Tootfinder

Prompt Attacks Reveal Superficial Knowledge Removal in Unlearning Methods
Yeonwoo Jang, Shariqah Hossain, Ashwin Sreevatsa, Diogo Cruz
https://arxiv.org/abs/2506.10236

Prompt Attacks Reveal Superficial Knowledge Removal in Unlearning Methods
In this work, we show that some machine unlearning methods may fail when subjected to straightforward prompt attacks. We systematically evaluate eight unlearning techniques across three model families, and employ output-based, logit-based, and probe analysis to determine to what extent supposedly unlearned knowledge can be retrieved. While methods like RMU and TAR demonstrate robust unlearning, ELM remains vulnerable to specific prompt attacks (e.g., Hindi filler text in original prompt recover…

Algorithmic methods of finite discrete structures. Topological graph drawing (part III)
Sergey Kurapov, Maxim Davidovsky
https://arxiv.org/abs/2506.10936 h…

Algorithmic methods of finite discrete structures. Topological graph drawing (part III)
The manuscript considers mathematical models for creating a topological drawing of a graph based on the methods of G. Ringel's vertex rotation theory. An algorithm is presented for generating a topological drawing of a flat part of a graph based on the selection of a basis for the cycle subspace C(G) using the Monte Carlo method. A steepest descent method for constructing a topological drawing of a flat subgraph is described in the manuscript. The topological drawing of a graph is constructed u…

Series A, Episode 09 - Project Avalon
AVON: [Avon looks at the screen] Not quite. I think I shall contain my enthusiasm here in the warm.
[Scene: A mutoid watches a guard below her position, then jumps him and after a brief struggle strangles him. Then she signals to various groups of mutoids with a morse-code like sender. The groups move through caves. A number of mutoids surround a group of people in a cavern. Travis is in the background.]

Blakes7Bot | Search
Companion site to the Blakes7Bot Twitter, Mastodon, and Bluesky accounts. Automated lines of Blakes 7 dialog. Includes a script search facility.

AI-Based Software Vulnerability Detection: A Systematic Literature Review
Samiha Shimmi, Hamed Okhravi, Mona Rahimi
https://arxiv.org/abs/2506.10280 https:…

AI-Based Software Vulnerability Detection: A Systematic Literature Review
Software vulnerabilities in source code pose serious cybersecurity risks, prompting a shift from traditional detection methods (e.g., static analysis, rule-based matching) to AI-driven approaches. This study presents a systematic review of software vulnerability detection (SVD) research from 2018 to 2023, offering a comprehensive taxonomy of techniques, feature representations, and embedding methods. Our analysis reveals that 91% of studies use AI-based methods, with graph-based models being th…

NSD-Imagery: A benchmark dataset for extending fMRI vision decoding methods to mental imagery https://arxiv.org/abs/2506.06898

NSD-Imagery: A benchmark dataset for extending fMRI vision decoding methods to mental imagery
We release NSD-Imagery, a benchmark dataset of human fMRI activity paired with mental images, to complement the existing Natural Scenes Dataset (NSD), a large-scale dataset of fMRI activity paired with seen images that enabled unprecedented improvements in fMRI-to-image reconstruction efforts. Recent models trained on NSD have been evaluated only on seen image reconstruction. Using NSD-Imagery, it is possible to assess how well these models perform on mental image reconstruction. This is a chal…

Convergence of adaptive boundary element methods driven by functional a posteriori error estimates
Alexander Freiszlinger, Dirk Pauly, Dirk Praetorius
https://arxiv.org/abs/2506.10499

Convergence of adaptive boundary element methods driven by functional a posteriori error estimates
The recent work [Kurz et al., Numer. Math., 147 (2021)] proposed functional a posteriori error estimates for boundary element methods (BEMs) together with a related adaptive mesh-refinement strategy. Unlike most a posteriori BEM error estimators, the proposed functional error estimators cover Galerkin as well as collocation BEM and, more importantly, do not control the error in the integral density on the boundary, but the error of the potential approximation in the domain, which is of greater …

Empirical and computer-aided robustness analysis of long-step and accelerated methods in smooth convex optimization
Pierre Vernimmen, Fran\c{c}ois Glineur
https://arxiv.org/abs/2506.09730

Empirical and computer-aided robustness analysis of long-step and accelerated methods in smooth convex optimization
This work assesses both empirically and theoretically, using the performance estimation methodology, how robust different first-order optimization methods are when subject to relative inexactness in their gradient computations. Relative inexactness occurs, for example, when compressing the gradient using fewer bits of information, which happens when dealing with large-scale problems on GPUs. Three major families of methods are analyzed: constant step gradient descent, long-step methods, and acc…

Speculative Design in Spiraling Time: Methods and Indigenous HCI
James Eschrich, Cole McMullen, Sarah Sterman
https://arxiv.org/abs/2506.10229 https://

Speculative Design in Spiraling Time: Methods and Indigenous HCI
In this position paper, we first discuss the uptake of speculative design as a method for Indigenous HCI. Then, we outline how a key assumption about temporality threatens to undermine the usefulness of speculative design in this context. Finally, we briefly sketch out a possible alternative understanding of speculative design, based on the concept of "spiraling time," which could be better suited for Indigenous HCI.

In Brazil, Volkswagen is once again being sued for its use of slave labor during the military dictatorship between 1974 - 1986. Looks like VW just exported the methods they had already pioneered during WW2 closer to home…
(article in German)
https://amerika21.de/2025/06/275586/vw-angezeigt-brasilien

Alternating steepest descent methods for tensor completion with applications to spectromicroscopy
Oliver Townsend, Sergey Dolgov, Silvia Gazzola, Misha Kilmer
https://arxiv.org/abs/2506.10661

Alternating steepest descent methods for tensor completion with applications to spectromicroscopy
In this paper we develop two new Tensor Alternating Steepest Descent algorithms for tensor completion in the low-rank $\star_{M}$-product format, whereby we aim to reconstruct an entire low-rank tensor from a small number of measurements thereof. Both algorithms are rooted in the Alternating Steepest Descent (ASD) method for matrix completion, first proposed in [J. Tanner and K. Wei, Appl. Comput. Harmon. Anal., 40 (2016), pp. 417-429]. In deriving the new methods we target the X-ray spectromic…