Abstract—The problem of binary classification of acoustic signals of biological origin recorded in the real world is considered. Information features such as entropy and statistical complexity are chosen as the characteristic description of objects. The solution methods are based on three neural network architectures modified by the authors (on the Inception core, on the Inception core with the Residual technology, and on the Self-Attention structure with LSTM blocks). A dataset from the Kaggle competition for detecting acoustic signatures of whales was used, and the models were compared in terms of the quality of solving the problem involved on a standard set of metrics. The AUC ROC value of more than 90% was obtained, which means that the problem of detecting a useful signal is solved successfully and indicates that information features can be applied to similar tasks.

摘要：考虑了现实世界中记录的生物源声信号的二元分类问题。选择熵、统计复杂度等信息特征作为对象的特征描述。解决方法基于作者修改的三种神经网络架构（Inception 核心、采用 Residual 技术的 Inception 核心以及采用 LSTM 块的 Self-Attention 结构）。使用来自 Kaggle 竞赛的数据集来检测鲸鱼的声学特征，并根据一组标准指标解决所涉及问题的质量对模型进行比较。获得了超过90%的AUC ROC值，这意味着成功解决了检测有用信号的问题，并表明信息特征可以应用于类似的任务。