Tootfinder

Given the reticence of its local affiliates to mention it's name, one could get the impression that ACT, the NZ Initiative, Curia, the NZ Taxpayers' 'Union', and the 'Free Speech' 'Union' are keen to avoid drawing attention to their Atlas Network (also responsible for the US' Project 2025) affiliation. Could it be they're keen to avoid more scrutiny of where their funding comes from? 🤔 Assuming their search is trustworthy, on the RNZ website, there'…

I find many folks today are fearful of confrontation. They see it as a proxy for violence. Confrontation is precisely not violence. In fact, in many cases, it's the only viable gambit for initiating communication with a goal of achieving understanding, defusing frustration. It may also be the only way to *avoid* future violence.
I think we should all be practicing confrontation, and get good at it. That's how the world becomes better. By avoiding it, we're just stockpiling…

Pure-Pass: Fine-Grained, Adaptive Masking for Dynamic Token-Mixing Routing in Lightweight Image Super-Resolution
Junyu Wu, Jie Tang, Jie Liu, Gangshan Wu
https://arxiv.org/abs/2510.01997

Pure-Pass: Fine-Grained, Adaptive Masking for Dynamic Token-Mixing Routing in Lightweight Image Super-Resolution
Image Super-Resolution (SR) aims to reconstruct high-resolution images from low-resolution counterparts, but the computational complexity of deep learning-based methods often hinders practical deployment. CAMixer is the pioneering work to integrate the advantages of existing lightweight SR methods and proposes a content-aware mixer to route token mixers of varied complexities according to the difficulty of content recovery. However, several limitations remain, such as poor adaptability, coarse-…

Ugh, listening to Kathryn Ryan on RNZ NineToNoon interviewing Sonny Taite(?) about an 'AI' scribe to take hospital staff clinical notes. Why - oh why - doesn't she ask the important question: where does the data go?! Which jurisdiction? Who's storing it? Who has access? Who cares for it and maintains its security & integrity? Her producers miss it every bloody time. Very disappointing.

Designing a Layered Framework to Secure Data via Improved Multi Stage Lightweight Cryptography in IoT Cloud Systems
Hojjat Farshadinia, Ali Barati, Hamid Barati
https://arxiv.org/abs/2509.01717

Designing a Layered Framework to Secure Data via Improved Multi Stage Lightweight Cryptography in IoT Cloud Systems
This paper presents a novel multi-layered hybrid security approach aimed at enhancing lightweight encryption for IoT-Cloud systems. The primary goal is to overcome limitations inherent in conventional solutions such as TPA, Blockchain, ECDSA and ZSS which often fall short in terms of data protection, computational efficiency and scalability. Our proposed method strategically refines and integrates these technologies to address their shortcomings while maximizing their individual strengths. By d…

AFFORD2ACT: Affordance-Guided Automatic Keypoint Selection for Generalizable and Lightweight Robotic Manipulation
Anukriti Singh, Kasra Torshizi, Khuzema Habib, Kelin Yu, Ruohan Gao, Pratap Tokekar
https://arxiv.org/abs/2510.01433

AFFORD2ACT: Affordance-Guided Automatic Keypoint Selection for Generalizable and Lightweight Robotic Manipulation
Vision-based robot learning often relies on dense image or point-cloud inputs, which are computationally heavy and entangle irrelevant background features. Existing keypoint-based approaches can focus on manipulation-centric features and be lightweight, but either depend on manual heuristics or task-coupled selection, limiting scalability and semantic understanding. To address this, we propose AFFORD2ACT, an affordance-guided framework that distills a minimal set of semantic 2D keypoints from a…

ArabEmoNet: A Lightweight Hybrid 2D CNN-BiLSTM Model with Attention for Robust Arabic Speech Emotion Recognition
Ali Abouzeid, Bilal Elbouardi, Mohamed Maged, Shady Shehata
https://arxiv.org/abs/2509.01401

ArabEmoNet: A Lightweight Hybrid 2D CNN-BiLSTM Model with Attention for Robust Arabic Speech Emotion Recognition
Speech emotion recognition is vital for human-computer interaction, particularly for low-resource languages like Arabic, which face challenges due to limited data and research. We introduce ArabEmoNet, a lightweight architecture designed to overcome these limitations and deliver state-of-the-art performance. Unlike previous systems relying on discrete MFCC features and 1D convolutions, which miss nuanced spectro-temporal patterns, ArabEmoNet uses Mel spectrograms processed through 2D convolutio…

Lightweight Error-Correction Code Encoders in Superconducting Electronic Systems
Yerzhan Mustafa, Berker Pek\"oz, Sel\c{c}uk K\"ose
https://arxiv.org/abs/2509.00962 ht…

Lightweight Error-Correction Code Encoders in Superconducting Electronic Systems
Data transmission from superconducting electronic circuits, such as single flux quantum (SFQ) logic, to room-temperature electronics is susceptible to bit errors, which may result from flux trapping, fabrication defects, and process parameter variations (PPV). Due to the cooling power budget at 4.2 K and constraints on the chip area, the size of the error-correction code encoders is limited. In this work, three lightweight error-correction code encoders are proposed that are based on Hamming(7,…

Designing a Lightweight GenAI Interface for Visual Data Analysis
Ratanond Koonchanok, Alex Kale, Khairi Reda
https://arxiv.org/abs/2509.02878 https://arxiv…

Designing a Lightweight GenAI Interface for Visual Data Analysis
Recent advances in Generative AI have transformed how users interact with data analysis through natural language interfaces. However, many systems rely too heavily on LLMs, creating risks of hallucination, opaque reasoning, and reduced user control. We present a hybrid visual analysis system that integrates GenAI in a constrained, high-level role to support statistical modeling while preserving transparency and user agency. GenAI translates natural language intent into formal statistical formul…

Wow, there's a mighty wind blowing in Ōtautahi right now - took my life in my hands putting out the washing... Chairs blowing off decks, trees and bedsheets whipping around. The Nor'west is formidable. And it's warm, with all the trees starting to leaf out...

Draugnet is a lightweight, open-source tool for anonymous cyber threat reporting. Built for the MISP ecosystem, it lets users submit and track reports — no accounts, no logins, just a simple token. Supports plain text, MISP JSON, and object templates. ideal for CSIRTs, ISACs, and trusted communities. 🔐 Built for anonymity 🧩 Easy MISP integration .
version 1.2.0 released with RTIR, Flowintel and Ollama integration.
Source code:

Hmm - seems that Hetzner in Falkenstein is having issues... anyone else noticing that?

LiLa-Net: Lightweight Latent LiDAR Autoencoder for 3D Point Cloud Reconstruction
Mario Resino, Borja P\'erez, Jaime Godoy, Abdulla Al-Kaff, Fernando Garc\'ia
https://arxiv.org/abs/2510.02028

LiLa-Net: Lightweight Latent LiDAR Autoencoder for 3D Point Cloud Reconstruction
This work proposed a 3D autoencoder architecture, named LiLa-Net, which encodes efficient features from real traffic environments, employing only the LiDAR's point clouds. For this purpose, we have real semi-autonomous vehicle, equipped with Velodyne LiDAR. The system leverage skip connections concept to improve the performance without using extensive resources as the state-of-the-art architectures. Key changes include reducing the number of encoder layers and simplifying the skip connections, …

LiFeChain: Lightweight Blockchain for Secure and Efficient Federated Lifelong Learning in IoT
Handi Chen, Jing Deng, Xiuzhe Wu, Zhihan Jiang, Xinchen Zhang, Xianhao Chen, Edith C. H. Ngai
https://arxiv.org/abs/2509.01434

LiFeChain: Lightweight Blockchain for Secure and Efficient Federated Lifelong Learning in IoT
The expansion of Internet of Things (IoT) devices constantly generates heterogeneous data streams, driving demand for continuous, decentralized intelligence. Federated Lifelong Learning (FLL) provides an ideal solution by incorporating federated and lifelong learning to overcome catastrophic forgetting. The extended lifecycle of FLL in IoT systems increases their vulnerability to persistent attacks, and these risks may be obscured by performance degradation caused by spatial-temporal data heter…

A Comparative Study of Controllability, Explainability, and Performance in Dysfluency Detection Models
Eric Zhang, Li Wei, Sarah Chen, Michael Wang
https://arxiv.org/abs/2509.00058

A Comparative Study of Controllability, Explainability, and Performance in Dysfluency Detection Models
Recent advances in dysfluency detection have introduced a variety of modeling paradigms, ranging from lightweight object-detection inspired networks (YOLOStutter) to modular interpretable frameworks (UDM). While performance on benchmark datasets continues to improve, clinical adoption requires more than accuracy: models must be controllable and explainable. In this paper, we present a systematic comparative analysis of four representative approaches--YOLO-Stutter, FluentNet, UDM, and SSDM--alon…

NeuroSwift: A Lightweight Cross-Subject Framework for fMRI Visual Reconstruction of Complex Scenes
Shiyi Zhang, Dong Liang, Yihang Zhou
https://arxiv.org/abs/2510.02266 https://…

NeuroSwift: A Lightweight Cross-Subject Framework for fMRI Visual Reconstruction of Complex Scenes
Reconstructing visual information from brain activity via computer vision technology provides an intuitive understanding of visual neural mechanisms. Despite progress in decoding fMRI data with generative models, achieving accurate cross-subject reconstruction of visual stimuli remains challenging and computationally demanding. This difficulty arises from inter-subject variability in neural representations and the brain's abstract encoding of core semantic features in complex visual inputs. To …

Accuracy vs Performance: An abstraction model for deadline constrained offloading at the mobile-edge
Jamie Cotter, Ignacio Castineiras, Victor Cionca
https://arxiv.org/abs/2510.01885

Accuracy vs Performance: An abstraction model for deadline constrained offloading at the mobile-edge
In this paper, we present a solution for low-latency deadline-constrained DNN offloading on mobile edge devices. We design a scheduling algorithm with lightweight network state representation, considering device availability, communication on the network link, priority-aware pre-emption, and task deadlines. The scheduling algorithm aims to reduce latency by designing a resource availability representation, as well as a network discretisation and a dynamic bandwidth estimation mechanism. We impl…

A Single Detect Focused YOLO Framework for Robust Mitotic Figure Detection
Yasemin Topuz, M. Taha G\"okcan, Serdar Y{\i}ld{\i}z, Song\"ul Varl{\i}
https://arxiv.org/abs/2509.02637

A Single Detect Focused YOLO Framework for Robust Mitotic Figure Detection
Mitotic figure detection is a crucial task in computational pathology, as mitotic activity serves as a strong prognostic marker for tumor aggressiveness. However, domain variability that arises from differences in scanners, tissue types, and staining protocols poses a major challenge to the robustness of automated detection methods. In this study, we introduce SDF-YOLO (Single Detect Focused YOLO), a lightweight yet domain-robust detection framework designed specifically for small, rare targets…

Wow. So much implicit trickle down economics. https://www.rnz.co.nz/national/programmes/afternoons/audio/2019002510/how-excited-should-we-be-about-the-amazon-announcement Come on, Paul, AWS and Microsoft having…

Portable Targeted Sampling Framework Using LLVM
Zhantong Qiu, Mahyar Samani, Jason Lowe-Power
https://arxiv.org/abs/2509.02873 https://arxiv.org/pdf/2509.0…

Portable Targeted Sampling Framework Using LLVM
Comprehensive architectural evaluation of full workloads is throttled by slow simulation and per-binary sampling pipelines. We present Nugget, a flexible framework for portable sampling across simulators and real hardware, ISAs, and libraries. Nugget operates at the LLVM IR level to perform binary-agnostic interval analysis, then emits lightweight, cross-platform executables--nuggets--that can be validated on real machines before driving simulation. Across SPEC CPU2017, NPB, and LSMS, Nugget cu…

Improved $\ell_{p}$ Regression via Iteratively Reweighted Least Squares
Alina Ene, Ta Duy Nguyen, Adrian Vladu
https://arxiv.org/abs/2510.01729 https://arx…

Improved $\ell_{p}$ Regression via Iteratively Reweighted Least Squares
We introduce fast algorithms for solving $\ell_{p}$ regression problems using the iteratively reweighted least squares (IRLS) method. Our approach achieves state-of-the-art iteration complexity, outperforming the IRLS algorithm by Adil-Peng-Sachdeva (NeurIPS 2019) and matching the theoretical bounds established by the complex algorithm of Adil-Kyng-Peng-Sachdeva (SODA 2019, J. ACM 2024) via a simpler lightweight iterative scheme. This bridges the existing gap between theoretical and practical a…

Crosslisted article(s) found for cs.NE. https://arxiv.org/list/cs.NE/new
[1/1]:
- StableSleep: Source-Free Test-Time Adaptation for Sleep Staging with Lightweight Safety Rails
Hritik Arasu, Faisal R Jahangiri

BioVERSE: Representation Alignment of Biomedical Modalities to LLMs for Multi-Modal Reasoning
Ching-Huei Tsou, Michal Ozery-Flato, Ella Barkan, Diwakar Mahajan, Ben Shapira
https://arxiv.org/abs/2510.01428

BioVERSE: Representation Alignment of Biomedical Modalities to LLMs for Multi-Modal Reasoning
Recent advances in large language models (LLMs) and biomedical foundation models (BioFMs) have achieved strong results in biological text reasoning, molecular modeling, and single-cell analysis, yet they remain siloed in disjoint embedding spaces, limiting cross-modal reasoning. We present BIOVERSE (Biomedical Vector Embedding Realignment for Semantic Engagement), a two-stage approach that adapts pretrained BioFMs as modality encoders and aligns them with LLMs through lightweight, modality-spec…

KTBox: A Modular LaTeX Framework for Semantic Color, Structured Highlighting, and Scholarly Communication
Bhaskar Mangal, Ashutosh Bhatia, Yashvardhan Sharma, Kamlesh Tiwari, Rashmi Verma
https://arxiv.org/abs/2510.01961

KTBox: A Modular LaTeX Framework for Semantic Color, Structured Highlighting, and Scholarly Communication
The communication of technical insight in scientific manuscripts often relies on ad-hoc formatting choices, resulting in inconsistent visual emphasis and limited portability across document classes. This paper introduces ktbox, a modular LaTeX framework that unifies semantic color palettes, structured highlight boxes, taxonomy trees, and author metadata utilities into a coherent system for scholarly writing. The framework is distributed as a set of lightweight, namespaced components: ktcolor.st…

from my link log —
Wiki Creole 1.0 lightweight markup language.
http://www.wikicreole.org/wiki/Creole1.0
saved 2018-12-03 https://dotat.at/:/69…

Prompt Curriculum Learning for Efficient LLM Post-Training
Zhaolin Gao, Joongwon Kim, Wen Sun, Thorsten Joachims, Sid Wang, Richard Yuanzhe Pang, Liang Tan
https://arxiv.org/abs/2510.01135

Prompt Curriculum Learning for Efficient LLM Post-Training
We introduce Prompt Curriculum Learning (PCL), a lightweight reinforcement learning (RL) algorithm that selects intermediate-difficulty prompts using a learned value model to post-train language models. Since post-training LLMs via RL remains sensitive to batching and prompt selection strategies, we first conduct a series of systematic experiments where we (1) determine the optimal training batch size that balances generation efficiency and gradient quality and (2) establish the importance of f…

Entropy-Guided Loop: Achieving Reasoning through Uncertainty-Aware Generation
Andrew G. A. Correa, Ana C. H de Matos
https://arxiv.org/abs/2509.00079 https://

Entropy-Guided Loop: Achieving Reasoning through Uncertainty-Aware Generation
Reasoning models often outperform smaller models but at 3--5$\times$ higher cost and added latency. We present entropy-guided refinement: a lightweight, test-time loop that uses token-level uncertainty to trigger a single, targeted refinement pass. We extract logprobs, compute Shannon entropy on top-$k$ alternatives, and apply a simple OR-logic trigger over perplexity, maximum token entropy, and low-confidence-token count. Unlike approaches that use entropy only for measurement or decoding, we …

Anybody out there who uses DigitalOcean cloud services, beware a phishing email telling you you've got an invoice for 'Support Services' that is DigitalOcean branded - check the URL of the payment button (which reveals a sham domain name when you hover over it).

WaveLLDM: Design and Development of a Lightweight Latent Diffusion Model for Speech Enhancement and Restoration
Kevin Putra Santoso, Rizka Wakhidatus Sholikah, Raden Venantius Hari Ginardi
https://arxiv.org/abs/2508.21153

WaveLLDM: Design and Development of a Lightweight Latent Diffusion Model for Speech Enhancement and Restoration
High-quality audio is essential in a wide range of applications, including online communication, virtual assistants, and the multimedia industry. However, degradation caused by noise, compression, and transmission artifacts remains a major challenge. While diffusion models have proven effective for audio restoration, they typically require significant computational resources and struggle to handle longer missing segments. This study introduces WaveLLDM (Wave Lightweight Latent Diffusion Model),…

Mycroft: Tracing Dependencies in Collective Communication Towards Reliable LLM Training
Yangtao Deng, Lei Zhang, Qinlong Wang, Xiaoyun Zhi, Xinlei Zhang, Zhuo Jiang, Haohan Xu, Lei Wang, Zuquan Song, Gaohong Liu, Yang Bai, Shuguang Wang, Wencong Xiao, Jianxi Ye, Minlan Yu, Hong Xu
https://arxiv.org/abs/2509.03018

Mycroft: Tracing Dependencies in Collective Communication Towards Reliable LLM Training
Reliability is essential for ensuring efficiency in LLM training. However, many real-world reliability issues remain difficult to resolve, resulting in wasted resources and degraded model performance. Unfortunately, today's collective communication libraries operate as black boxes, hiding critical information needed for effective root cause analysis. We propose Mycroft, a lightweight distributed tracing and root cause analysis system designed to address previously hidden reliability issues in c…

And, on this second day of 'spring' (according to convention in Aotearoa), we've got mostly beautiful sunshine and mild temperatures here in Ōtautahi... We've got the first daffodils blooming and I just saw 5 native skinks in the garden, 3 sunning themselves on the tiles in our lizardarium. At least one was gravid (carrying eggs)... With indigenous NZ skinks, the eggs hatch in the female and are born live.

Comparison between Supervised and Unsupervised Learning in Deep Unfolded Sparse Signal Recovery
Koshi Nagahisa, Ryo Hayakawa, Youji Iiguni
https://arxiv.org/abs/2509.01331 https…

Comparison between Supervised and Unsupervised Learning in Deep Unfolded Sparse Signal Recovery
This paper investigates the impact of loss function selection in deep unfolding techniques for sparse signal recovery algorithms. Deep unfolding transforms iterative optimization algorithms into trainable lightweight neural networks by unfolding their iterations as network layers, with various loss functions employed for parameter learning depending on application contexts. We focus on deep unfolded versions of the fundamental iterative shrinkage thresholding algorithm (ISTA) and the iterative …

MODE: Mixture of Document Experts for RAG
Rahul Anand
https://arxiv.org/abs/2509.00100 https://arxiv.org/pdf/2509.00100

MODE: Mixture of Document Experts for RAG
Retrieval-Augmented Generation (RAG) often relies on large vector databases and cross-encoders tuned for large-scale corpora, which can be excessive for small, domain-specific collections. We present MODE (Mixture of Document Experts), a lightweight alternative that replaces fine-grained nearest-neighbor search with cluster-and-route retrieval. Documents are embedded, grouped into semantically coherent clusters, and represented by cached centroids. At query time, we route to the top centroid(s)…

Replaced article(s) found for cs.CR. https://arxiv.org/list/cs.CR/new
[1/1]:
- A Quantitative Security Analysis of S-boxes in the NIST Lightweight Cryptography Finalists
Mahnoor Naseer, Sundas Tariq, Naveed Riaz, Naveed Ahmed, Mureed Hussain

One More Glance with Sharp Eyes: Rethinking Lightweight Captioning as a Practical Visual Specialist
Junha Song, Yongsik Jo, So Yeon Min, Quanting Xie, Taehwan Kim, Yonatan Bisk, Jaegul Choo
https://arxiv.org/abs/2508.21451

One More Glance with Sharp Eyes: Rethinking Lightweight Captioning as a Practical Visual Specialist
Image captioning is fundamental for applications like video instruction systems and exploration robots, yet deploying such models on local devices is challenging due to the high computational demands of multimodal large language models (MLLMs). To address this, we first explore lightweight captioning by implementing a specialist based on a 125M-parameter language model, 56 times smaller than LLaMA-7B, and evaluating its performance on both single-sentence and detailed captioning tasks. Surprisi…

Counterfactual Sensitivity for Faithful Reasoning in Language Models
Ibne Farabi Shihab, Sanjeda Akter, Anuj Sharma
https://arxiv.org/abs/2509.01544 https://

Counterfactual Sensitivity for Faithful Reasoning in Language Models
Large language models (LLMs) often produce correct answers while relying on flawed or irrelevant reasoning traces, undermining their trustworthiness in high-stakes domains. We propose Counterfactual Sensitivity Regularization (CSR), a lightweight training objective that enforces dependence between intermediate reasoning and final outputs. CSR introduces automated, operator-level counterfactual interventions (e.g., swapping "+" with "-") during training and penalizes models that preserve the sam…

CabinSep: IR-Augmented Mask-Based MVDR for Real-Time In-Car Speech Separation with Distributed Heterogeneous Arrays
Runduo Han, Yanxin Hu, Yihui Fu, Zihan Zhang, Yukai Jv, Li Chen, Lei Xie
https://arxiv.org/abs/2509.01399

CabinSep: IR-Augmented Mask-Based MVDR for Real-Time In-Car Speech Separation with Distributed Heterogeneous Arrays
Separating overlapping speech from multiple speakers is crucial for effective human-vehicle interaction. This paper proposes CabinSep, a lightweight neural mask-based minimum variance distortionless response (MVDR) speech separation approach, to reduce speech recognition errors in back-end automatic speech recognition (ASR) models. Our contributions are threefold: First, we utilize channel information to extract spatial features, which improves the estimation of speech and noise masks. Second, …

DEPTHOR : Robust Depth Enhancement from a Real-World Lightweight dToF and RGB Guidance
Jijun Xiang, Longliang Liu, Xuan Zhu, Xianqi Wang, Min Lin, Xin Yang
https://arxiv.org/abs/2509.26498

DEPTHOR++: Robust Depth Enhancement from a Real-World Lightweight dToF and RGB Guidance
Depth enhancement, which converts raw dToF signals into dense depth maps using RGB guidance, is crucial for improving depth perception in high-precision tasks such as 3D reconstruction and SLAM. However, existing methods often assume ideal dToF inputs and perfect dToF-RGB alignment, overlooking calibration errors and anomalies, thus limiting real-world applicability. This work systematically analyzes the noise characteristics of real-world lightweight dToF sensors and proposes a practical and n…

Unlocking the Effectiveness of LoRA-FP for Seamless Transfer Implantation of Fingerprints in Downstream Models
Zhenhua Xu, Zhaokun Yan, Binhan Xu, Xin Tong, Haitao Xu, Yourong Chen, Meng Han
https://arxiv.org/abs/2509.00820

Unlocking the Effectiveness of LoRA-FP for Seamless Transfer Implantation of Fingerprints in Downstream Models
With the rapid advancement of large language models (LLMs), safeguarding intellectual property (IP) has become increasingly critical. To address the challenges of high costs and potential contamination in fingerprint integration, we propose LoRA-FP, a lightweight, plug-and-play framework that embeds backdoor fingerprints into LoRA adapters through constrained fine-tuning. This design enables seamless fingerprint transplantation via parameter fusion, eliminating the need for full-parameter updat…

This is frickin' brilliant: #TimMinchin, you lil' beauty.

ToolBrain: A Flexible Reinforcement Learning Framework for Agentic Tools
Quy Minh Le, Minh Sao Khue Luu, Khanh-Tung Tran, Duc-Hai Nguyen, Hoang-Quoc-Viet Pham, Quan Le, Hoang Thanh Lam, Hoang D. Nguyen
https://arxiv.org/abs/2510.00023

ToolBrain: A Flexible Reinforcement Learning Framework for Agentic Tools
Effective tool use is essential for agentic AI, yet training agents to utilize tools remains challenging due to manually designed rewards, limited training data, and poor multi-tool selection, resulting in slow adaptation, wasted computational resources, and suboptimal performance. We introduce ToolBrain, a lightweight and user-friendly framework for coaching tool use in agentic models with flexible reinforcement learning (RL), easing the barriers for researchers and practitioners to adapt LLM-…

Bingo: "awareness is growing that digital autonomy is no longer a luxury but a prerequisite for a well-functioning democracy." https://ginaplat.com/posts/20251031-digital-sovereignty-and-gov/ Thanks for this, @…

Damn - https://www.rnz.co.nz/news/national/577448/watch-the-moment-a-dramatic-blaze-ignites-in-whangarei-hospital-s-carpark imagine if the person in that ICE ute had a blanket and had thrown it over over the fire at th…

Faster and Better: Reinforced Collaborative Distillation and Self-Learning for Infrared-Visible Image Fusion
Yuhao Wang, Lingjuan Miao, Zhiqiang Zhou, Yajun Qiao, Lei Zhang
https://arxiv.org/abs/2509.02424

Faster and Better: Reinforced Collaborative Distillation and Self-Learning for Infrared-Visible Image Fusion
Infrared and visible image fusion plays a critical role in enhancing scene perception by combining complementary information from different modalities. Despite recent advances, achieving high-quality image fusion with lightweight models remains a significant challenge. To bridge this gap, we propose a novel collaborative distillation and self-learning framework for image fusion driven by reinforcement learning. Unlike conventional distillation, this approach not only enables the student model t…

The Name-Free Gap: Policy-Aware Stylistic Control in Music Generation
Ashwin Nagarajan, Hao-Wen Dong
https://arxiv.org/abs/2509.00654 https://arxiv.org/pdf…

The Name-Free Gap: Policy-Aware Stylistic Control in Music Generation
Text-to-music models capture broad attributes such as instrumentation or mood, but fine-grained stylistic control remains an open challenge. Existing stylization methods typically require retraining or specialized conditioning, which complicates reproducibility and limits policy compliance when artist names are restricted. We study whether lightweight, human-readable modifiers sampled from a large language model can provide a policy-robust alternative for stylistic control. Using MusicGen-small…

Multi-Agent Data Visualization and Narrative Generation
Anton Wolter, Georgios Vidalakis, Michael Yu, Ankit Grover, Vaishali Dhanoa
https://arxiv.org/abs/2509.00481 https://

Multi-Agent Data Visualization and Narrative Generation
Recent advancements in the field of AI agents have impacted the way we work, enabling greater automation and collaboration between humans and agents. In the data visualization field, multi-agent systems can be useful for employing agents throughout the entire data-to-communication pipeline. We present a lightweight multi-agent system that automates the data analysis workflow, from data exploration to generating coherent visual narratives for insight communication. Our approach combines a hybrid…

Enhancing Fitness Movement Recognition with Attention Mechanism and Pre-Trained Feature Extractors
Shanjid Hasan Nishat, Srabonti Deb, Mohiuddin Ahmed
https://arxiv.org/abs/2509.02511

Enhancing Fitness Movement Recognition with Attention Mechanism and Pre-Trained Feature Extractors
Fitness movement recognition, a focused subdomain of human activity recognition (HAR), plays a vital role in health monitoring, rehabilitation, and personalized fitness training by enabling automated exercise classification from video data. However, many existing deep learning approaches rely on computationally intensive 3D models, limiting their feasibility in real-time or resource-constrained settings. In this paper, we present a lightweight and effective framework that integrates pre-trained…

Welp, that was a short and very damp (hail?!) dog walk here in Ōtautahi.

Decoupling Bidirectional Geometric Representations of 4D cost volume with 2D convolution
Xiaobao Wei, Changyong Shu, Zhaokun Yue, Chang Huang, Weiwei Liu, Shuai Yang, Lirong Yang, Peng Gao, Wenbin Zhang, Gaochao Zhu, Chengxiang Wang
https://arxiv.org/abs/2509.02415

Decoupling Bidirectional Geometric Representations of 4D cost volume with 2D convolution
High-performance real-time stereo matching methods invariably rely on 3D regularization of the cost volume, which is unfriendly to mobile devices. And 2D regularization based methods struggle in ill-posed regions. In this paper, we present a deployment-friendly 4D cost aggregation network DBStereo, which is based on pure 2D convolutions. Specifically, we first provide a thorough analysis of the decoupling characteristics of 4D cost volume. And design a lightweight bidirectional geometry aggrega…

Very good to see this! https://www.zdnet.com/article/this-european-military-just-ditched-microsoft-for-open-source-libreoffice-heres-why/ h/t to @…

ProfVLM: A Lightweight Video-Language Model for Multi-View Proficiency Estimation
Edoardo Bianchi, Jacopo Staiano, Antonio Liotta
https://arxiv.org/abs/2509.26278 https://

ProfVLM: A Lightweight Video-Language Model for Multi-View Proficiency Estimation
Existing approaches to skill proficiency estimation often rely on black-box video classifiers, ignoring multi-view context and lacking explainability. We present ProfVLM, a compact vision-language model that reformulates this task as generative reasoning: it jointly predicts skill level and generates expert-like feedback from egocentric and exocentric videos. Central to our method is an AttentiveGatedProjector that dynamically fuses multi-view features, projected from a frozen TimeSformer backb…

It's inspiring to see someone as sharp and insightful - and principled - as this woman from Cambridge. She is a formidable debate opponent. And it's delightful to see how unabashedly self-possessed she is. https://yewtu.be/watch?v=Zn0_2iACV-A or

Was it you, @…, who mentioned #Beszel (https://beszel.dev) as a monitoring solution?…

Interpret, prune and distill Donut : towards lightweight VLMs for VQA on document
Adnan Ben Mansour, Ayoub Karine, David Naccache
https://arxiv.org/abs/2509.26235 https://

Interpret, prune and distill Donut : towards lightweight VLMs for VQA on document
Recent advances in Visually-rich Document Understanding rely on large Vision-Language Models like Donut, which perform document-level Visual Question Answering without Optical Character Recognition. Despite their effectiveness, these models are too costly for real-time or resource-constrained applications. We investigate model compression through knowledge distillation, training compact student models from a larger teacher. We leverage mechanistic interpretability to drive student architecture …

For the record, I'm pretty sure that no one should care what Bill Gates thinks. He's missed/misjudged every single technological and social leap in the past 30 years. I'm amazed he's made any money at all, but I guess that's what multi-million $ trust funds and a willingness to ruthlessly exploit ill-gotten monopolies will get you.

Can someone remind me why no national leaders (so far as I know) are clamouring for a global, independent tax authority to rein in exploitative global corporations (and their attendant billionaires)?

On a side note, where are the Chinese-made EV station wagons? That most perfect of form factors seems to be a dying breed (and all of the current models European except for maybe Japanese Subarus, but even they're looking very bubbly-SUVy to me these days) - it's all hatches and inflated SUVs... yuck.

Damn, that was lovely.

I'm right chuffed! Even just after watering it, and having the door open for a while, our greenhouse is between 25-40 degrees C, in the last week of winter. We've already got a bunch of seedlings coming up. I think it's going to be a good summer for veggies in our household. Now if I can just find a good spot to plant that Avocado tree... 🤔

It looks like the world might finally be doing something. The words from the Colombian and Finnish presidents are especially poignant! The US (like Russia and China) is quickly isolating itself on the global stage. https://yewtu.be/watch?v=vvSsGE6UwsI or

This is a great talk to the UN General Assembly from the Nobel Peace Prize winner, Maria Ressa: https://yewtu.be/watch?v=d4kizE2LgUU or https://www.you…

This was a very good interview on climate change and implications in Aotearoa NZ: https://www.rnz.co.nz/national/programmes/saturday/audio/2019006040/tim-lenton-how-to-fix-the-climate-crisis

And, boom, with that, all my servers and VPSs are monitored (25 total, with 3 hubs). Thank you, henrygd - https://github.com/henrygd - for Beszel - https://beszel.dev/

Yeah, folks getting locked out of bigtech socials is too bad... https://www.rnz.co.nz/news/national/571326/musician-pacific-leader-northland-grandmother-dozens-of-kiwis-social-accounts-banned but it's en…

@… I've noticed that most of the recent posts on the OEGlobal Connect forum are boosting 'AI'. I'm wondering if folks are familiar with this rather comprehensive article from @…:

Nice, fairly long gentle swaying just now #eqnz... anyone else in Ōtautahi notice it?

Over the weekend, I saw the first skink of the season taking in the sun in our "Lizardarium"...

Oh happy day! Saw my first ever tūī in Ōtautahi! It was feeding in a flowering tree I didn't recognize on the north eastern hillside of our valley, below the Sign of the Kiwi...

What the? Just saw a DeLorean driving on the road in Ōtautahi!

Whelp, after a stunning summer morning here in Ōtautahi, the temp's just suddenly plummeted, dark clouds rapidly gathered, and the rain has begun.

Ooo, my small parcel form King's Seeds has just arrived **rubs hands**... my greenhouse is about to have some new residents...

Hah - interesting that this was considered newsworthy: https://www.rnz.co.nz/news/national/576752/picton-man-uses-electric-car-to-cook-hot-breakfast-amid-power-outages during the ongoing windstorm that's knocked out po…

Huh, how time flies. 6 or 7 years ago, at the OER Foundation, for whom I work as Open Source Technologist, we decided to replace our MailChimp with, eventually, a self-hosted instance of Mautic (https://mautic.org) to communicate with learners taking courses in our online open university courses (

Hmm - judging by the smoke, looks like there's a bush fire on the hills to the east of us somewhere (south Ōtautahi)... I'm hoping someone's on that.

This is something we should all be pondering: Big tech has most of my email, because it has *all* of yours: https://mako.cc/copyrighteous/google-has-most-of-my-email-because-it-has-all-of-yours

Managed to dig out my cardboard eclipse glasses from last years total eclipse in the US. Had a great view! Lovely morning here in Ōtautahi!

Puking in my mouth, hearing about the lionisation of Charlie Kirk. So entirely wrong. He's no hero. Just an unfortunate victim of US-style gun violence, loud-mouthed, largely ignorant, wrong-thinking agitator. Based on my exposure to him (more than I'd like) he did very little in 'good faith'. Nothing admirable to be seen.

Fascinating - https://www.rnz.co.nz/news/te-manu-korihi/573694/from-new-jersey-to-aotearoa-american-teen-dedicates-year-to-te-reo-maori - good on Charlie, and Te Wānanga Takiura! I come from near where she does, w…

There are at least two Aotearoa native grass skinks in this picture... Can you see them?
#Herptophile

Wow, it's a stunner here in Ōtautahi, but that nor'west wind is formidable! Hope the greenhouse holds up to the onslaught!

I wonder how many digital service businesses who have been affected by the AWS outage... have all their comms to their customers also happening through that same service... Wonder how many of them recognised that issue and went 🤷‍♀️ we'll be fine.

Most days I'm reminded of this: 'expedience is the absence of principle'...

Many parts of the Internet I know and loath are mostly down. I hope this helps shake confidence in US BigTech. Probably won't, because if people were rational, they wouldn't have built the digital world we're currently forced to inhabit, but who knows, it might. Even the vague possibility of it has made my day better as I head to bed.

For those not afraid of foreign language (Deutsch) music, this is superb (in my opinion). At the start he says "just taking the voice for a walk". If you don't already know, you'll soon hear why. https://www.youtube.com/watch?v=NANdY_68PFU

A bit of understated genius for you on a Weds evening. https://yewtu.be/watch?v=ASi7lyg5gRI or https://www.youtube.com/watch?v=ASi7lyg5gRI

Anybody here heard of 'FlexOS'? Apparently it's 'free and lightweight'... https://www.education.govt.nz/bulletins/digital-download/30-07-25 It's what the Min of Ed is recommending for school devices that can't be upgraded to Windows 11 when …

This is quite splendid. Why are we settling for either Chris when we could have someone like this steering the waka?! https://www.youtube.com/watch?v=rtSsFV9eC1Y

Cool, after having it sitting on my desk for a few months, doing a whole lotta nuthin', I've just updated the firmware on my Liligo-T-echo LoRA device, and got it connected to my phone... Local community-based (locally sovereign) coms, here we come! #meshtastic Thanks to @…

To all our Aotearoa-based Fediverse community - if you're eligible, have you voted in our local elections?

Hazzah & yikes! Just saw the matriarch skink in our garden - we call her 3-Legs, because she's missing her left front leg. This is my first time seeing her since last fall. She's looking rather worse for the wear. From my quick glimpse, seems she's lost most of her tail and appears to have a mostly healed wound on her side. Here's hoping she's able to recover - and perhaps even parent another brood of skinklets - this summer! It's a pretty harsh world for skinks o…

I just wrote a new blog post... "It's Aotearoa".... https://davelane.nz/its-aotearoa