Excited to share that our paper showing printed organic optoelectronics can be used for machine learning and in-sensor classification is featured on the cover of Issue 3 of Matter & Light, Cell Press — a cover illustration of an array of organic phototransistors simultaneously sensing and classifying a handwritten digit. Congratulations to Mahsa Sadeghi for leading this work, and it’s been a pleasure continuing to collaborate with my PhD advisor, Ana Claudia Arias, eight years after graduation. 🎉📖

Paper title: Machine learning classifier based on printed organic electronics

Publication:

  1. Machine learning classifier based on printed organic electronics Mahsa Sadeghi, Rozhan Rabbani, Ali Moin, Burak Eminoglu, Seiya Ono, Anju Toor, Yasser Khan, Jan Rabaey, and Ana C. Arias Matter & Light, 2026 Cover Article

    Confronting the challenge of dense data processing in AI hardware, this research introduces an in-sensor computing approach that reduces bandwidth by transmitting only salient features. Using flexible printed electronics, it enables sensory platforms beyond the limits of conventional rigid electronics. The study presents two in-sensor classifier models: one based on organic thin-film transistors (OTFTs) that classify sensor input data and one based on organic phototransistors (OPTs) that integrate sensing and classification within a single device. On the MNIST dataset, the OTFT-based circuit achieved 91.3% average accuracy for binary classification of digits 0 and 7, while the OPT-based system achieved 94.1%. Extending the framework to 10-class classification without hardware modification yielded average accuracies of 87.1% (OTFT) and 91.5% (OPT). Evaluation on Fashion-MNIST demonstrated generalizability, achieving 86.5% (OTFT) and 81.1% (OPT), highlighting the feasibility of AI sensory hardware based on printed organic electronics.

    @article{sadeghi2026machine, title = {Machine learning classifier based on printed organic electronics}, author = {Sadeghi, Mahsa and Rabbani, Rozhan and Moin, Ali and Eminoglu, Burak and Ono, Seiya and Toor, Anju and Khan, Yasser and Rabaey, Jan and Arias, Ana C.}, journal = {Matter \& Light}, year = {2026}, publisher = {Elsevier}, doi = {10.1016/j.matlit.2026.100064}, thumbnail = {sadeghi2026machine.jpg}, url = {https://doi.org/10.1016/j.matlit.2026.100064}, pdf = {sadeghi2026machine.pdf}, note = {Cover Article} }

    PDF