# Tackling Distribution Shifts in Task-Oriented Communication with Information Bottleneck arXiv: https://arxiv.org/abs/2405.09514 ยท HTML: https://mlsanity.com/paper/2405.09514/ ## Key facts - Cited by 7 papers in the map (data through 14 Oct 2026). - Builds on 38 of its 53 references that are in the map. - First posted on arXiv 15 May 2024. ## Builds on (most cited first, 30 of 38) - [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017) - [FaceNet: A Unified Embedding for Face Recognition and Clustering](https://mlsanity.com/paper/1503.03832/) (2015) - [Learning Deep Features for Discriminative Localization](https://mlsanity.com/paper/1512.04150/) (2015) - [Invariant Risk Minimization](https://mlsanity.com/paper/1907.02893/) (2019) - [Learning deep representations by mutual information estimation and maximization](https://mlsanity.com/paper/1808.06670/) (2018) - [In Search of Lost Domain Generalization](https://mlsanity.com/paper/2007.01434/) (2020) - [Deep Variational Information Bottleneck](https://mlsanity.com/paper/1612.00410/) (2016) - [Generalized Out-of-Distribution Detection: A Survey](https://mlsanity.com/paper/2110.11334/) (2021) - [Which Training Methods for GANs do actually Converge?](https://mlsanity.com/paper/1801.04406/) (2018) - [Generalized ODIN: Detecting Out-of-distribution Image without Learning from Out-of-distribution Data](https://mlsanity.com/paper/2002.11297/) (2020) - [Deep Learning Enabled Semantic Communication Systems](https://mlsanity.com/paper/2006.10685/) (2020) - [An Introduction to Deep Learning for the Physical Layer](https://mlsanity.com/paper/1702.00832/) (2017) - [The Roadmap to 6G -- AI Empowered Wireless Networks](https://mlsanity.com/paper/1904.11686/) (2019) - [Learning Task-Oriented Communication for Edge Inference: An Information Bottleneck Approach](https://mlsanity.com/paper/2102.04170/) (2021) - [Conditional Gaussian Distribution Learning for Open Set Recognition](https://mlsanity.com/paper/2003.08823/) (2020) - [Semantic Communication Systems for Speech Transmission](https://mlsanity.com/paper/2102.12605/) (2021) - [Communication-Efficient Edge AI: Algorithms and Systems](https://mlsanity.com/paper/2002.09668/) (2020) - [Towards an Intelligent Edge: Wireless Communication Meets Machine Learning](https://mlsanity.com/paper/1809.00343/) (2018) - [BottleNet++: An End-to-End Approach for Feature Compression in Device-Edge Co-Inference Systems](https://mlsanity.com/paper/1910.14315/) (2019) - [Invariant Information Bottleneck for Domain Generalization](https://mlsanity.com/paper/2106.06333/) (2021) - [Graph Neural Networks for Scalable Radio Resource Management: Architecture Design and Theoretical Analysis](https://mlsanity.com/paper/2007.07632/) (2020) - [Task-Oriented Multi-User Semantic Communications](https://mlsanity.com/paper/2112.10255/) (2021) - [Deep Learning-Based Communication Over the Air](https://mlsanity.com/paper/1707.03384/) (2017) - [Wireless Image Retrieval at the Edge](https://mlsanity.com/paper/2007.10915/) (2020) - [Task-Oriented Communication for Multi-Device Cooperative Edge Inference](https://mlsanity.com/paper/2109.00172/) (2021) - [Wireless Semantic Communications for Video Conferencing](https://mlsanity.com/paper/2204.07790/) (2022) - [A Unified Multi-Task Semantic Communication System for Multimodal Data](https://mlsanity.com/paper/2209.07689/) (2022) - [Deep Learning-Enabled Semantic Communication Systems with Task-Unaware Transmitter and Dynamic Data](https://mlsanity.com/paper/2205.00271/) (2022) - [A Minimax Approach to Supervised Learning](https://mlsanity.com/paper/1606.02206/) (2016) - [Distributed Variational Representation Learning](https://mlsanity.com/paper/1807.04193/) (2018) ## Built on by (most cited first, 7 of 7) - [Toward Real-Time Edge AI: Model-Agnostic Task-Oriented Communication with Visual Feature Alignment](https://mlsanity.com/paper/2412.00862/) (2024) - [Multi-Modal Self-Supervised Semantic Communication](https://mlsanity.com/paper/2503.13940/) (2025) - [Distributed Information Bottleneck Theory for Multi-Modal Task-Aware Semantic Communication](https://mlsanity.com/paper/2510.04000/) (2025) - [Task-Oriented Communication for Human Action Understanding via Edge-Cloud Co-Inference](https://mlsanity.com/paper/2605.07354/) (2026) - [Orchestrating Multimodal DNN Workloads in Wireless Neural Processing](https://mlsanity.com/paper/2603.02109/) (2026) - [Domain-Adaptive Deep Joint Source-Channel Coding for Image Classification](https://mlsanity.com/paper/2607.28907/) (2026) - [Domain-Generalized Adaptive Semantic Communication for Collaborative Perception](https://mlsanity.com/paper/2608.00056/) (2026) ## Abstract (arXiv) Task-oriented communication aims to extract and transmit task-relevant information to significantly reduce the communication overhead and transmission latency. However, the unpredictable distribution shifts between training and test data, including domain shift and semantic shift, can dramatically undermine the system performance. In order to tackle these challenges, it is crucial to ensure that the encoded features can generalize to domain-shifted data and detect semanticshifted data, while remaining compact for transmission. In this paper, we propose a novel approach based on the information bottleneck (IB) principle and invariant risk minimization (IRM) framework. The proposed method aims to extract compact and informative features that possess high capability for effective domain-shift generalization and accurate semantic-shift detection without any knowledge of the test data during training. Specifically, we propose an invariant feature encoding approach based on the IB principle and IRM framework for domainshift generalization, which aims to find the causal relationship between the input data and task result by minimizing the complexity and domain dependence of the encoded feature. Furthermore, we enhance the task-oriented communication with the label-dependent feature encoding approach for semanticshift detection which achieves joint gains in IB optimization and detection performance. To avoid the intractable computation of the IB-based objective, we leverage variational approximation to derive a tractable upper bound for optimization. Extensive simulation results on image classification tasks demonstrate that the proposed scheme outperforms state-of-the-art approaches and achieves a better rate-distortion tradeoff.