Tootfinder

Scaling Spoken Language Models with Syllabic Speech Tokenization
Nicholas Lee, Cheol Jun Cho, Alan W Black, Gopala K. Anumanchipalli
https://arxiv.org/abs/2509.26634 https://

Scaling Spoken Language Models with Syllabic Speech Tokenization
Spoken language models (SLMs) typically discretize speech into high-frame-rate tokens extracted from SSL speech models. As the most successful LMs are based on the Transformer architecture, processing these long token streams with self-attention is expensive, as attention scales quadratically with sequence length. A recent SSL work introduces acoustic tokenization of speech at the syllable level, which is more interpretable and potentially more scalable with significant compression in token len…

Are there any admins on social.tchncs.de.? Is anonymously spreading lies about over 30 families in Gaza allowed on this instance?
This anonymous coward has been chased off of every server he’s been on and yet he’s still here, still attacking people suffering from genocide and famine.
(And yes, some folks copy posts from others. We’ve asked them not to. But they’re trying to survive genocide, not win social media awards.)
Also (see his second post) we’re not attempting to r…

Linux Is Best (@Linux@social.tchncs.de)
Attached: 2 images 1 of 2 First, let me say: what’s happening in Gaza is absolutely awful. But what’s also awful is that there are dishonest people taking advantage of others during a crisis. In one photo, from Facebook (my job, where I work), we had a woman claiming to be a 36-year-old mother of two children, ages 4 and 7. She spaced out the URL in her post, hoping it wouldn’t get flagged by spam filters (it did). The very next day, someone on the Fediverse posted the exact same messa…

»Decreasing Certificate Lifetimes to 45 Days:
Let’s Encrypt will be reducing the validity period of the certificates we issue. We currently issue certificates valid for 90 days, which will be cut in half to 45 days by 2028.«
The @… halving the validity of the keys makes the web more secure (not more anonymous).
🔐

Decreasing Certificate Lifetimes to 45 Days
Let’s Encrypt will be reducing the validity period of the certificates we issue. We currently issue certificates valid for 90 days, which will be cut in half to 45 days by 2028. This change is being made along with the rest of the industry, as required by the CA/Browser Forum Baseline Requirements, which set the technical requirements that we must follow. All publicly-trusted Certificate Authorities like Let’s Encrypt will be making similar changes. Reducing how long certificates are valid …

Fascinating story from a software dev Fedi friend, shared with permission to keep it anonymous:
❝A couple of days ago, I had an experience at work that made me understand one of the reasons why the chasm of opinion about LLMs is so deep and wide.
My department mostly does fiddly lowlevel work, [close to hardware]. A few of us don't use LLMs at all, a few use them sparingly, and one member is absolutely all-in. So during one of our morning meetings he suddenly started going off on a deeply disturbing diatribe about how we need to treat the LLMs “like slaves”.❞
1/

Anonymous Quantum Tokens with Classical Verification
Dmytro Gavinsky, Dar Gilboa, Siddhartha Jain, Dmitri Maslov, Jarrod R. McClean
https://arxiv.org/abs/2510.06212 https://

Anonymous Quantum Tokens with Classical Verification
The no-cloning theorem can be used as a basis for quantum money constructions which guarantee unconditionally unforgeable currency. Existing schemes, however, either (i) require long-term quantum memory and quantum communication between the user and the bank in order to verify the validity of a bill or (ii) fail to protect user privacy due to the uniqueness of each bill issued by the bank, which can allow its usage to be tracked. We introduce a construction of single-use quantum money that give…

Remember, remember! The 5th of November. #anonymous #VforVendetta #3Dprinting

This young producer is a friend of mine, and has finally been persuaded to release his very good materials. :) [albeit in as low-key, anonymous as possible a way, it seems. 😄]
I find it really hallucinogenic and special - in a super fragmented, "implied music"-but-still-groove-based kind of way.
https://www.youtube.…

Investigation of perception inconsistency in speaker embedding for asynchronous voice anonymization
Rui Wang, Liping Chen, Kong Aik Lee, Zhengpeng Zha, Zhenhua Ling
https://arxiv.org/abs/2510.05718

Investigation of perception inconsistency in speaker embedding for asynchronous voice anonymization
Given the speech generation framework that represents the speaker attribute with an embedding vector, asynchronous voice anonymization can be achieved by modifying the speaker embedding derived from the original speech. However, the inconsistency between machine and human perceptions of the speaker attribute within the speaker embedding remains unexplored, limiting its performance in asynchronous voice anonymization. To this end, this study investigates this inconsistency via modifications to s…

AV-EMO-Reasoning: Benchmarking Emotional Reasoning Capabilities in Omni-modal LLMS with Audio-visual Cues
Krish Patel, Dingkun Zhou, Ajay Kankipati, Akshaj Gupta, Zeyi Austin Li, Mohul Shukla, Vibhor Narang, Sara Kofman, Zongli Ye, Grace Wang, Xiaoyu Shi, Tingle Li, Guan-Ting Lin, Kan Jen Cheng, Huang-Cheng Chou, Jiachen Lian, Gopala Anumanchipalli
https://

AV-EMO-Reasoning: Benchmarking Emotional Reasoning Capabilities in Omni-modal LLMS with Audio-visual Cues
Emotions conveyed through voice and face shape engagement and context in human-AI interaction. Despite rapid progress in omni-modal large language models (LLMs), the holistic evaluation of emotional reasoning with audiovisual cues remains limited. To address this gap, we introduce AV-EMO-Reasoning, a benchmark designed to systematically assess emotional coherence in LLMs. The framework leverages a curated, single- and multi-turn synthetic audiovisual corpus with a real-world set and is assessed…

TravelBench : Exploring LLM Performance in Low-Resource Domains
Srinivas Billa, Xiaonan Jing
https://arxiv.org/abs/2510.02719 https://arxiv.org/pdf/2510.02…

TravelBench : Exploring LLM Performance in Low-Resource Domains
Results on existing LLM benchmarks capture little information over the model capabilities in low-resource tasks, making it difficult to develop effective solutions in these domains. To address these challenges, we curated 14 travel-domain datasets spanning 7 common NLP tasks using anonymised data from real-world scenarios, and analysed the performance across LLMs. We report on the accuracy, scaling behaviour, and reasoning capabilities of LLMs in a variety of tasks. Our results confirm that gen…