Dear colleagues,

We invite you to participate in the LIBS 2026 data analysis contest, organized as a benchmarking activity associated with the LIBS 2026 conference in Guangzhou. The purpose of this activity is to provide a common dataset and evaluation protocol for comparing data-processing and classification methods in laser-induced breakdown spectroscopy.

The contest focuses on a steel aging-state classification problem. The goal is to predict the aging level of a steel sample from its LIBS spectra. The spectral data provided for this contest are raw signals directly obtained from the spectrometer.

The dataset contains 180 anonymized samples representing five aging grades:

• 120 labeled training samples;
• 60 test samples for which participants must predict the associated aging levels;
• 12282 wavelength values for each spectrum;
• 200 spectra for each sample.

The dataset was acquired under measurement conditions designed to reflect practical constraints. Spectra were collected in ambient air using a three-channel fiber spectrometer equipped with CMOS detectors, with remote measurements performed at a stand-off distance of 0.5 m. The contest therefore evaluates classification performance in the presence of measurement variability relevant to field-oriented LIBS applications.

All necessary data for the challenge are available for download here:
https://pan.jiangnan.edu.cn/link/AA3B0954F24EAA44C0A71484571BB364EF

Contest rules:

1. The classification accuracy on the test set will be used as the primary figure of merit for ranking submitted results.
2. Participants should submit a CSV file containing the predicted aging level for each test sample. The required file format is provided in sample_submission/predictions.csv.
3. Submission files should be sent by email to libs2026@hjsmeeting.cn.
4. The submission deadline is October 31, 2026.
5. Only registered LIBS 2026 conference attendees may participate in the contest.
6. Each participant or team may submit only one final CSV file before the deadline.
7. There are no restrictions on the software tools or classification methodologies used.
8. The top-performing teams will be acknowledged in a special session at LIBS 2026 in Guangzhou and will be invited to present their methods and results.
9. The top-performing teams will also be invited to contribute, together with the contest organizers, to a manuscript for a special issue of Spectrochimica Acta Part B.

For questions regarding the contest, please contact libs2026@hjsmeeting.cn.