Papers & Preprints

Conference & journal papers

1. Accepted

   The Rate–Distortion–Deception Tradeoff

   S. Akkoc, S. Liyanaarachchi, S. Ulukus, and A. Yener

   62nd Allerton Conference on Communication, Control, and Computing · Sep 2026

   Introduces deception as a third axis of the rate–distortion–perception framework: a compressed reconstruction stays within a target fidelity of the source while provably resembling a sample from a different distribution. Relevant to generative-model watermarking, output auditing, and privacy-preserving compression.

   arXiv Bib

   @inproceedings{akkoc2026deception,
     title = {The Rate--Distortion--Deception Tradeoff},
     author = {Akkoc, S. and Liyanaarachchi, S. and Ulukus, S. and Yener, A.},
     booktitle = {Proc. 62nd Annual Allerton Conference on Communication, Control, and Computing},
     location = {Urbana, IL, USA},
     month = sep,
     year = {2026},
   }

2. Published

   Practical Hardware Demonstration of a Multi-Sensor Goal-Oriented Semantic Signal Processing and Communications Network

   S. Akkoc^†, A. Çınar^†, B. Ercan^†, M. Kalfa, and O. Arıkan

   Journal of the Franklin Institute, vol. 362, no. 1 · 2024

   † Equal contribution

   A hardware realization of a goal-oriented semantic communication network, showing that low-power commodity edge devices can sustain real-time semantic extraction and multi-sensor fusion.

   DOI Bib

   @article{akkoc2024semantic,
     title = {Practical Hardware Demonstration of a Multi-Sensor Goal-Oriented Semantic Signal Processing and Communications Network},
     author = {Akkoc, S. and Çınar, A. and Ercan, B. and Kalfa, M. and Arıkan, O.},
     journal = {Journal of the Franklin Institute},
     volume = {362},
     number = {1},
     year = {2024},
     doi = {10.1016/j.jfranklin.2024.107363},
   }

Preprints & under review

1. Under Review

   On the Diverse Dynamical Behaviors Arising in Deep Linear Transformers

   S. Li, T. J. Maranzatto, J. Peszek, T. Teolis, S. Akkoc, K. Riedl, S. Ulukus, and N. García Trillos

   Foundations of Computational Mathematics (FoCM) · Jul 2026

   Recasts transformer inference as an interacting-particle system and proves that, in embedding dimension two, the depth-wise dynamics reduce exactly to a generalized Kuramoto model, giving a closed-form surrogate for stacked linear self-attention and explaining token clustering, oscillation, and bifurcation.

   arXiv Bib

   @misc{li2026dynamical,
     title = {On the Diverse Dynamical Behaviors Arising in Deep Linear Transformers},
     author = {Li, S. and Maranzatto, T. J. and Peszek, J. and Teolis, T. and Akkoc, S. and Riedl, K. and Ulukus, S. and García Trillos, N.},
     month = jul,
     year = {2026},
   }

2. Submitted

   Task-aware Compression with Distribution Matching

   S. Liyanaarachchi, S. Akkoc, S. Ulukus, and A. Yener

   ICLR 2027 · Sep 2026

   Derives the optimal compression rate when the reconstruction is consumed by a mismatched downstream classifier, and recasts the design as a rate–distortion–perception problem with a KL constraint matching the classifier’s training distribution.

   arXiv Bib

   @misc{liyanaarachchi2026taskaware,
     title = {Task-aware Compression with Distribution Matching},
     author = {Liyanaarachchi, S. and Akkoc, S. and Ulukus, S. and Yener, A.},
     month = sep,
     year = {2026},
   }

3. Submitted

   The Role of Feed-Forward Layers in Transformer Dynamics

   T. J. Maranzatto, S. Akkoc, and S. Ulukus

   AISTATS 2027 · Sep 2026

   Extends the interacting-particle analysis of transformers beyond attention to the feed-forward block, isolating how pointwise nonlinearities reshape token trajectories and which layers carry the compressive work of the network.

   arXiv Bib

   @misc{maranzatto2026feedforward,
     title = {The Role of Feed-Forward Layers in Transformer Dynamics},
     author = {Maranzatto, T. J. and Akkoc, S. and Ulukus, S.},
     month = sep,
     year = {2026},
   }