Skip to main content
Search

University of Shanghai for Science and Technology

3 related articles

Subscribe to updates
  1. Co-frequency signal separation at 235K parameters: complex attention cuts collapse and keeps phase

    Single-channel blind separation of co-frequency communication signals has relied on real-valued networks or attention modules blind to complex structure, which cannot exploit the phase coupling between the I and Q rails. ComplexLightweightSepNet embeds a complex squeeze-and-excitation block in a complex convolutional encoder-decoder, scaling complex features with a real-valued channel weight that preserves phase, at 235,525 parameters. On 5 Hz co-frequency synthetic mixtures across four modulations it reaches a five-seed mean SDR of 2.31 dB with 0.62 dB standard deviation; paired Wilcoxon tests do not reach alpha = 0.05, so the result is positioned as a working point at the 235K scale, not an inferred advantage.

  2. Arctic sea ice forecasts drop recurrent nets: causal time convolution plus state-space encoding beats MT-IceNet at every lead from one to six months

    Arctic sea ice concentration forecasting has been limited by the difficulty of fusing multi-timescale information, with statistical models missing nonlinear relationships and numerical models depending on initial conditions. STICEnet replaces the U-Net encoder-decoder path with strictly unidirectional temporal causal and transposed convolutions and embeds channel and spatial attention in the skip connections, trained on NSIDC sea ice concentration and ERA5 atmospheric and oceanic variables. Across one- to six-month leads its RMSE, MAE and R2 all beat CNN, ConvLSTM, IceNet and MT-IceNet, reaching 7.51% RMSE and 0.925 R2 at month six.

  3. U-Net-Based Automated Quality Control of Knee Radiographs: Dual-Center Validation and Clinical Intervention

    This study developed and dual-center validated an interpretable U-Net-based AI framework for automated quality control (QC) of knee anteroposterior (AP) and lateral (LAT) radiographs, generating QC indices through anatomical segmentation and landmark localization, with mean Dice similarity coefficients of 0.964 and 0.936 in internal and external validation, most intraclass correlation coefficients exceeding 0.90, QC sensitivity of 91.67% to 98.36% and specificity of 89.13% to 99.10%; separately, six radiographers who received 4 weeks of AI-based feedback on 948 radiographs from 474 patients showed exploratory numerical gains in sensitivity, with effect sizes of 0.30 to 0.73.

You’re all caught up