As network attacks are constantly and dramatically evolving, demonstrating new patterns, intelligent Network Intrusion Detection Systems (NIDS), using deep-learning techniques, have been actively studied to tackle these problems. Recently, various autoencoders have been used for NIDS in order to accurately and promptly detect unknown types of attacks (i.e., zero-day attacks) and also alleviate the burden of the laborious labeling task. Although the autoencoders are effective in detecting unknown types of attacks, it takes tremendous time and effort to find the optimal model architecture and hyperparameter settings of the autoencoders that result in the best detection performance. This can be an obstacle that hinders practical applications of autoencoder-based NIDS. To address this challenge, we rigorously study autoencoders using the benchmark datasets, NSL-KDD, IoTID20, and N-BaIoT. We evaluate multiple combinations of different model structures and latent sizes, using a simple autoencoder model. The results indicate that the latent size of an autoencoder model can have a significant impact on the IDS performance.

Dongjun Lim, Yun-Gyung Cheong. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.

KSII Transactions on Internet and Information Systems TIIS Paper Details

Understanding how urban parks are utilized and perceived by the public is crucial for effective urban planning and management. This study introduces a novel dataset derived from Instagram, using 42,187 images tagged with #Seoul and #Park hashtags from 2017 to 2023. These images were filtered using InternLM-XComposer2, a Multimodal Large Language Model (MLLM), to confirm they depicted park scenes. GPT-4 then annotated the filtered images, resulting in 29,866 valid image annotations of physical elements, human activities, animals, and emotions. The dataset is publicly available at https://huggingface.co/datasets/RedBall/seoul-urban-park-analysis-by-llm.