Cyber risk quantification as a really neat use case of building out a custom annotation toolchain and methodology to bootstrap a dataset you’ve aggregated -- how to dedup, clean up, and annotate with new features. https://t.co/VHVhn8sbku
This post shares a method and toolchain for human annotation on top of aggregation to bootstrap data for cyber risk quantification. Curious about scaling this to an open community effort; anyone seen something like that work? cc @CISecurity@FAIRInstitute https://t.co/VHVhn8sbku
Second in the series as we open our cyber risk dataset; how we bootstrapped more robust breach records with an annotation toolchain for @CISecurity controls and loss info for modeling severity to drive @FAIRInstitute Cyber Risk Quantification models https://t.co/VHVhn8sbku
Efficiently Bootstrapping Datasets with Machine Aggregation and Human Annotation: a Cyber Risk Case Study that annotates @CISecurity controls to power @FAIRInstitute cyber risk quantification models https://t.co/gb4mb054jv
Analysis suggests what CISOs have been saying all along; insurers may not be looking at the security controls when assessing security risk
https://t.co/gVTHSiF7UV
Which CIS Controls Matter Most: An Empirical Analysis of Cyber Insurance Risk Assessment https://t.co/qzSbfZhjeX << Analysis suggests what CISOs have been saying all along; insurers may not be looking at the best data to access security risk.
Deep learning on big dense data has been a bigger focus in the ML community recently, but problems like Cyber Insurance with small sparse data require Bayesian methods. Details here https://t.co/X4Wj4d0AZK
A high level overview of our underwriting models, to be followed by more technical posts: Bayesian Cyber Risk Quantification With Industry-Specific Models https://t.co/X4Wj4cJ08c
Deep learning on big dense data has been a bigger focus in the ML community recently, but problems like cyber insurance with small sparse data require Bayesian methods. Walk through an example: https://t.co/Lz6ULZA7Kx
Cybersecurity risk modelers and cyber insurers might want to sign up for the private beta of the world’s most complete breach dataset for cybersecurity risk models https://t.co/dzrapZ2KRt
hello friends @CISecurity we have just released a new cybersecurity risk dataset which includes human annotated CIS controls affected at time of breach. Maybe this could be helpful for an upcoming CIS controls revision? https://t.co/dzrapZ2KRt
Security, Risk, and Insurance Peoples: Cost-per-record is a pretty limited model for losses stemming from security incidents. We’re looking for feedback on our industry-specific incident loss modeling. Please share comments on our white paper https://t.co/YY027gFyGu
Cost-per-record isn’t a legit model to unlock real capacity in the cyber insurance market. We’ve researched around cyber loss costs and summarised in a white paper we want to continuously improve. Please send us your comments and feedback. https://t.co/GS7NljND8P
The world's largest insurance and reinsurance balance sheets might like to take on more cyber risk, but they can’t without industry-specific modeling on both the security side and on the financial side. https://t.co/ak51UT1JMR
the $690M in losses announced so far may grow over time. a new bill proposed by warren and others "would have resulted in a $1.5 billion penalty after the Equifax breach revealed in 2017, the bill’s sponsors estimate." https://t.co/NZ1mjZFxix
#Cybersecurity is not a problem to be solved but a risk to be managed. Learn more from this @towerstreetHQ whitepaper https://t.co/541WgBQSDN and watch our on-demand webinar https://t.co/WMdxxo4AkT