Auditing

Machine Learning

Machine learning has revolutionised the way we process data. The combination of the logical, human and mathematical modelling presents opportunities on the battle field, in healthcare, automotives and space technology.

Hunnic Cyber works with you to challenge your assumptions, tease apart your conclusions and help you to address the mathematical limitations of your algorithms.

We work with you to establish your goals, intended approach and mathematics. We then start our audit, playing the devil's advocate at each level, dimension and structure of complexity.

Once the audit concludes, we provide you with actionable recommendations to improve.



Our Approach

Comprehensive

Model Audit

image

Mathematical

The mathematical elements of Machine Learning seek to break down complex mathematical, logical and procedural structures in order to anticipate. On the battle field, space exploration and healthcare, minute mathematical deviations save lives.

image

Structural

Machine Learning deals with input and output. While the assumptions made and mathematics can be sound, the "plumbing " can be the source of deficiencies. Integrity, security, interface - the full stack must be considered.

image

Logical

Whether working with healthcare, or modelling complex human behaviour, the first step is to make an assumption about the cause and effect. Bias, can result in flawed data, that is either unethical, or worse puts lives at risk.

Intellectual

Partnerships

The machine learning boom will have global, regional and local implications.

Already partnered in Saudi Arabia, Hunnic Cyber seeks to grow, foster and enrich relationships with think tanks, academic insitutations and promote collarboration.

The old Hungarian culture, political order, and science, was one that promoted inclusion, ethics and a pursuit for the right answer.

We seek to know the unknown, anticipate the random, and simplify the complex.


RANDOM

UNKNOWN

COMPLEX

Promoting

Research

We recently partnered with Fiduciam Global and LFD to provide a continous audit of their Machine Learning solution already fighting coronavirus in Pakistan.

We are also releasing a research post around using a Generative Adversarial Network to audit one of the leading Gartner Endpoint, Detection & Response solutions on the market.



REQUEST RESEARCH

Public

Speaking

Hackers Academy Virtual Online Conference.

We will be speaking at HAVOC around the use of Machine Learning in Adversary Simulation engagements and how threat actors can leverage these technologies.



IN MEMORIAM - Czakó Emerencia Ágnes