Arash Behboodi I am machine learning research scientist (Principal Engineer/Manager) at Qualcomm AI Research. I am doing research on machine learning design for wireless communication, learning theory and machine learning for inverse problems. My research interests include also information theory, compressed sensing and mathematical signal processing.

News and Highlights

• Workshop on Resource-Constrained Learning in Wireless Networks got accepted at Sixth Conference on Machine Learning and Systems (MLSys 2023) - more to follow.

• A new paper accepted at ICLR 2023: WiNeRT: Towards Neural Ray Tracing for Wireless Channel Modelling.

• Two papers got accepted at NeurIPS 2022: PAC-Bayesian Generalization Bounds for Equivariant Networks and On the symmetries of the synchronization problem in Cryo-EM: Multi-Frequency Vector Diffusion Maps on the Projective Plane.

• Two papers got accepted at Globecom 2022: Learning Perturbations for Soft-Output Linear MIMO Demappers and Beyond Codebook-Based Analog Beamforming at mmWave: Compressed Sensing and Machine Learning Methods.

• My chat with Sam Charrington at TWIML podcast around equivariant priors for compressed senseing.

• DeepSense competition: Multi Modal Beam Prediction Challenge 2022: Towards Generalization. Also see this paper.

• Our blog post on Bringing AI research to wireless communication and sensing.

• I am giving a talk at Qualcomm webinar on Bringing AI research to wireless communication and sensing.

• Accepted paper at ICML 2022: Equivariant priors for compressed sensing with unknown orientation.

• Two papers got accepted at ICC 2022: MIMO-GAN and Neural RF SLAM.

• CVPR 2022 workshop on Wireless AI Perception organized by my colleagues at Qualcomm.

Selected recent papers

• Tribhuvanesh Orekondy, Pratik Kumar, Shreya Kadambi, Hao Ye, Joseph Soriaga, Arash Behboodi, WiNeRT: Towards Neural Ray Tracing for Wireless Channel Modelling, ICLR 2023

• Arash Behboodi, Gabriele Cesa, Taco Cohen, PAC-Bayesian Generalization Bounds for Equivariant Networks, NeurIPS 2022

• Gabriele Cesa, Arash Behboodi, Taco Cohen, Max Welling On the symmetries of the synchronization problem in Cryo-EM: Multi-Frequency Vector Diffusion Maps on the Projective Plane, NeurIPS 2022

• Hamed Pezeshki, Fabio Valerio Massoli, Arash Behboodi, Taesang Yoo, Arumugam Kannan, Mahmoud Taherzadeh Boroujeni, Qiaoyu Li, Tao Luo, Joseph B. Soriaga, Beyond Codebook-Based Analog Beamforming at mmWave: Compressed Sensing and Machine Learning Methods, Globecom 2022

• Daniel E Worrall, Markus Peschl, Arash Behboodi, Roberto Bondesan, Learning Perturbations for Soft-Output Linear MIMO Demappers, Globecom 2022

• Anna Kuzina, Kumar Pratik, Fabio Valerio Massoli, Arash Behboodi, Equivariant priors for compressed sensing with unknown orientation, ICML 2022

• Tribhuvanesh Orekondy, Arash Behboodi, Joseph B. Soriaga, MIMO-GAN: Generative MIMO Channel Modeling, ICC 2022

• Shreya Kadambi, Arash Behboodi, Joseph B. Soriaga, Max Welling, Roohollah Amiri, Srinivas Yerramalli, Taesang Yoo, Neural RF SLAM for unsupervised positioning and mapping with channel state information, ICC 2022

• Niklas Koep, Arash Behboodi, Rudolf Mathar, The Restricted Isometry Property of Block Diagonal Matrices for Group-Sparse Signal Recovery, Applied and Computational Harmonic Analysis, Vol. 60, September 2022, (Arxiv)

• Ekkehard Schnoor, Arash Behboodi, Holger Rauhut, Generalization Error Bounds for Iterative Recovery Algorithms Unfolded as Neural Networks , Dec. 2021 (Arxiv)