Note: I have added a (***) to blogs which have been viewed 1000+ times
Building NLG Systems
- Does Quality Matter in Training Data?
- Embedding Machine Learning in a Rules-Based NLG System
- How do I Build an NLG System: Requirements and Corpora
- How do I Build an NLG System: Testing and Quality Assurance
- How do I Build an NLG System: Tools? (***)
- Natural Language Generation and Machine Learning (***)
- NLG vs Templates: Levels of Sophistication in Generating Text
- Should I Use Deep Learning? (Latest blog!)
- Testing Multiple Hypotheses
- The story of simplenlg
- You Need to Understand your Corpora! The Weathergov Example
Evaluating NLG Systems
- A Consumer Perspective on Evaluation
- Evaluation in Medicine and NLG/NLP
- How to do an NLG Evaluation: Metrics
- How to do an NLG Evaluation: Human Ratings in Artificial Context
- How to do an NLG Evaluation: Human Ratings in Real-World Context
- How to do an NLG Evaluation: Task-Based (Extrinsic) Performance in Real-World Context
- Is BLEU valid? First observations and concerns
- My Guidelines for Evaluating AI Systems
- Research Ethics of A/B Testing
- Regression to Mean
- Study Design for Systematic Review of BLEU Validity: Comments Welcome!
- Types of NLG Evaluation: Which is Right for Me?
Other NLG Topics
- Academic Teaching vs Commercial Training Courses
- Commercial and Academic Perspectives on NLG (and AI?)
- Good Papers are Hard to Publish
- Exciting NLG Research Topics (June 2017)
- How do Users React to NLG?
- Managing Research Projects is Painful but Necessary
- My PhD Students: Where Are They Now (June 2017)
- Non-Experts Struggle with Information Graphics
- Text or Graphics?
- Response to Goldberg’s Blog on Deep Learning for NLG
- Why isnt there More Open-Source NLG Software? (***)
- Writing NLG Pages for Wikipedia