Strain sensors have been attracting tremendous attention for the promising application of wearable devices in recent years. However, trade-off between high resolution, sensitivity, and broad detection range is a great challenge strain sensors. Herein, novel design hierarchical synergistic structure (HSS) Au micro cracks carbon black (CB) nanoparticles reported to overcome this challenge. The sensor based on designed HSS exhibit sensitivity (GF > 2400), resolution (0.2%) even under large...

Wearable strain sensors hold immense promise in monitoring human motion activities due to their low cost, lightweight design, and excellent biocompatibility. For example, continuous real-time of neck activity can effectively prevent the onset acute torticollis. However, current approaches sleep posture primarily depend on technologies such as computer vision, which are characterized by limited wearability portability issues. Herein, this work introduces a cost-effective, highly sensitive...

We propose a compressive light field display of wide viewing angle with multi-directional backlight. Displayed layer images sub-viewing zones are synchronized the Viewers can perceive three-dimensional scene large based on persistence vision.

Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities across a range of multimodal tasks. However, their inference efficiency is constrained by the large number visual tokens processed during decoding. To address this challenge, we propose Per-Layer Per-Head Vision Token Pruning (PLPHP), two-level fine-grained pruning method including Layer-Level Retention Rate Allocation and Head-Level Pruning. Motivated Re-attention phenomenon decoder layers, dynamically adjust...

ABSTRACT Stretchable electronic devices play a key role in bridging various components, especially flexible and complex multifunctional logic circuits, as well information recognition transmission systems. The stretchability stability of solid metal‐based stretchable electrodes are still limited due to their intrinsic rigidity. Liquid metals one the most popular materials for high conductivity, flexibility, machinability. However, surface tension liquid is extremely at room temperature,...

To address the challenges of high labor intensity and expensive manpower in morel harvesting, a specialized picking machine was designed. It is capable performing tasks grasping, cutting, collecting morels. A three-degree-of-freedom robotic arm incorporated to ensure precise positioning claw. The kinematic model this established determine operational radius for effective picking. three-finger claw developed, with synchronous opening closing three fingers through wire-pulling mechanism...

This research proposes a deep-learning paradigm, termed functional learning (FL), to physically train loose neuron array, group of non-handcrafted, non-differentiable, and loosely connected physical neurons whose connections gradients are beyond explicit expression. The paradigm targets training non-differentiable hardware, therefore solves many interdisciplinary challenges at once: the precise modeling control high-dimensional systems, on-site calibration multimodal hardware imperfectness,...

Glasses-free light field displays have significantly progressed due to advances in high-resolution microdisplays and high-end graphics processing units (GPUs). However, for near-eye light-field requiring portability, the fundamental trade-off regarding achieved spatial resolution remains: retinal blur quality must be degraded; otherwise, computational consumption increases. This has prevented synthesizing high-quality from being fast. By integrating off-the-shelf gaze tracking modules into...

A multi-directional backlight autostereoscopic display system with high resolution, low crosstalk, and motion parallax is developed in this paper. The proposed based on the Bragg mismatched reconstruction of volume holographic optical element (VHOE), includes a set light sources which are uniformly arrayed along one direction. Each source produces its corresponding directional lighting to follow human eye position detected by an tracker. Two scenarios presented build system. prism-type...

Membership Inference Attacks (MIA) aim to infer whether a target data record has been utilized for model training or not. Prior attempts have quantified the privacy risks of language models (LMs) via MIAs, but there is still no consensus on existing MIA algorithms can cause remarkable leakage practical Large Language Models (LLMs). Existing MIAs designed LMs be classified into two categories: reference-free and reference-based attacks. They are both based hypothesis that records consistently...

Abstract Additive light field displays are transparent autostereoscopic three‐dimensional without backlights, thus suitable for augmented reality applications. However, when the parallax between viewpoint images becomes large with increase of viewing angle, optimization algorithm is hard to handle too many dissimilarities evenly on all viewpoints, resulting in poor reconstructed quality. This paper presents an additive display using weighted simultaneous algebraic reconstruction technique...

As the most commonly used exterior-protecting components for buildings, window glass, due to its typical brittle characteristics, is more prone break under action of blast loads and produce high-speed flying fragments, which can pose a great threat personal safety. This paper studies influence size glasses on their failure characteristics. A simplified numerical simulation-based method evaluating P–I (pressure–impulsive) curve glass proposed. The influences length–width ratio, area,...

Membership Inference Attack (MIA) identifies whether a record exists in machine learning model's training set by querying the model. MIAs on classic classification models have been well-studied, and recent works started to explore how transplant MIA onto generative models. Our investigation indicates that existing designed for mainly depend overfitting target However, can be avoided employing various regularization techniques, whereas demonstrate poor performance practice. Unlike...

Numerical simulation was performed on an axi-symmetric metallic die during a whole fife cycle. It shown that most important stresses are generated by the thermal gradients heating up and repeated removal from press furnace also clamping pressure forming. The forming itself causes non dimensioning stresses. Creep relaxation induces plastic deformation distortion of mould after cooling down. To increase relevance, new behaviour models investigated for three classes materials candidate SPF...

Photonic memory provides optics the opportunity for upgrading data storage hardware and shows great promise in future computing due to its innate superiority speed, power consumption bandwidth. Chalcogenide phase-change materials (PCMs) such as Ge 2 Sb Te 5 (GST) are one of most promising non-volatile have been used on-chip photonic integration circuits (PIC). Here, we explore transmission contrast thermal distribution produced by different spatial structures GST cells including geometric...

We apply an iterative weighting scheme for additive light field synthesis. Unlike previous work optimizing evenly over viewpoints, we constrain the optimization to deliver a reconstructed of high image quality viewpoints large weight.

Augmented reality (AR) near-eye displays have significantly progressed due to advances in nanostructure fabrication. However, for diffractive waveguide AR requiring exit pupil expansion, the angular uniformity of each position still needs improve. In this paper, an improvement method based on region geometry optimization is proposed. This essentially introduces interaction number light with grating as one variables manipulate energy distribution. distribution obtained by rigorous coupled...

Abstract Glioblastoma multiforme (GBM), one of the most aggressive brain cancers, presents substantial therapeutic challenges, particularly concerning postsurgical recurrence and inherent drug resistance. In this study, a nanofiber‐based delivery system integrating platinum pro‐drugs hemoglobin into mesoporous silica nanoparticles (HB/HSPt@MS NPs) is reported, which are surface‐modified with poly(ethylene glycol) (PEG) targeting molecules, subsequently embedded nanofibers using an...

Stance detection automatically detects the stance in a text towards target, vital for content analysis web and social media research. Despite their promising capabilities, LLMs encounter challenges when directly applied to detection. First, demands multi-aspect knowledge, from deciphering event-related terminologies understanding expression styles platforms. Second, requires advanced reasoning infer authors' implicit viewpoints, as are often subtly embedded rather than overtly stated text....

We implement a graphic processing unit acceleration program of weighted simultaneous algebra reconstruction technique solver for additive light field synthesis. The factorized results are verified by comparing them with MATLAB implementation.

Recently, glass-free light field displays of multi-layer architecture have gradually entered the commercial stage. However, for near-eye displays, rendering still suffers from expensive computational costs. It can hardly achieve an acceptable framerate real-time displays. This work develops a novel display pipeline that uses two gaze maps to reconstruct patterns foveated vision effect. With acceleration GPU and emerging eye-tracking technique, cone be updated instantaneously. The...