Four journal papers published in four months

Academic publishing is like training a deep learning model with a dataset full of noisy labels on a computer out of the 90s: the peer review process takes forever, and you’re not even sure it’ll converge. When you finally get reviewer feedback, it's like realising that your model has been optimising the wrong objective function all along. Further iterations often lead not so much to the improvement in the quality of your manuscript, but rather cause you to overfit to the reviewer's biases.

Yet peer review remains the cornerstone of the academic process, thus we throw ourselves onto this great barrier reef again and again and again. And eventually, we persevere.

Year 2024 has been exceptionally fruitful for CIRG: journal papers first submitted for review in 2022 and 2023 finally saw the light of day, and not one or two of them, but four in a row, in a span of just four months, with the first one being accepted in May 2024, and the last one in August 2024. The proud first authors are Reinhard Booysen (PhD student), Savvas Panagiotou (MSc Alumni), Matthys Ellis (MSc Alumni) and Arne Schreuder (PhD student). Head to the Publications page for full reference information.

CIRG proudly presents:

• Reinhard Booysen, Anna S. Bosman, Multi-objective Evolutionary Neural Architecture Search for Recurrent Neural Networks. Neural Process Letters 56, 200 (2024).
• Mathys Ellis, Anna S. Bosman, Andries P. Engelbrecht, Regularised feed forward neural networks for streamed data classification problems. Engineering Applications of Artificial Intelligence, Volume 133, Part F, 2024, 108555.
• Savvas Panagiotou, Anna S. Bosman, Denoising diffusion post-processing for low-light image enhancement. Pattern Recognition, 2024, 110799, ISSN 0031-3203.
• A.N. Schreuder, A.S. Bosman, A.P. Engelbrecht, C.W. Cleghorn, Training feedforward neural networks with Bayesian hyper-heuristics, Information Sciences, Volume 686, 2025, 121363.