Humanist Discussion Group

Humanist Archives: Oct. 29, 2021, 9:23 a.m. Humanist 35.331 - learning about machine learning

              Humanist Discussion Group, Vol. 35, No. 331.
        Department of Digital Humanities, University of Cologne
                      Hosted by DH-Cologne
                Submit to:

        Date: 2021-10-28 08:58:26+00:00
        From: Willard McCarty <>
        Subject: learning about machine learning

Allow me to recommend an online lecture by Charles Isbell, Peofessor at
Georgia Institute of Technology (U.S.), "You Can’t Escape
Hyperparameters and Latent Variables: Machine Learning as a Software
Engineering Enterprise", delivered for the Neural Information Processing
Systems conference last year. Don't be put off by the title -- it's a
hugely entertaining and intelligent attempt to dig into the
inseparability of algorithmic machine-learning systems from the world in
which we all live. It may be found at:

To get to the lecture, you will have to advance the video forward to
26:18, at which Isbell is introduced and his show begins.

During the lecture, and even more now, I found myself wanting the book
that he could obviously write if he took the time that he's unlikely to
have to write it. One of his included commentators remarks on how
important it is to have "the right people in the room", esp when
important decisions are made (e.g. about design of machine-learning
software that recommends whether prison-sentences are extended). It
seems to me that we should be among them, but that we are unlikely to be
motivated to be there unless the inclusiveness he and others talk about
extends to disciplines beyond the computational sciences. We need close,
careful and critical argument written out so it can be considered, as
well as spoken aloud so that it stirrs someone like me to recommend it.

Anyhow, the online lecture is more than worth the candle.


Willard McCarty,
Professor emeritus, King's College London;
Editor, Interdisciplinary Science Reviews;  Humanist

Unsubscribe at:
List posts to:
List info and archives at at:
Listmember interface at:
Subscribe at: