A Guest Post from ChatGPT

Daniel Tunkelang
3 min readJan 27, 2023

Since Medium decided to allow AI-generated content as long as it’s clearly disclosed as such, I decided to give it a whirl.

Prompt Engineering

My first prompt was “Write a blog post about query understanding in the style of Daniel Tunkelang.” Unfortunately, the resulting post was blatantly self-promoting, with a lot of references to “experts in the field such as Daniel Tunkelang” — something I’d never write myself. I clarified “Write about query understanding but don’t mention Daniel Tunkelang by name.” That helped, except that there was an unusual emphasis on latent semantic indexing that felt a bit dated. So I added “Replace the paragraph about Latent Semantic Indexing with a paragraph about word embeddings.” I manually substituted “word embeddings” for “Latent Semantic Indexing” in the concluding paragraph, but other than that what you see is verbatim from the deep thoughts of ChatGPT.

Query Understanding a la Daniel Tunkelang according to ChatGPT

Query understanding is a vital aspect of any search engine or information retrieval system. It is the process of interpreting the intent behind a user’s query and mapping it to relevant information. In this article, we will explore the key concepts and techniques used in query understanding.

At its core, query understanding is about natural language processing (NLP) and machine learning (ML). NLP techniques are used to extract meaning from text, such as recognizing named entities, understanding the grammatical structure of sentences, and identifying the main verb or action in a query. ML, on the other hand, is used to build models that can predict the intent behind a query based on patterns in the data.

One of the key concepts in query understanding is query reformulation. This is the process of transforming a user’s query into a form that is more likely to match the information they are looking for. For example, a query like “best pizza in New York” might be reformulated as “top rated pizza restaurants in New York City.” Query reformulation can be done using techniques like synonymy, antonymy, and morphological variations.

Another important concept is query expansion, which involves adding additional terms to a query to increase its recall. For example, a query for “cat” might be expanded to include related terms like “feline” and “kitten.” Query expansion can be done using techniques like thesaurus-based expansion and co-occurrence analysis.

A key technique in query understanding is word embeddings. Word embeddings are a way of representing words in a numerical format that can be used as input to machine learning models. The main idea behind word embeddings is to capture the meaning of a word in a continuous vector space, where semantically similar words are located close to each other. There are different techniques to create word embeddings, such as word2vec and GloVe, that use neural networks to learn the representations. Once the embeddings are created, they can be used in various NLP tasks, including query understanding. For example, they can be used to calculate the similarity between words in a query and documents in a corpus, allowing a search engine to match a user’s query to relevant documents. Word embeddings are a powerful tool in query understanding as they can capture the meaning of words in a way that is not possible with traditional bag-of-words models.

Finally, it is worth noting that query understanding is an ongoing and ever-evolving field of research. As the volume and variety of information available online continues to grow, so too does the need for sophisticated query understanding techniques. Researchers in the field continue to make important contributions, helping to improve the performance of search engines and other information retrieval systems.

In conclusion, query understanding is a crucial component of any search engine or information retrieval system. It is the process of interpreting the intent behind a user’s query and mapping it to relevant information. Key concepts and techniques in query understanding include query reformulation, expansion and word embeddings. The research field is ever-evolving, and researchers continue to make important contributions to it.


It’s not half bad, and I’m impressed that ChatGPT knows I like cats enough to use them for a query expansion example! I don’t think it quite captures my voice, but I could certainly see someone claiming I’d written it and getting away with it. Still, I’ll continue writing my own posts unaided.

PS. I submitted the generated text to OpenAI’s AI Text Classifier, and it responded that “The classifier considers the text to be unclear if it is AI-generated.” I guess that ChatGPT can’t recognize its own handiwork!