Graph Neural Architecture Search (GNAS) has achieved superior performance on various graph-structured tasks. However, existing GNAS studies overlook the applications of GNAS in resource-constrained scenarios. This paper proposes to design a joint graph data and architecture mechanism, which identifies important sub-architectures via the valuable graph data. To search for optimal lightweight Graph Neural Networks (GNNs), we propose a Lightweight Graph Neural Architecture Search with Graph SparsIfication and Network Pruning (GASSIP) method. In particular, GASSIP comprises an operation-pruned architecture search module to enable efficient lightweight GNN search. Meanwhile, we design a novel curriculum graph data sparsification module with an architecture-aware edge-removing difficulty measurement to help select optimal sub-architectures. With the aid of two differentiable masks, we iteratively optimize these two modules to efficiently search for the optimal lightweight architecture. Extensive experiments on five benchmarks demonstrate the effectiveness of GASSIP. Particularly, our method achieves on-par or even higher node classification performance with half or fewer model parameters of searched GNNs and a sparser graph.
Continuous monitoring of arterial pulse has great significance for detecting the early onset of cardiovascular disease and assessing health status, while needs pressure sensors with high sensitivity and signal-to-noise ratio (SNR) to accurately capture more health information concealed in pulse waves. Field effect transistors (FETs) combined with the piezoelectric film is an ultrahigh sensitive pressure sensor category, especially when the FET works in the subthreshold regime, where the signal enhancement effect on the piezoelectric response is the most effective. However, controlling the work regime of FET needs extra external bias assistance which will interfere with the piezoelectric response signal and complicate the test system thus making the scheme difficult to implement. Here, we described a gate dielectric modulation strategy to match the subthreshold region of the FET with the piezoelectric output voltage without external gate bias, finally enhancing the sensitivity of the pressure sensor. A carbon nanotube field effect transistor and polyvinylidene fluoride (PVDF) together form the pressure sensor with a high sensitivity of 7 × 10−1 kPa−1 for a pressure range of 0.038–0.467 kPa and 6.86 × 10−2 kPa−1 for a pressure range of 0.467–15.5 kPa, SNR, and the ability to continuously monitor pulse in real-time. Additionally, the sensor enables high-resolution detection of weak pulse signals under large static pressure.