Tootfinder

Weekend #Plankton Factoid 🦠🦐
While algal #blooms are often associated with summer, they can occur in any season. SE #Australia is experiencing a prolonged bloom of Karenia mikimotoi (same genus as Flor…

QuickSilver -- Speeding up LLM Inference through Dynamic Token Halting, KV Skipping, Contextual Token Fusion, and Adaptive Matryoshka Quantization
Danush Khanna, Aditya Kumar Guru, Srivarshinee Sridhar, Zidan Ahmed, Rubhav Bahirwani, Meetu Malhotra, Vinija Jain, Aman Chadha, Amitava Das, Kripabandhu Ghosh
https://arxiv.org/abs/2…

QuickSilver -- Speeding up LLM Inference through Dynamic Token Halting, KV Skipping, Contextual Token Fusion, and Adaptive Matryoshka Quantization
Inference accounts for the majority of latency and energy consumption in large language model (LLM) deployments, often exceeding 90% of total cost. While training-time efficiency has seen extensive progress, runtime optimization remains a key bottleneck, particularly under autoregressive decoding. Existing approaches -- such as pruning, quantization, early exits, and speculative decoding -- often require retraining, architectural changes, or disrupt decoding compatibility. We introduce QuickSil…

Optimal Estimation of Watermark Proportions in Hybrid AI-Human Texts
Xiang Li, Garrett Wen, Weiqing He, Jiayuan Wu, Qi Long, Weijie J. Su
https://arxiv.org/abs/2506.22343

Optimal Estimation of Watermark Proportions in Hybrid AI-Human Texts
Text watermarks in large language models (LLMs) are an increasingly important tool for detecting synthetic text and distinguishing human-written content from LLM-generated text. While most existing studies focus on determining whether entire texts are watermarked, many real-world scenarios involve mixed-source texts, which blend human-written and watermarked content. In this paper, we address the problem of optimally estimating the watermark proportion in mixed-source texts. We cast this proble…

The Tate-Shafarevich group of a polarised K3 surface
Daniel Huybrechts, Dominique Mattei
https://arxiv.org/abs/2507.22703 https://arxiv.org/pdf/2507.22703

The Tate-Shafarevich group of a polarised K3 surface
In an earlier paper we generalised the notion of the Tate-Shafarevich group of an elliptic K3 surface to the Tate-Shafarevich group of a polarised K3 surface. In the present note, we complement the result by proving that the Tate-Shafarevich group of a polarised K3 surface (S,h) with h primitive parametrises bijectively all torsors for the Jacobian of the generic curve in the linear system |h| that admit a good hyperkähler compactification. The result is seen as the analogue of the c…

A triptych on Starhawk, who developed some of the most poignantly insightful ideas in structural theories of power, among the many philosophers, sociologists, and psychologists who wrote about theories of power in the 20th century
Note that the second and third image come from the *same book*, in the first and second chapters
I…did not expect woo poetry on magic and sexuality in a book that's highly cited by writers on theories of power

#WoodyGuthrie recorded some stuff while suffering from ALS that’s never been released. Now it has been.
#giftArticle https://wapo.st/44Z5SBz

The Woody Guthrie tapes you weren’t supposed to hear
“Woody at Home,” a new CD of some of Woody Guthrie’s last recordings, shows the effect of Huntington’s disease on the “This Land Is Your Land” and “Deportee” singer.

Geometric Kernels of Proper Maps Between Non-Compact Surfaces
Sumanta Das
https://arxiv.org/abs/2508.21057 https://arxiv.org/pdf/2508.21057

Geometric Kernels of Proper Maps Between Non-Compact Surfaces
A map between connected $2$-manifolds has a geometric kernel if it sends a non-contractible simple loop to a null-homotopic loop. While every non-$π_1$-injective map between compact surfaces admits a geometric kernel, this generally fails for compact bordered or non-compact surfaces. In this paper, we use Brown's proper fundamental group to give a sufficient condition under which a degree-one map between non-compact surfaces admits a geometric kernel. Furthermore, we characterize conjugacy cla…

Between Markov and restriction: Two more monads on categories for relations
Cipriano Junior Cioffo, Fabio Gadducci, Davide Trotta
https://arxiv.org/abs/2508.20054 https://

Between Markov and restriction: Two more monads on categories for relations
The study of categories abstracting the structural properties of relations has been extensively developed over the years, resulting in a rich and diverse body of work. A previous paper offered a survey providing a modern and comprehensive presentation of these ``categories for relations'' as instances of gs-monoidal categories, showing how they arise as Kleisli categories of suitable symmetric monoidal monads. The end result was a taxonomy that organised numerous related concepts in the literat…

Languages Still Left Behind: Toward a Better Multilingual Machine Translation Benchmark
Chihiro Taguchi, Seng Mai, Keita Kurabe, Yusuke Sakai, Georgina Agyei, Soudabeh Eslami, David Chiang
https://arxiv.org/abs/2508.20511

Languages Still Left Behind: Toward a Better Multilingual Machine Translation Benchmark
Multilingual machine translation (MT) benchmarks play a central role in evaluating the capabilities of modern MT systems. Among them, the FLORES+ benchmark is widely used, offering English-to-many translation data for over 200 languages, curated with strict quality control protocols. However, we study data in four languages (Asante Twi, Japanese, Jinghpaw, and South Azerbaijani) and uncover critical shortcomings in the benchmark's suitability for truly multilingual evaluation. Human assessments…

Efficient Attention Mechanisms for Large Language Models: A Survey
Yutao Sun, Zhenyu Li, Yike Zhang, Tengyu Pan, Bowen Dong, Yuyi Guo, Jianyong Wang
https://arxiv.org/abs/2507.19595

Efficient Attention Mechanisms for Large Language Models: A Survey
Transformer-based architectures have become the prevailing backbone of large language models. However, the quadratic time and memory complexity of self-attention remains a fundamental obstacle to efficient long-context modeling. To address this limitation, recent research has introduced two principal categories of efficient attention mechanisms. Linear attention methods achieve linear complexity through kernel approximations, recurrent formulations, or fastweight dynamics, thereby enabling scal…