Tootfinder

Pushback grows against nickel mining in Indonesian marine paradise of Raja Ampat https://news.mongabay.com/2025/06/pushback-grows-against-nickel-mining-in-indonesian-marine-paradise-of-raja-ampat/

Pushback grows against nickel mining in Indonesian marine paradise of Raja Ampat
JAKARTA — The Indonesian government has suspended nickel mining in the Raja Ampat archipelago, citing multiple environmental violations by companies operating in the ecologically sensitive islands. The move follows viral social media posts highlighting the expansion of nickel mines and the attendant environmental degradation in the region famous for its picturesque islets and turquoise waters, […]

On-the-Fly Adaptive Distillation of Transformer to Dual-State Linear Attention
Yeonju Ro, Zhenyu Zhang, Souvik Kundu, Zhangyang Wang, Aditya Akella
https://arxiv.org/abs/2506.09316

On-the-Fly Adaptive Distillation of Transformer to Dual-State Linear Attention
Large language models (LLMs) excel at capturing global token dependencies via self-attention but face prohibitive compute and memory costs on lengthy inputs. While sub-quadratic methods (e.g., linear attention) can reduce these costs, they often degrade accuracy due to overemphasizing recent tokens. In this work, we first propose \textit{dual-state linear attention} (\textbf{\dsla}), a novel design that maintains two specialized hidden states-one for preserving historical context and one for tr…

AR2: Attention-Guided Repair for the Robustness of CNNs Against Common Corruptions
Fuyuan Zhang, Qichen Wang, Jianjun Zhao
https://arxiv.org/abs/2507.06332

AR2: Attention-Guided Repair for the Robustness of CNNs Against Common Corruptions
Deep neural networks suffer from significant performance degradation when exposed to common corruptions such as noise, blur, weather, and digital distortions, limiting their reliability in real-world applications. In this paper, we propose AR2 (Attention-Guided Repair for Robustness), a simple yet effective method to enhance the corruption robustness of pretrained CNNs. AR2 operates by explicitly aligning the class activation maps (CAMs) between clean and corrupted images, encouraging the model…

This https://arxiv.org/abs/2506.06278 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csLG_…

Distillation Robustifies Unlearning
Current LLM unlearning methods are not robust: they can be reverted easily with a few steps of finetuning. This is true even for the idealized unlearning method of training to imitate an oracle model that was never exposed to unwanted information, suggesting that output-based finetuning is insufficient to achieve robust unlearning. In a similar vein, we find that training a randomly initialized student to imitate an unlearned model transfers desired behaviors while leaving undesired capabilitie…

Gluten-Free Vegan Chocolate Cheesecake #plantbased

Gluten-Free Vegan Chocolate Cheesecake
A vegan chocolate cheesecake is just what you need for a summer dessert. It's practically no bake save for the crust.

"As with many digital [vibe coding] doesn't fully stand up to scrutiny but show a deep misunderstanding of how software is made, the potential externalities (and internalities) software brings and a disdain for experience and embodied knowledge."
(Original title: On “Vibe Coding”)
https://tant…

On "Vibe Coding"
The hype about the potentials (it’s always future potential, never real current use) of AI has discarded its last cycle (“reasoning models”/”deep research”, both terms being factually untrue and deeply deceiving at best) and moved to a new double whammy of “agentic AI” and “Vibe Coding”. Now “agentic AI” basically just means that some LLM […]

Witnessing and guiding sets of tangles
Annegret Seibt
https://arxiv.org/abs/2507.04394 https://arxiv.org/pdf/2507.04394

Witnessing and guiding sets of tangles
Tangles o er a way to indirectly but precisely capture cluster-like though possibly fuzzy substructures in discrete data. In this paper, we analyze witnessing and guiding sets of tangles that can help to find proper cluster candidates for given tangles. We show that every k-tangle has a witnessing set whose size is bounded in an exponential function in k which improves a result of Grohe and Schweizer. Further, we generalize a result of Diestel, Elbracht and Jacobs by providing a characterizatio…

This https://arxiv.org/abs/2503.02321 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_ees…

Rapid Bone Scintigraphy Enhancement via Semantic Prior Distillation from Segment Anything Model
Rapid bone scintigraphy is crucial for diagnosing skeletal disorders and detecting tumor metastases in children, as it shortens scan duration and reduces discomfort. However, accelerated acquisition often degrades image quality, impairing the visibility of fine anatomical details and potentially compromising diagnosis. To overcome this limitation, we introduce the first application of SAM-based semantic priors for medical image restoration, utilizing the Segment Anything Model (SAM) to enhance …

Distillation Robustifies Unlearning
Bruce W. Lee, Addie Foote, Alex Infanger, Leni Shor, Harish Kamath, Jacob Goldman-Wetzler, Bryce Woodworth, Alex Cloud, Alexander Matt Turner
https://arxiv.org/abs/2506.06278

Distillation Robustifies Unlearning
Current LLM unlearning methods are not robust: they can be reverted easily with a few steps of finetuning. This is true even for the idealized unlearning method of training to imitate an oracle model that was never exposed to unwanted information, suggesting that output-based finetuning is insufficient to achieve robust unlearning. In a similar vein, we find that training a randomly initialized student to imitate an unlearned model transfers desired behaviors while leaving undesired capabilitie…

Cardamom Pistachio Ice Cream (Dairy-Free And Vegan) #plantbased

Cardamom Pistachio Ice Cream (Dairy-Free And Vegan)
This cardamom pistachio ice cream is made with coconut milk and based on the Indian dessert Kulfi - dense ice cream.