Jove
Visualize
お問い合わせ
JoVE
x logofacebook logolinkedin logoyoutube logo
JoVEについて
概要リーダーシップブログJoVEヘルプセンター
著者向け
出版プロセス編集委員会範囲と方針査読よくある質問投稿
図書館員向け
推薦の声購読アクセスリソース図書館諮問委員会よくある質問
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experimentsアーカイブ
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教員リソースセンター教員サイト
利用規約
プライバシーポリシー
ポリシー

関連する概念動画

Stereotype Threat and Self-fulfilling Prophecies02:09

Stereotype Threat and Self-fulfilling Prophecies

42.4K
When we hold a stereotype about a person, we have expectations that he or she will fulfill that stereotype. A self-fulfilling prophecy is an expectation held by a person that alters his or her behavior in a way that tends to make it true. When we hold stereotypes about a person, we tend to treat the person according to our expectations. This treatment can influence the person to act according to our stereotypic expectations, thus confirming our stereotypic beliefs. Research by Rosenthal and...
42.4K
Threats to Biodiversity01:50

Threats to Biodiversity

27.1K
There have been five major extinction events throughout geological history, resulting in the elimination of biodiversity, followed by a rebound of species that adapted to the new conditions. In the current geological epoch, the Holocene, there is a sixth extinction event in progress. This mass extinction has been attributed to human activities and is thus provisionally called the Anthropocene. In 2019 the human population reached 7.7 billion people and is projected to comprise 10 billion by...
27.1K
Intelligence01:27

Intelligence

8.6K
The term "intelligence" is complex because it refers to both behavior and individuals, and its interpretation varies across cultures. European Americans tend to link intelligence with reasoning and cognitive skills, while in Kenya, it is tied to responsible participation in family and social life. In Uganda, intelligence is seen as the ability to know the right actions and carry them out effectively, while the Iatmul people of Papua New Guinea associate it with the capacity to remember...
8.6K
Protein Networks02:26

Protein Networks

4.5K
An organism can have thousands of different proteins, and these proteins must cooperate to ensure the health of an organism. Proteins bind to other proteins and form complexes to carry out their functions. Many proteins interact with multiple other proteins creating a complex network of protein interactions.
These interactions can be represented through maps depicting protein-protein interaction networks, represented as nodes and edges. Nodes are circles that are representative of a protein,...
4.5K
Network Covalent Solids02:18

Network Covalent Solids

16.2K
Network covalent solids contain a three-dimensional network of covalently bonded atoms as found in the crystal structures of nonmetals like diamond, graphite, silicon, and some covalent compounds, such as silicon dioxide (sand) and silicon carbide (carborundum, the abrasive on sandpaper). Many minerals have networks of covalent bonds.
To break or to melt a covalent network solid, covalent bonds must be broken. Because covalent bonds are relatively strong, covalent network solids are typically...
16.2K
Measures of Intelligence01:29

Measures of Intelligence

8.4K
Psychologists measure intelligence by using standardized tests that produce a score known as the intelligence quotient or IQ. To understand IQ tests, it's important to recognize the key principles behind their construction: validity, reliability, and standardization.
Validity refers to how well a test measures what it claims to measure. An intelligence test should accurately assess intelligence rather than another characteristic, like anxiety. Criterion validity is one way to evaluate this;...
8.4K

こちらも読む

関連記事

共著者、ジャーナル、引用グラフによってこの研究に関連する記事。

並び替え
Same author

Knockdown of Slc11a1 alleviates myocardial ischemia reperfusion-induced injury in mice.

Histology and histopathology·2026
Same author

The genetic repertoire of deep-sea microbiome: From sequence to structure and function.

Cell host & microbe·2026
Same author

Invasion/metastasis-related differentially methylated genes predict prognosis in diffuse gliomas.

Journal of neuro-oncology·2026
Same author

Investigating Bidirectional Causal Relationships Between Imaging-Derived Brain Phenotypes and Sedative-Hypnotic Use Disorder: A Mendelian Randomization Study.

Addiction biology·2026
Same author

Protist-dominated hard substrate faunas thrive at the deepest ocean depths.

Science (New York, N.Y.)·2026
Same author

Effects of Short-Term 5-MeO-DiPT Exposure on the Brain Metabolome of Zebrafish (Danio rerio).

Journal of applied toxicology : JAT·2026
Same journal

Therapeutic potential of crude protein extracts from two Egyptian freshwater snails Lanistes carinatus and Bellamya unicolor.

Scientific reports·2026
Same journal

Microbial contamination of donor corneas and post-keratoplasty endophthalmitis: a comparison between Japanese and U.S. eye banks using cold storage.

Scientific reports·2026
Same journal

Prevalence and contributing factors of virological non-suppression among adult patients on first-line antiretroviral therapy in tertiary hospitals in Ethiopia.

Scientific reports·2026
Same journal

An in vitro comparison of color stability between alkasite and different restorative materials in various staining solutions.

Scientific reports·2026
Same journal

Toward accessible mRNA LNP formulation: systematic evaluation of mixing strategies and key parameters.

Scientific reports·2026
Same journal

A network analysis of personality traits, mentalizing, and psychological health in Chinese college students.

Scientific reports·2026
関連記事をすべて見る

関連する実験動画

Updated: Jan 31, 2026

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

5.4K

RoBERTa-Largeとマルチソースネットワーク脅威インテリジェンスを用いたメタデータ駆動型悪意のあるURL検出

Lina Chen1, Liang Meng2

  • 1Guangxi Power Grid Co. Ltd., Nanning, 530022, China.

Scientific reports
|January 29, 2026
PubMed
まとめ
この要約は機械生成です。

この研究は、RoBERTa-Largeトランスフォーマーを使用した高度な悪意のあるURL検出のための新しいアプローチを導入し、精度98%を達成しました。この方法は、フィッシングやマルウェアの脅威を特定する上で、従来の機械学習および深層学習モデルを大幅に上回っています。

キーワード:
人工知能サイバーセキュリティ深層学習悪意のあるURL検出RoBERTaトランスフォーマー

さらに関連する動画

Author Spotlight: Emerging Technologies and Advanced Tools for Decoding Metabolomics Data Analysis
07:11

Author Spotlight: Emerging Technologies and Advanced Tools for Decoding Metabolomics Data Analysis

Published on: November 10, 2023

3.3K
Using the Threat Probability Task to Assess Anxiety and Fear During Uncertain and Certain Threat
11:18

Using the Threat Probability Task to Assess Anxiety and Fear During Uncertain and Certain Threat

Published on: September 12, 2014

15.8K

関連する実験動画

Last Updated: Jan 31, 2026

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

5.4K
Author Spotlight: Emerging Technologies and Advanced Tools for Decoding Metabolomics Data Analysis
07:11

Author Spotlight: Emerging Technologies and Advanced Tools for Decoding Metabolomics Data Analysis

Published on: November 10, 2023

3.3K
Using the Threat Probability Task to Assess Anxiety and Fear During Uncertain and Certain Threat
11:18

Using the Threat Probability Task to Assess Anxiety and Fear During Uncertain and Certain Threat

Published on: September 12, 2014

15.8K

科学分野:

  • サイバーセキュリティ
  • 人工知能
  • 自然言語処理

背景:

  • 悪意のあるURLは、フィッシングやマルウェア配布などのサイバー攻撃の主要なベクトルです。
  • 既存の機械学習（ML）および深層学習（DL）モデルは、新しい敵対的改変に苦労しています。
  • 逐次モデルは文字レベルのパターンを捉えますが、洗練された脅威に対する堅牢性が欠けています。

研究 の 目的:

  • 堅牢で正確な悪意のあるURL検出方法を開発すること。
  • 高度なセキュリティのために最先端の大規模言語モデルを活用すること。
  • 悪意のあるURL検出モデルの解釈可能性を向上させること。

主な方法:

  • デュアルメカニズム大規模言語モデルであるRoBERTa-Largeトランスフォーマーの適用。
  • 注意層を介したコンテキスト化されたサブワード埋め込みとメタデータ信号の統合。
  • 良性、改ざん、フィッシング、マルウェアURLのバランスの取れたデータセットでのモデルのファインチューニング。

主要な成果:

  • 悪意のあるURL検出において98%の全体的な精度を達成しました。
  • 既存のMLおよびDLモデルを大幅に上回りました。
  • SHAPおよびLIME分析により、主要な機能（URLの長さ、スラッシュの深さ、エントロピー）が確認され、注意ヘッドを介して微妙な語彙上の異常が特定されました。

結論:

  • メタデータ注意とマスク言語モデルの統合は、悪意のあるURL検出のための最先端のパフォーマンスを提供します。
  • 提案された方法は、実際のアプリケーションのための透明な意思決定を提供します。
  • RoBERTa-Largeトランスフォーマーは、サイバーセキュリティ脅威検出における重要な進歩を表しています。