Tootfinder

GradOT: Training-free Gradient-preserving Offsite-tuning for Large Language Models
Kai Yao, Zhaorui Tan, Penglei Gao, Lichun Li, Kaixin Wu, Yinggui Wang, Yuan Zhao, Yixin Ji, Wei Wang, Jianke Zhu
https://arxiv.org/abs/2507.04455

GradOT: Training-free Gradient-preserving Offsite-tuning for Large Language Models
The rapid growth of large language models (LLMs) with traditional centralized fine-tuning emerges as a key technique for adapting these models to domain-specific challenges, yielding privacy risks for both model and data owners. One promising solution, called offsite-tuning (OT), is proposed to address these challenges, where a weaker emulator is compressed from the original model and further fine-tuned with adapter to enhance privacy. However, the existing OT-based methods require high computa…

Velocity optimization of self-equilibrated obstacles in a two-dimensional viscous flow
Gilles A. Francfort, Alessandro Giacomini, Scott Weady
https://arxiv.org/abs/2508.05481 ht…

Velocity optimization of self-equilibrated obstacles in a two-dimensional viscous flow
An obstacle is immersed in an externally driven 2D Stokes or Navier-Stokes fluid. We study the self-equilibration conditions for that obstacle under steady state assumptions on the flow. We then seek to optimize the translational and/or angular velocity of the obstacle by varying its shape. To allow general variations, we must consider a very large class of obstacles for which the notion of trace is meaningless. This forces us to revisit the notion of self-equilibration for both Stokes and Navi…

ECOLogic: Enabling Circular, Obfuscated, and Adaptive Logic via eFPGA-Augmented SoCs
Ishraq Tashdid, Dewan Saiham, Nafisa Anjum, Tasnuva Farheen, Sazadur Rahman
https://arxiv.org/abs/2508.04516

ECOLogic: Enabling Circular, Obfuscated, and Adaptive Logic via eFPGA-Augmented SoCs
Traditional hardware platforms - ASICs and FPGAs - offer competing trade-offs among performance, flexibility, and sustainability. ASICs provide high efficiency but are inflexible post-fabrication, require costly re-spins for updates, and expose IPs to piracy risks. FPGAs offer reconfigurability and reuse, yet suffer from substantial area, power, and performance overheads, resulting in higher carbon footprints. We present ECOLogic, a hybrid design paradigm that embeds lightweight eFPGA fabric wi…

Safe Bimanual Teleoperation with Language-Guided Collision Avoidance
Dionis Totsila, Clemente Donoso, Enrico Mingo Hoffman, Jean-Baptiste Mouret, Serena Ivaldi
https://arxiv.org/abs/2507.04791

Safe Bimanual Teleoperation with Language-Guided Collision Avoidance
Teleoperating precise bimanual manipulations in cluttered environments is challenging for operators, who often struggle with limited spatial perception and difficulty estimating distances between target objects, the robot's body, obstacles, and the surrounding environment. To address these challenges, local robot perception and control should assist the operator during teleoperation. In this work, we introduce a safe teleoperation system that enhances operator control by preventing collisions i…

RunPacer: A Smartwatch-Based Vibrotactile Feedback System for Symmetric Co-Running by Visually Impaired Individuals and Guides
Yichen Yu, Huan-Song Xu
https://arxiv.org/abs/2507.04241

RunPacer: A Smartwatch-Based Vibrotactile Feedback System for Symmetric Co-Running by Visually Impaired Individuals and Guides
Visually impaired individuals often require a guide runner to safely participate in outdoor running. However, maintaining synchronized pacing with verbal cues or tethers can be mentally taxing and physically restrictive. Existing solutions primarily focus on navigation or obstacle avoidance but overlook the importance of real-time interpersonal rhythm coordination during running. We introduce RunPacer, a smartwatch-based vibrotactile feedback system that delivers synchronized rhythmic pulses to…

UWB TDoA Error Correction using Transformers: Patching and Positional Encoding Strategies
Dieter Coppens, Adnan Shahid, Eli De Poorter
https://arxiv.org/abs/2507.03523

UWB TDoA Error Correction using Transformers: Patching and Positional Encoding Strategies
Despite their high accuracy, UWB-based localization systems suffer inaccuracies when deployed in industrial locations with many obstacles due to multipath effects and non-line-of-sight (NLOS) conditions. In such environments, current error mitigation approaches for time difference of arrival (TDoA) localization typically exclude NLOS links. However, this exclusion approach leads to geometric dilution of precision problems and this approach is infeasible when the majority of links are NLOS. To a…

Complex symplectic structures: deformations and cohomology
Giovanni Bazzoni, Marco Freibert, Adela Latorre, Nicoletta Tardini
https://arxiv.org/abs/2507.03940

Complex symplectic structures: deformations and cohomology
We show that complex symplectic structures need not be preserved under small deformations, and we find sufficient conditions for this to happen. We study various cohomologies of compact complex symplectic manifolds, obtaining some topological obstructions to their existence.

Fractional Infinity Laplacian with Obstacle
Samer Dweik, Ahmad Sabra
https://arxiv.org/abs/2507.04328 https://arxiv.org/pdf/2507.0432…

Fractional Infinity Laplacian with Obstacle
This paper deals with the obstacle problem for the fractional infinity Laplacian with nonhomogeneous term $f(u)$, where $f:\mathbb{R}^+ \mapsto \mathbb{R}^+$: $$\begin{cases} L[u]=f(u) &\qquad in \{u>0\}\\ u \geq 0 &\qquad in\, Ω\\ u=g &\qquad on\, \partial Ω\end{cases},$$ with $$L[u](x)=\sup_{y\in Ω,\,y\neq x}\dfrac{u(y)-u(x)}{|y-x|^α}+\inf_{y\in Ω,\,y\neq x} \dfrac{u(y)-u(x)}{|y-x|^α},\qquad 0<α<1.$$ Under the assumptions that $f$ is a continuous and monotone functi…

Rapid and Safe Trajectory Planning over Diverse Scenes through Diffusion Composition
Wule Mao, Zhouheng Li, Yunhao Luo, Yilun Du, Lei Xie
https://arxiv.org/abs/2507.04384

Rapid and Safe Trajectory Planning over Diverse Scenes through Diffusion Composition
Safe trajectory planning remains a significant challenge in complex environments, where traditional methods often trade off computational efficiency for safety. Comprehensive obstacle modeling improves safety but is computationally expensive, while approximate methods are more efficient but may compromise safety. To address this issue, this paper introduces a rapid and safe trajectory planning framework based on state-based diffusion models. Leveraging only low-dimensional vehicle states, the d…

"Hi AirStar, Guide Me to the Badminton Court."
Ziqin Wang, Jinyu Chen, Xiangyi Zheng, Qinan Liao, Linjiang Huang, Si Liu
https://arxiv.org/abs/2507.04430

"Hi AirStar, Guide Me to the Badminton Court."
Unmanned Aerial Vehicles, operating in environments with relatively few obstacles, offer high maneuverability and full three-dimensional mobility. This allows them to rapidly approach objects and perform a wide range of tasks often challenging for ground robots, making them ideal for exploration, inspection, aerial imaging, and everyday assistance. In this paper, we introduce AirStar, a UAV-centric embodied platform that turns a UAV into an intelligent aerial assistant: a large language model a…