TSM: Measuring the Enticement of Honeyfiles with Natural Language Processing

Timmer, Roelien C.; Liebowitz, David; Nepal, Surya; Kanhere, Salil S.

doi:10.48550/arxiv.2203.07580

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We found that all these methods have various disadvantages -they lack complete automation, most of them are domain-specific, and would need to be built up from scratch if a new domain is to be included & they are seldom of the same quality as an original document would be. In a similar attempt, Roelien C. Timmer et al (18) present Topic Semantic Matching (TSM) as a means to find the enticement of honeyfiles. The contents of local contexts were used to calculate the inter-corpus similarity using vector embeddings.…”

Section: High-interactionmentioning

confidence: 99%

Exploring Applications of NLP to Create Deceptive Sandbox Environments with Honeyfiles

Ravi¹,

Surve²

2023

Preprint

View full text Add to dashboard Cite

Various techniques are in practice to create a sandboxed environment to study malicious actors and/or divert them from gaining access to critical assets. Honeypots and honeynets have long been used to create decoy systems that lure adversaries to attack them, thereby revealing tactics, techniques, and procedures. When mimicking enterprise systems and networks, certain artifacts will have to be strategically placed so that they look organically generated, yet do not contain confidential data. Text assets such as word documents, audit reports, and email conversations are usually targeted for their potential value. This research explores the possibilities of generating fake, yet convincingly real documents based on relevant publicly available data. The research gaps in the existing state-of-the-art and appropriate technologies to address them would be evaluated. The project would also include proposed empirical research to verify if these NLP-generated documents are better than random gibberish text.

show abstract

Section: High-interactionmentioning

confidence: 99%