@alan not really an independent position...

I agree with the missed opportunity to sell "economically harnessing constantly evolving data" as a business case. Although they did pitch on "data lakes"!

@ohyran when I joined CWI (Amsterdam, NL) some 20 years ago all CS staff including the secretaries ran on Linux and this has never changed!

Here's Paul Baran's RAND-published September 1962 justification for distributed communications networks summed up in a single chart. It's also the same paper where the famous "centralized vs decentralized vs distributed" triptych of graphs comes from. You still see this exact diagram, uncited, in modern presentations on the decentralized/distributed web.

Full paper: rand.org/content/dam/rand/pubs

Interesting CHI 2019 paper:

Google Top Stories box provides most of the attention and traffic to a handful of news organizations.

I.e., the algorithms concentrate attention to a few sources, and that has implications in ideological diversity.


psst, markov chains are the solarpunk GPT-2, pass it on

[ICYMI on twitter] The Deep Learning track at will feature two tasks—a DOCUMENT RANKING task and a PASSAGE RANKING task—both with large training datasets.

Approximate dataset stats: 367K training queries + 3.2M documents for the document ranking task; and 500K training queries + 8.8M passages for the passage ranking task.

Official guidelines: microsoft.github.io/TREC-2019-
Coordinators: Nick Craswell, Bhaskar Mitra, Emine Yilmaz, Daniel Campos.

Please share.

Explain Yourself Leveraging Language Models for Commonsense Reasoning - arxiv.org/abs/1906.02361

Personal best-of:

The Raconteurs
Sports Team
Kurt Vile & The Violators
Snail Mail
Personal Trainer
The Beths
Miya Folick
Bon Iver
Indian Askin
Mac de Marco
Caroline Rose
Ron Gallo

Aldous Harding: I wish I had seen more, what I saw was very good too - but only last two songs.

Big Thief: I hope to see at another occasion!

someday I want to make a book that just collects diagrams from machine learning papers (especially nlproc-related)—they're always so elaborate and beautiful and bizarre (this one's from arxiv.org/pdf/1808.10805.pdf)

PS: slaat krantenartikel over (mis-/ge-) bruik verschoningsrecht.

Ik zeg CC de jurist op alles!

Lekker bezig Shell

Show more

The "unofficial" Information Retrieval Mastodon Instance.

Goal: Make idf.social a viable and valuable social space for anyone working in Information Retrieval and related scientific research.

Everyone welcome but expect some level of geekiness on the instance and federated timelines.