Ehud Reiter's Blog

Desk rejection is an unfortunate necessity

Cycling in Northern Ireland

AI for Healthcare: Great benchmarks but minimal impact

Memories of being an NL researcher in 1990

What is the purpose of ACL conferences?

Future of NLG evaluation

I am worried by NLP research culture

Software engineering of prompts

AI and CS Teaching

Achieving my vision of personal AI health assistants

Real-world safety and harms from patient-facing LLMs

Comparing performance of LLMs is not very interesting

Please follow the rules for ARR/ACL papers

Questions from readers of my book

Dont ignore omissions!

My Eureka moments in research

Lets use AI to help people manage illness

Retirement Plans: Travel and some academics

Do a sanity check on your experiments

Do LLMs cheat on benchmarks

Hard to Change Poor Research Culture

Understanding what users want from NLG

Most common uses of AI in Healthcare

Good diagrams for research papers

Reflections on blogging

Defining hallucination is not straightforward

Encouraging safer driving with NLG apps

I hate pay-to-publish

More on evaluating impact

Cycling in Netherlands

Patients want to know what information an AI model considers

The Aberdeen NLP Research Group

Key messages from my NLG book

Even good leaderboards may not be useful, because they are gamed