Tootfinder

The Curious Case of Visual Grounding: Different Effects for Speech- and Text-based Language Encoders
Adrian Sauter, Willem Zuidema, Marianne de Heer Kloots
https://arxiv.org/abs/2509.15837

The Curious Case of Visual Grounding: Different Effects for Speech- and Text-based Language Encoders
How does visual information included in training affect language processing in audio- and text-based deep learning models? We explore how such visual grounding affects model-internal representations of words, and find substantially different effects in speech- vs. text-based language encoders. Firstly, global representational comparisons reveal that visual grounding increases alignment between representations of spoken and written language, but this effect seems mainly driven by enhanced encodi…

AFC West Power Rankings: Can Anyone Catch The Chiefs Now? https://raiderramble.com/2025/11/01/afc-west-power-rankings-can-anyone-catch-the-chiefs-now/

I do like #NextCloud but sometimes it can drive me crazy. For example, fresh update, hundreds of errors with astoundingly unhelpful messages. Consider:
chmod(): Operation not permitted at /var/www/nextcloud/apps/recognize/lib/Migration/InstallDeps.php#169
ok, so what do we find at line 169? a call to chmod() on presumably a filename, wrapped in an exception Catch.
I realize this is very much non-trivial software with oodles of hours of difficult work by so many kind people, but, really, how much effort would it have been to include the errant target FILENAME in the exception message?
Similarly, a deprecation warning is, by definition, a Warning, not an error.
None of this affects my using Nextcloud, but it does really clutter the log and who knows, maybe some of these Errors are actually important?

How long does it take an Elephant Random Walk to forget its training
Zheng Fang
https://arxiv.org/abs/2509.15049 https://arxiv.org/pdf/2509.15049

How long does it take an Elephant Random Walk to forget its training
We study how conditioning affects the long-term behavior of the Elephant Random Walk, focusing on its first return to the origin, namely, the time it takes to forget the training. When the walker is conditioned to be at position $n$ at time $n$, this return time scales as $n^{(4-4p)/(3-4p)}$ in the diffusive regime, and grows exponentially in the critical regime.

Speak Your Mind: The Speech Continuation Task as a Probe of Voice-Based Model Bias
Shree Harsha Bokkahalli Satish, Harm Lameris, Olivier Perrotin, Gustav Eje Henter, \'Eva Sz\'ekely
https://arxiv.org/abs/2509.22061

Speak Your Mind: The Speech Continuation Task as a Probe of Voice-Based Model Bias
Speech Continuation (SC) is the task of generating a coherent extension of a spoken prompt while preserving both semantic context and speaker identity. Because SC is constrained to a single audio stream, it offers a more direct setting for probing biases in speech foundation models than dialogue does. In this work we present the first systematic evaluation of bias in SC, investigating how gender and phonation type (breathy, creaky, end-creak) affect continuation behaviour. We evaluate three rec…