Tootfinder

Trade-offs in Data Memorization via Strong Data Processing Inequalities
Vitaly Feldman, Guy Kornowski, Xin Lyu
https://arxiv.org/abs/2506.01855 https://

Trade-offs in Data Memorization via Strong Data Processing Inequalities
Recent research demonstrated that training large language models involves memorization of a significant fraction of training data. Such memorization can lead to privacy violations when training on sensitive user data and thus motivates the study of data memorization's role in learning. In this work, we develop a general approach for proving lower bounds on excess data memorization, that relies on a new connection between strong data processing inequalities and data memorization. We then demonst…

Atomreaktor für US-Armee: BWXT arbeitet an 1,5-MW-Minireaktor als Demonstrator
Ein transportabler Minireaktor soll die Stromversorgung militärischer US-Stützpunkte sicherstellen. BWXT arbeitet mit Rolls-Royce an einem Prototypen.

Impact of Fine-Tuning Methods on Memorization in Large Language Models
Jie Hou, Chuxiong Wu, Lannan Luo, Qiang Zeng
https://arxiv.org/abs/2507.00258 https:…

Impact of Fine-Tuning Methods on Memorization in Large Language Models
As the capabilities of pre-trained large language models (LLMs) continue to advance, the "pre-train and fine-tune" paradigm has become increasingly mainstream, leading to the development of various fine-tuning methods. However, the privacy risks arising from memorization during fine-tuning have received relatively little attention. To address this gap, we categorize popular fine-tuning approaches and assess their impact on memorization through the lens of membership inference attacks (MIAs). Ou…

A Novel Discrete Memristor-Coupled Heterogeneous Dual-Neuron Model and Its Application in Multi-Scenario Image Encryption
Yi Zou, Mengjiao Wang, Xinan Zhang, Herbert Ho-Ching Iu
https://arxiv.org/abs/2505.24294

A Novel Discrete Memristor-Coupled Heterogeneous Dual-Neuron Model and Its Application in Multi-Scenario Image Encryption
Simulating brain functions using neural networks is an important area of research. Recently, discrete memristor-coupled neurons have attracted significant attention, as memristors effectively mimic synaptic behavior, which is essential for learning and memory. This highlights the biological relevance of such models. This study introduces a discrete memristive heterogeneous dual-neuron network (MHDNN). The stability of the MHDNN is analyzed with respect to initial conditions and a range of neuro…

Paris' public prosecutor says it charged 25 people, including six minors, over a spate of kidnappings and attempted abductions targeting crypto executives (Le Monde)
https://www.lemonde.fr/en/france/article/2

France charges 25 suspects in crypto abduction cases
The Paris public prosecutor said 18 people were placed in pre-trial detention on Saturday after a spate of kidnappings, successful and attempted, related to those in the crypto world.

"The administration’s claims to monarchical power are a real threat to America’s constitutional order. But its executive orders and policy feints are so haphazard and poorly articulated that they amount to a kind of autocratic takeover written in smudge-able crayon: terrifying, cartoonish, and vulnerable to erasure, all at once.
This is not to say that Americans should ignore Trump’s efforts to make confetti of the Constitution. Rather, when evaluating any one Trump policy, one has to keep front of mind the possibility that it simply won’t exist by the end of the week. Despite an energetic effort by some right-wing intellectuals to make Trump out to be some kind of 14-dimensional-chess player, his approach doesn’t resemble chess so much as a denial-of-service attack on a functioning government."
#USPolitcs
https://www.theatlantic.com/politics/archive/2025/06/trump-self-destructive-agenda/683013/

AI-guided digital intervention with physiological monitoring reduces intrusive memories after experimental trauma
Megan T. deBettencourt, Sruthi Sakthivel, Emily A. Holmes, Mark Chevillet
https://arxiv.org/abs/2507.01081

AI-guided digital intervention with physiological monitoring reduces intrusive memories after experimental trauma
Trauma prevalence is vast globally. Evidence-based digital treatments can help, but most require human guidance. Human guides provide tailored instructions and responsiveness to internal cognitive states, but limit scalability. Can generative AI and neurotechnology provide a scalable alternative? Here we test ANTIDOTE, combining AI guidance and pupillometry to automatically deliver and monitor an evidence-based digital treatment, specifically the Imagery Competing Task Intervention (ICTI), to r…

Toward Memristor-like Resonant Sensors: Observation of Pinched Hysteresis within MEMS Resonators
Erion Uka, Chun Zhao
https://arxiv.org/abs/2506.01749 http…

Toward Memristor-like Resonant Sensors: Observation of Pinched Hysteresis within MEMS Resonators
Memristors, uniquely characterized by their pinched hysteresis loop fingerprints, have attracted significant research interest over the past decade, due to their enormous potential for novel computation and artificial intelligence applications. Memristors are widely regarded as the fourth fundamental electrical component, with voltage and current being their input and output signals. In broader terms, similar pinched hysteresis behavior should also exist in other physical systems across domains…

A federal judge temporarily blocks Florida from enforcing a law banning social media accounts for children under 14, saying the law is "likely unconstitutional" (Kate Payne/Associated Press)
https://apnews.com/article/florida-soc

Rote Learning Considered Useful: Generalizing over Memorized Data in LLMs
Qinyuan Wu, Soumi Das, Mahsa Amani, Bishwamittra Ghosh, Mohammad Aflah Khan, Krishna P. Gummadi, Muhammad Bilal Zafar
https://arxiv.org/abs/2507.21914

Rote Learning Considered Useful: Generalizing over Memorized Data in LLMs
Rote learning is a memorization technique based on repetition. It is commonly believed to hinder generalization by encouraging verbatim memorization rather than deeper understanding. This insight holds for even learning factual knowledge that inevitably requires a certain degree of memorization. In this work, we demonstrate that LLMs can be trained to generalize from rote memorized data. We introduce a two-phase memorize-then-generalize framework, where the model first rote memorizes factual su…