[Topic-models] LDA paper by Blei - some more questions

Mateusz Berezecki mateuszb at gmail.com
Sun Nov 23 06:57:43 EST 2008


Hi Veena,

Thanks a lot for the response.

On Sun, Nov 23, 2008 at 7:27 AM, Veena T <veenat2005 at gmail.com> wrote:

>>So the first step of the algorithm explicitly relies on the inference
>>phase being completed as the variational parameters are valid
>>(optimized) only after inference phase (?).
>
> See, here I have a doubt as to your undertsadning of the word 'inference'. I
> feel that you are referring to the word in the context of infering some
> results from the model for the unseen examples. Is that true?

I guess you nailed it. I was more than convinced that inference means
predicting/infering something from the trained model. What is
"inference" supposed to mean in the context of the LDA paper then?

I am, to be honest, really surprised I did not understand the meaning correctly.

Mateusz


More information about the Topic-models mailing list