Hasil Pencarian  ::  Simpan CSV :: Kembali

"Secara umum praktik praktik web phishing semakin kompleks dan sulit dideteksi, sehingga diperlukan pendekatan yang lebih maju untuk menghadapinya. Selain itu kesadaran dan kepercayaan pengguna terhadap sistem deteksi phishing perlu ditingkatkan dalam kerangka pemahaman tentang bagaimana sistem pendeteksian bekerja dan alasan di balik hasil yang diberikan. Penelitian penggunaan metode pendeteksian phishing menggunakan Machine Learning sifat pemrosesannya masih berupa Black-box yaitu proses pengambilan keputusannya tidak diketahui. Penelitian ini bertujuan membantu memberikan penjelasan pemahaman terhadap hasil pendeteksian phishing. Pendekatan XAI (eXplainable Artificial Intelligence) untuk pendeteksian phishing memberikan tambahan penjelasan fitur-fitur yang diekstraksi dari web berupa URL dan struktur HTML yang berkontribusi dalam hasil pendeteksian phishing. Selain memberikan kejelasan kontribusi juga dapat digunakan untuk menilai model menggunakan metrik kinerja, dan efek kausalitas dari fitur-fitur tersebut.Penelitian menggunakan kombinasi algoritma yang termasuk dalam model White-box seperti algoritma Explainable Boosting Machine (EBM) dan model Black-box populer yaitu Random Forest, XGBoost dan CatBoost. Algoritma tersebut akan diukur dan dibandingkan kinerjanya pada sekumpulan dataset Phishing bersumber dari beberapa situs. Selain diukur kinerjanya, model yang dihasilkan diuraikan atau dijelaskan menggunakan metode XAI yaitu SHAP (Shapley Additives exPlanation). Adapun penjelasan hasil XAI divisualisasikan menggunakan Plot Nilai Penting Fitur dan Efek Kasualitas Model. Setiap Plot yang dihasilkan dianalisis dan dirangkum untuk mendapatkan temuan atau jawaban atas pertanyaan riset yang relevan. Hasil yang didapatkan dari penelitian ini menunjukkan beberapa temuan penting bahwa ada korelasi dimensi dataset dengan metrik kinerja. Dari metrik penjelas, plot Nilai Penting Fitur memberikan beberapa gambaran kemampuan model terhadap dataset baru dan juga memberikan rekomendasi fitur-fitur yang perlu digunakan kembali dalam model. Penjelasan hasil metode XAI bersifat global artinya merupakan gambaran utuh perilaku dan kontribusi fitur terhadap model. Secara umum algoritma Black-box XGBoost mempunyai kinerja yang baik. Dalam perspektif fitur dalam dataset phishing, Fitur-fitur seperti url_length, n_slash, n_dots, SSLfinal_State, dan URL_of_anchor konsisten menjadi fitur paling signifikan di berbagai model.

---

A more sophisticated strategy is required to combat web phishing tactics since they are generally growing more intricate and challenging to identify. Furthermore, in order to better comprehend how detection systems operate and the rationale behind the findings they produce, user awareness and trust in phishing detection systems must be raised. The processing nature of research on machine learning-based phishing detection techniques is still a "black box," meaning that the approach used to make decisions is unknown. The purpose of this study is to contribute to the understanding and explanation of phishing detection outcomes. Additional explanations of online elements like URLs and HTML structures that are extracted and contribute to the phishing detection findings are provided by the XAI (eXplainable Artificial Intelligence) approach. It can be used to assess the model using performance metrics and the causal impacts of certain features, in addition to offering clarity on contributions.

In this study, the Explainable Boosting Machine (EBM) algorithm and other White-box algorithms are combined with well-known Black-box models including Random Forest, XGBoost, and CatBoost. The algorithms' performance will be evaluated and contrasted using a collection of phishing datasets from various websites. The resulting model is described or explained using the XAI method, namely SHAP (Shapley Additives exPlanation), in addition to its performance being measured. Model Causality Effects and Feature Importance Value Plots are used to illustrate the explanation of the XAI results. Every plot that is produced is examined and condensed to produce conclusions or responses to pertinent research questions.

There is a correlation between the dataset's dimensions and performance measures, according to the study's conclusions, which highlight numerous significant findings. The Feature Importance plot derived from the explanatory metrics In addition to suggesting which characteristics should be kept in the model, features give some insight into how well the model performs with new datasets. The outcomes of the XAI approach are explained globally, which gives a thorough understanding of how features behave and contribute to the model. The Black-box XGBoost algorithm works well overall. Features like url_length, n_slash, n_dots, SSLfinal_State, and URL_of_anchor are consistently the most important features in the phishing dataset when viewed from the standpoint of features in different models."

"Penerapan teknik Explainable AI (XAI) telah menjadi fokus utama penelitian dalam upaya untuk meningkatkan interpretabilitas dan kepercayaan dalam model AI, khususnya pada bidang outlier detection. Penelitian ini bertujuan untuk mengungkapkan proses pengambilan keputusan yang kompleks di balik proses outlier detection, serta untuk memberikan pemahaman yang lebih dalam tentang faktor-faktor yang mempengaruhi keputusan tersebut. Dalam penelitian ini, diselidiki berbagai teknik XAI yang dapat digunakan dalam konteks outlier detection. Penelitian ini memberikan evaluasi komprehensif tentang aplikasi XAI dalam outlier detection, dengan mengevaluasi kelebihan dan kelemahan dari setiap teknik yang digunakan. Hasil eksperimen menunjukkan bahwa penerapan XAI dalam outlier detection dapat memberikan wawasan yang berharga tentang faktor-faktor yang mempengaruhi keputusan model, dan dapat meningkatkan interpretabilitas dan kepercayaan dalam model outlier detection.

---

The application of Explainable AI (XAI) techniques has been the main focus of research to improve interpretability and trust in AI models, particularly in the field of outlier detection. This study aims to uncover the complex decision-making process behind outlier detection and provide a deeper understanding of the factors influencing these decisions. Various XAI techniques that can be used in outlier detection are investigated in this research. This study provides a comprehensive evaluation of XAI applications in outlier detection by assessing the strengths and weaknesses of each technique used. The experimental results indicate that the implementation of XAI in outlier detection can provide valuable insights into the factors influencing model decisions and can enhance the interpretability and trustworthiness of outlier detection models."

"This book mainly discusses the most important issues in artificial intelligence-aided future networks, such as applying different ML approaches to investigate solutions to intelligently monitor, control and optimize networking. The authors focus on four scenarios of successfully applying machine learning in network space. It also discusses the main challenge of network traffic intelligent awareness and introduces several machine learning-based traffic awareness algorithms, such as traffic classification, anomaly traffic identification and traffic prediction. The authors introduce some ML approaches like reinforcement learning to deal with network control problem in this book."

"Penggunaan analisis sentimen semakin umum digunakan. Dalam pengembangan analisis sentimen ini banyak tantangan yang perlu dihadapi. Karena analisis ini termasuk Natural Language Processing NLP, hal yang perlu dimengerti adalah kompleksitas bahasa. Dengan berkembangnya teknologi Artificial Neural Network, ANN semakin banyak permasalahan yang bisa diselesaikan. Ada banyak contoh struktur ANN dan untuk penelitian ini yang digunakan adalah Convolutional Neural Network CNN dan Recurrent Neural Network RNN. Kedua jenis ANN tersebut sudah menunjukkan performa yang baik untuk beberapa tugas NLP. Maka akan dilakukan analisis sentimen dengan menggunakan kedua jenis ANN tersebut dan dibandingkan kedua performa ANN tersebut. Untuk data yang akan digunakan diambil dari publikasi stanford dan untuk mengubah data tersebut bisa digunakan pada ANN digunakan word2vec. Hasil dari analisis menunjukkan bahwa RNN menunjukkan hasil yang lebih baik dari CNN. Walaupun akurasi tidak terlalu terlihat perbedaan yaitu pada RNN yang mencapai 88.35 0.07 dan CNN 87.11 0.50, tetapi waktu pelatihan RNN hanya membutuhkan waktu 8.256 detik sedangkan CNN membutuhkan waktu 544.366 detik.

---

Term of sentiment analysis become popular lately. There are many challenges developing sentiment analysis that need to be addressed. Because this kind analysis is including Natural Language Processing, the thing need to understand is the complexity of the language. With the current development of Artificial Neural Network ANN, more problems can be solved.

There are many type of ANN and for this research Convolutional Neural Network CNN and Recurrent Neural Network will be used. Both already showing great result for several NLP tasks. Data taken from stanford publication and transform it with word2vec so could be used for ANN.

The result shows that RNN is better than CNN. Even the difference of accuracy is not significant with 88.35 0.07 for RNN and 87.11 0.50 for CNN, the training time for RNN only need 8.256 secods while CNN need 544.366 seconds."

"The term "Artificial Intelligence" has been around since the 1950s, but a lot has changed since then. Today, AI is referenced in the news, books, movies, and TV shows, and the exact definition is often misinterpreted. Artificial Intelligence For Dummies provides a clear introduction to AI and how it's being used today. Inside, you'll get a clear overview of the technology, the common misconceptions surrounding it, and a fascinating look at its applications in everything from self-driving cars and drones to its contributions in the medical field."

"In the chapters in Part I of this textbook the author introduces the fundamental ideas of artificial intelligence and computational intelligence. In Part II he explains key AI methods such as search, evolutionary computing, logic-based reasoning, knowledge representation, rule-based systems, pattern recognition, neural networks, and cognitive architectures. Finally, in Part III, he expands the context to discuss theories of intelligence in philosophy and psychology, key applications of AI systems, and the likely future of artificial intelligence. A key feature of the author's approach is historical and biographical footnotes, stressing the multidisciplinary character of the field and its pioneers.The book is appropriate for advanced undergraduate and graduate courses in computer science, engineering, and other applied sciences, and the appendices offer short formal, mathematical models and notes to support the reader."