Tootfinder

USC professor Allison Marsh writes this delightful short article for IEEE Spectrum about Marvin Minsky and Seymour Papert's British counterparts in Cambridge and how, in their case, a document (the Lighthill Report) precipitated cuts in British AI research funding in ways similar to how funding was temporarily cut in the U.S. following MIT's publishing of Minsky and Papert's "Perceptrons" in 1969.
"Freddy the Robot Was the Fall Guy for British AI"

Freddy the Robot Was the Fall Guy for British AI

Its creator lost a power struggle that led to an AI winter in the 1970s

Building on the 90s, statistical n-gram language models, trained on vast text collections, became the backbone of NLP research. They fueled advancements in nearly all NLP techniques of the era, laying the groundwork for today's AI.
F. Jelinek (1997), Statistical Methods for Speech Recognition, MIT Press, Cambridge, MA
#NLP

Understanding the Error Sensitivity of Privacy-Aware Computing
Mat\'ias Mazzanti (University of Buenos Aires), Esteban Mocskos (University of Buenos Aires), Augusto Vega (IBM T. J. Watson Research Center), Pradip Bose (IBM T. J. Watson Research Center)
https://arxiv.org/abs/2506.07957

Understanding the Error Sensitivity of Privacy-Aware Computing
Homomorphic Encryption (HE) enables secure computation on encrypted data without decryption, allowing a great opportunity for privacy-preserving computation. In particular, domains such as healthcare, finance, and government, where data privacy and security are of utmost importance, can benefit from HE by enabling third-party computation and services on sensitive data. In other words, HE constitutes the "Holy Grail" of cryptography: data remains encrypted all the time, being protected while in …

New around here, hi!
For those who'd like a big old dose of "yes we can" energy,
please check out
"Words to Win By "
where every episode details a campaign we've won somewhere
and delves into the research, messaging & strategies used to do it.
https://wordstowinby-pod.com/

Words to Win By
Words to Win By, formerly Brave New Words, takes listeners on a journey around the globe with renowned communications researcher and campaign advisor Anat Shenker-Osorio as she unpacks real-world narrative shifts that led to real-world victories.

This https://arxiv.org/abs/2504.21398 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csIR_…

In a Few Words: Comparing Weak Supervision and LLMs for Short Query Intent Classification
User intent classification is an important task in information retrieval. Previously, user intents were classified manually and automatically; the latter helped to avoid hand labelling of large datasets. Recent studies explored whether LLMs can reliably determine user intent. However, researchers have recognized the limitations of using generative LLMs for classification tasks. In this study, we empirically compare user intent classification into informational, navigational, and transactional c…

"In Empire of #AI, journalist Karen Hao writes about the rise of #OpenAI and the impacts of AI around the world. Below is an extract from the book on the effects on Chile's mineral reserves and water resources."

Schaden durch Bürgergeldmissbrauch:
000 260 Mio. €
Sicherstellungen durch Steuerfahndungen:
002 500 Mio. €
Schätzung des Schadens aus #Steuerhinterziehung
0XX 000 Mio. € (zweistelliger Milliardenbetrag, Bundesrechnungshof)
200 000 Mio. € (Deutschen Steuer-Gewerkschaft)
"Der erwartete Personalabbau um ein Drittel bis 2030 trage zur Verschärf…

„Nehmen den Staat auf allen Ebenen aus“ - Report entlarvt Deutschlands wahre Sozialschmarotzer
Ein Bericht des Bundesrechnungshofs zeigt: Nicht Bürgergeld-Empfänger, sondern Steuerbetrüger kosten den Staat jährlich Milliarden.

For Zotero users, the Android app is finally out of closed beta. It supports saving papers directly from the browser or other apps, which takes several steps out of my workflow for increasing my TBR pile.
https://www.zotero.org/blog/zotero-for-android/

Oh dear -- within a day of hearing about this from another professor, I got my first LLM-generated email from an aspiring grad student. It mischaracterized my research (apparently I write lots of mathematical formulas, ha!) and ascribed papers to me that I never wrote, which are basically remixes of common words in my paper titles. If you do this, you will insult the people you want to impress.

'Failure Imminent': When LLMs In a Long-Running Vending Business Simulation Went Berserk
https://slashdot.org/story/25/05/31/2112240/failure-imminent-when-llms-in-a-long-running-vending-business-simulati…

'Failure Imminent': When LLMs In a Long-Running Vending Business Simulation Went Berserk - Slashdot
Long-time Slashdot reader lunchlady55 writes: A pair of researchers investigating the ability of LLMs to coherently operate a simulated vending machine business have recorded hilariously unhinged behavior in many of the current "advanced" LLMs. The LLMs were equipped with several "tools" (code the ...