©2023 Autonym Pte Ltd - Singapore
A method inspired by neuroscience
Xenomorphic cryptography, or unusual form of cryptography, is a form of mulvariate
cryptography using cognive compung that combines incoming signal processing with a racing
arcial neural network for contextual passkey and private key generaon, with dynamic
evoluon by elongaon and retracon over me and usage. The user’s cognive variability itself
becomes the passkey and decrypon key’s originator, so such generated passkey and key can
take unpredictable forms (i.e., xenomorphic). With this cryptographic method, the user will self-
create a chain of nodes based on self-captured informaon – from reacon me to facial
expression and age from birth – during a series of simple tasks, such as taking a sele, making a
visual choice, or making an intuive choice. Each task will create one node, one to one. All nodes
will be racing against me and the cipher will exploit the many-valued output of the whole
structure, such as the angle value and the angular velocity of each component to generate a one-
me cipher for single use. Each new login will add a node to the chain of nodes unl preset
boundaries are met, aer which the chain will start decaying in order to limit predictability.
Generang a one-me decrypon key at the user’s me of use and adopng this self-generated
key to create a one-me-pair of keys for end-to-end encrypon constutes the core of the new
method, since decrypng the ciphertext of the senders will require using the self-generated
(nave) key to regenerate the plaintext at the receiver’s end.
Compung the Human Unclonable Funcon
Inspired by chip’s Physically Unclonable Funcon used for Internet of Things’ cryptography,
xenomorphic cryptography makes use of the Human Unclonable Funcon to generate its user-
centric cipher. The chain of concatenated nodes at the core of the cipher generaon is based on
tri-dimensional nodes, i.e., with two spaal dimensions and one temporal. Accordingly, nodes are
drawn from three inputs related to the performance of the user when exposed to a simple
cognive task, and repeatedly, including reacon me, captured by a mer between the task’s
request and the response (motor core of the node), facial expression, when recognized by a
convoluonal neural network as the user completes the task (sensory core of the node) and
velocity, determined from the age of the user since birth. When mulple nodes are programmed,
they form a chain, and the nodes in the chain will start racing in a network which has dynamics
that serve as an encrypon device or cipher.
Grand complicaons
The xenobase for encoding contains tasks such as sele task, single answer queson, single visual
choice, mulvariate choice replacing the usual set of number, leer, and symbol of tradional
bases. Forming in such an unusual matrix, the nodes allow dynamic encoding of unusual forms
that can sasfy any preset security requirements such as logic, complexity, mutability, and
meliness. A decay mechanism, inspired by biology, is another core component introduced to
prevent reverse engineering while liming compung eorts necessary to maintain encrypon
complexity. Many types of decay can be implemented such as linear decay, random decay, preset
decay, and binomial distributed decay. A new form of arithmec called Posit was implemented to
make computaonal operaons faster, improving accuracy and lowering compung power needs
in parcular for portable soluons. Finally, an imaginary transform was created to benet from
posits’ strengths.