Real-time speech communication over wireless networks remains challenging, as conventional channel protection mechanisms cannot effectively counter packet loss under stringent bandwidth and latency constraints. Semantic communication has emerged as a promising paradigm for enhancing the robustness of speech transmission by means of joint source channel coding (JSCC). However, its cross-layer design hinders practical deployment due to the incompatibility with existing digital communication systems. To address this, we perform JSCC over the network layer to combat packet loss and support real deployment. Inspired by the generative latent modeling, we propose Glaris, a generative latent-prior-based resilient speech semantic communication framework that performs resilient transform coding in the generative latent space. Generative latent priors enable high-quality packet loss concealment (PLC) at the receiver side, well-balancing semantic consistency and reconstruction fidelity. Additionally, an integrated error resilience mechanism is designed to mitigate the error propagation and improve the effectiveness of PLC. Compared with traditional packet-level forward error correction (FEC) strategies, our new method achieves enhanced robustness over dynamic wireless networks while reducing redundancy overhead significantly. Experimental results on the LibriSpeech dataset demonstrate that Glaris consistently outperforms existing error-resilient codecs, achieving JSCC-level robustness while maintaining seamless compatibility with existing systems, and it also strikes a favorable balance between transmission efficiency and error resilience.

Deep learning (DL)-based channel state information (CSI) feedback holds substantial promise for boosting spectrum efficiency in massive MIMO systems. However, prevailing studies often treat compressed CSI bits uniformly, assuming their accurate transmission over noisy channels. Such assumptions falter when confronted with bandwidth limitations or low signal-to-noise ratios (SNRs), significantly impairing CSI reconstruction quality. In this paper, we introduce a novel soft variational approach to implement deep joint source-channel coding for CSI feedback within noisy environments, conceptualizing the system as an end-to-end rate-distortion (RD) optimization framework. Specifically, our model utilizes a nonlinear transform to extract the latent representation of CSI and employs an entropy model as a prior to guide the variable-rate transmission of latent features across noisy channels. Experimental results demonstrate that our approach achieves substantial improvements over benchmark schemes, saving 40% in feedback bandwidth under normalized mean squared error (NMSE) and ensuring the robustness to varying wireless channels.