Markov Models (HMM) and their applications in text analytics. Discover the potential of distributional semantic HMMs for summarization tasks, and analyze sample summaries produced by these models. Investigate a new evaluation framework focusing on argument structure invariance, and compare various distributional semantic models. Gain insights into opinion summarization, semantic knowledge induction, and the challenges of abstractive summarization in natural language processing.

Intro
Automatic Summarization
Patient Record Data for Depression
Natural Language Inference
Logical Semantics
Distributional Semantics (DS)
Compositional DS
Evaluating DS by Similarity Judgments
Structure Induction
Basic Structure of DSHMM
Text Analytics Conference 2010
Slot Induction Results
DSHMM for Summarization
Summarization Results
Sample Summary
A New Evaluation Framework
Argument Structure Invariance
Task Descriptions
Distributional Semantic Models
Evaluation Results
Opinion Summarization
Semantic Knowledge Induction
Abstractive Summarization