Learning to Interfere in Non-Orthogonal Multiple-Access Joint Source-Channel Coding
Authors: Selim F. Yilmaz, Can Karamanli, Deniz Gunduz
Organizations: Department of Electrical and Electronic Engineering, Imperial College London, United Kingdom. · School of Biomedical Engineering & Imaging Sciences, King’s College London, United Kingdom.
We consider multiple transmitters aiming to communicate their source signals (e.g., images) over a multiple access channel (MAC). Conventional communication systems minimize interference by orthogonally allocating resources (time and/or bandwidth) among users, which limits their capacity. We introduce a machine learning (ML)-aided wireless image transmission method that merges compression and channel coding using a multi-view autoencoder, which allows the transmitters to use all the available channel resources simultaneously, resulting in a non-orthogonal multiple access (NOMA) scheme. The receiver must recover all the images from the received superposed signal, while also associating each image with its transmitter. Traditional ML models deal with individual samples, whereas our model allows signals from different users to interfere in order to leverage gains from NOMA under limited bandwidth and power constraints. We introduce a progressive fine-tuning algorithm that doubles the number of users at each iteration, maintaining initial performance with orthogonalized user-specific projections, which is then improved through fine-tuning steps. Remarkably, our method supports up to 64 users with potential for further scaling, with only a 2.4% increase in the number of trainable parameters compared to a single-user model, significantly enhancing recovered image quality and outperforming existing NOMA-based methods over a wide range of datasets, metrics, and channel conditions. Our approach paves the way for more efficient and robust multi-user communication systems, leveraging innovative ML components and strategies.
Figures & tables
Fig. 1: Extension of DeepJSCC to two users via TDMA.
Paper
Demonstrated Max # Users
Dataset(s)
Channel(s)
Coding
Modality
Uplink/Downlink
Ours
64
CIFAR,TinyImagenet, Cityscapes,Kodak
AWGN, Rayleigh
JSCC
Image
Uplink
DeepJSCC-NOMA [ 19 ]
2
CIFAR
AWGN
JSCC
Image
Uplink
DBC Aware JSCC [ 20 ]
2
CIFAR
AWGN
JSCC
Image
Downlink
NOMASC [ 21 ]
3
MNIST,CIFAR,Europarl
AWGN, Rayleigh
JSCC
Image, Text
Downlink
DeepSCM [ 22 ]
2
CIFAR
AWGN
JSCC
Image
Downlink
IS-SNOMA [ 23 ]
2
Cityscapes
Rician
JSCC
Image
Downlink
TABLE I: Comparison of Deep Learning Based Transmission Methods for NOMA
Fig. 2: Overall architecture of our DeepJSCC-PNOMA method. Due to the design of the user-specific projection matrices in DeepJSCC-PNOMA—which preserve the power of each encoded signal and enforce mutual orthogonality among signals—the system with n users initially behaves as if it were time-sharing among 2n users under fixed per-user average power Pˉavg and per-user bandwidth ratio ρˉ . In this configuration, each user’s signal retains its original power and remains free from interference, enabling independent decoding. Consequently, the network can begin fine-tuning without any performance degradation from inter-user interference.
Fig. 3: Encoder architecture, user-specific projection layer and power normalization layer of the introduced method.
Fig. 4: User-specific projection layer and decoder architecture of the introduced method.
Algorithm 1 Training and Fine-tuning of DeepJSCC-PNOMA
Fig. 5: Progressive fine-tuning methodology for scaling DeepJSCC-PNOMA from 1 to 16 users. The process iteratively doubles the number of users while maintaining training stability through orthogonal initialization. Each stage preserves signal power and ensures initial orthogonality between user groups, enabling convergence from interference-free configurations.
Algorithm 2 Progressive fine-tuning for doubling the number of users in DeepJSCC-PNOMA
Fig. 6: Comparison with point-to-point DeepJSCC and separation-based methods.
ρˉ
Pˉavg
σ2
Method
n
ρ
Pavg
1/6
1
1
DeepJSCC-TDMA
1
1/6
1
DeepJSCC-PNOMA
1
1/6
1
2
2/6
1/2
4
4/6
1/4
8
8/6
1/8
16
16/6
1/16
TABLE II: Employed hyperparameters for fair comparison between scenarios with different number of users
Fig. 7: Comparison with NOMA-based methods for different bandwidth ratios on CIFAR-10 [ 51 ] .
Fig. 8: Performance comparison in terms of PSNRon AWGNand Rayleigh channels for correlated source images from Cityscapes dataset.
Fig. 9: Comparison of DeepJSCC-PNOMA with number of users n∈{1,2,4,8,16} .
Method
# Users
ρˉ=1/6
ρˉ=1/12
DeepJSCC-TDMA
All
22200211
22149011
DeepJSCC-PNOMA (Ours)
1
22200211
22149011
2
22202259
22149523
4
22208403
22151059
8
22232979
22157203
16
22331283
22181779
TABLE III: Number of trainable model parameters
Fig. 10: Ablation study of the number of the sampled training pairs and the introduced components.
Fig. 11: Analysis of orthogonality between transmitted filters on Kodak.
Fig. 12: Qualitative comparison of reconstructed images.
Fig. 13: Visualization of transmitted filters for DeepJSCC-TDMA and DeepJSCC-PNOMA.
Fig. 14: Visualization of losses throughout the training or fine-tuning steps.
SNR (dB)
Bandwidth
Users
0
3
7
10
Ratio
ρˉ=1/6
n=16
25.96
28.37
31.28
33.27
n=32
26.03
28.42
31.33
33.32
n=64
26.05
28.46
31.34
33.29
ρˉ=1/12
n=16
23.04
25.07
27.49
29.03
TABLE IV: Scalability analysis: Performance comparison on CIFAR-10 under AWGN channel for different bandwidth ratios. Values shown as PSNR (dB). Bold values indicate best performance within each bandwidth ratio group.
Training Duration
Number of Epochs
Method
ρˉ=\nicefrac16
ρˉ=\nicefrac112
ρˉ=\nicefrac16
ρˉ=\nicefrac112
DeepJSCC-PNOMA ( n=1 )
1:04:38
1:50:16
54
80
DeepJSCC-PNOMA ( n=2 )
4:42:27
2:09:58
86
40
DeepJSCC-PNOMA ( n=4 )
3:49:57
2:34:43
42
28
DeepJSCC-PNOMA ( n=8 )
10:41:58
7:14:31
31
21
DeepJSCC-PNOMA ( n=16 )
20:24:06
3:40:45
33
11
TABLE V: Training durations and number of epochs passed before early stopping. Duration format is hours:minutes:seconds.
Fig. 15: Analysis of fairness between users on Kodak dataset.
Fig. 16: Comparison with point-to-point DeepJSCC and separation-based methods over SSIM, MS-SSIM and LPIPS metrics.
Fig. 17: Comparison with point-to-point DeepJSCC and separation-based methods with capacity over PSNR, SSIM, MS-SSIM and LPIPS metrics.
Fig. 18: Performance comparison in terms of MS-SSIMand LPIPSon AWGNand Rayleigh channels for correlated source images from Cityscapes dataset.
Fig. 19: Comparison of DeepJSCC-PNOMA for number of active users ∣P∣∈{1,2,4,8,16} to demonstrate the comprehensiveness of our method according to the Definition 2 .
Joint image compression and wireless transmission remain relatively underexplored compared to generic image restoration, despite its importance in practical communication systems. We formulate this problem under an equivalent linear model, and propose Diffusion-OAMP, a training-free reconstruction framework that embeds a pre-trained diffusion model into the OAMP algorithm. In Diffusion-OAMP, the OAMP linear estimator produces pseudo-AWGN observations, while the diffusion model serves as a nonlinear estimator under an SNR-matching rule. This framework offers a way to incorporate multiple generative priors into OAMP. Experiments with varying compression ratios and noise levels show that Diffusion-OAMP performs favorably against classic methods in the evaluated settings.
Wentao Hou, Yimin Bai, Zelei Luo +2
College of Information Science and Electronic Engineering, Zhejiang University, China
Digital image transmission typically optimizes source coding, channel coding, and modulation separately, limiting how source representations adapt to channel impairments. We propose SemQAM, a joint semantic-channel coding and modulation scheme for generative image transmission converting source to modulated symbols, such as standard QAM symbols. Channel impairments create two challenges: they can corrupt the latent representation at the receiver, and this corruption can compound through autoregressive coding, degrading reconstruction quality. First, SemQAM represents latent features with vector quantization (VQ), maps each VQ index to a QAM symbol, and updates the codeword corresponding to the detected symbol so that channel impairments directly affect codebook learning. Moreover, SemQAM uses the conditional expectation of received codewords to approximate the receiver reconstruction and reduce error propagation during autoregressive coding. SemQAM supports multiple SNRs and CBRs in a unified model, while also supporting conventional source coding through SemQAM-C. Experiments show that (i) under the same LDPC and QAM settings, SemQAM-C outperforms OneDC in LPIPS and DISTS with substantially lower complexity, showing that the unified model retains competitive capability for source coding. (ii) Compared with this separated counterpart, SemQAM reduces channel uses by 18.57% at matched LPIPS and outperforms the compared JSCC methods, demonstrating the benefit of joint coding.
Hao Cao, Jingkai Ying, Yinhuan Huang +2
Department of Electronic Engineering, Tsinghua University · State Key Laboratory of Space Network and Communications
Dense IoT networks require reliable communication despite limited spectrum and substantial multi-user interference while maintaining manageable receiver complexity. This work introduces a deep-learning-based end-to-end multi-user communication design for interference-limited finite-blocklength IoT scenarios, focusing on short and medium blocklengths. We extend a prior 2-user SiameseNet transceiver framework to accommodate 2, 4, and 8 users, leveraging learned redundancy for interference suppression and noise robustness. Compared to conventional non-orthogonal access baselines, our method demonstrates strong Block Error Rate (BLER) performance across various scenarios without resorting to joint detection; the per-user decoder scales roughly linearly with the number of users. Further, we examine the robustness under interference mismatch and unequal interference strengths, critical for practical deployments with heterogeneous devices. The Latent-space analysis reveals that the learned codeword distance increases as the effective per-user rate decreases, corroborating with the observed BLER improvements. In addition, we also present preliminary results for a 2X2 MIMO setup under fixed-channel CSIT and CSIR, indicating potential for extending the framework to IoT gateways with multiple antennas.