Thesis Domains Method Team Research Contact
Independent Research Laboratory

We reframe problems
the world gave up on

Lateral Frame solves long-standing problems in mathematics, medicine, and machine intelligence by changing the frame, not the effort.

View Research Our Thesis
Fluid Dynamics Clinical AI Cybersecurity Mathematical Physics Financial Early Warning Autonomous Agents Pattern Recognition Critical Transitions Fluid Dynamics Clinical AI Cybersecurity Mathematical Physics Financial Early Warning Autonomous Agents Pattern Recognition Critical Transitions
01 / Thesis
The problem is never
the problem

Most hard problems remain unsolved not because they lack effort, but because they are framed incorrectly. The same question, seen from a different angle, often yields an answer that was always there.

Lateral Frame is an independent research laboratory that identifies problems stuck for years or decades and applies cross-domain reframing to unlock them. We draw on mathematics, physics, information theory, and clinical science, not as separate fields, but as different lenses on the same underlying structures.

Every system approaching failure exhibits the same signature. We find it.

8
Patents Filed
6
Domains Covered
40+
Year-old Problems Addressed
25
Years Experience
02 / Domains
One method, many fields
001
Mathematical Physics
Novel results in vortex sheet singularity theory, reframing classical open problems through condition number analysis and spectral methods.
002
Clinical AI
Early warning systems for patient deterioration using entropy-based monitoring, validated across multiple critical care datasets.
003
Cybersecurity
Cross-domain threat detection and vulnerability intelligence, applying critical transition theory to network security monitoring.
004
Financial Systems
Early warning frameworks for systemic risk, detecting the statistical signatures of approaching market regime changes.
005
AI Agent Monitoring
Runtime monitoring of autonomous AI systems, detecting drift and failure modes before they cascade into harmful outputs.
006
Drug Response
Cellular-level prediction of therapeutic response, identifying patient-specific treatment trajectories from molecular signals.
03 / Method
How we think
01
Identify the stuck frame
We find problems where decades of effort in the conventional frame have failed to produce a solution. The longer the stalemate, the more likely the frame itself is wrong.
02
Reframe laterally
We apply cross-domain pattern recognition to reformulate the problem. Techniques from information theory illuminate fluid dynamics. Entropy analysis reveals clinical signals. The method transfers.
03
Prove and protect
Solutions are formalised as rigorous proofs or validated through empirical testing, then protected through patent filings. Every insight becomes intellectual property.
04 / Team
The people behind
the reframe
KW
Kevin Wharram
Co-founder
Independent researcher and inventor with 25+ years in cybersecurity and 18 years specialising in SIEM systems. Kevin's cross-domain pattern recognition, connecting information theory, fluid dynamics, and clinical science, is the foundation of Lateral Frame's method. He manages a growing patent portfolio spanning eight filings across six domains.
LinkedIn
TP
Theodore Pezas
Co-founder
Surgeon and clinical researcher based in London with expertise in plastic and reconstructive surgery. Theodore brings frontline clinical insight to Lateral Frame's medical research, bridging the gap between algorithmic detection and real-world patient outcomes. His published research spans surgical outcomes, complications analysis, and evidence-based clinical practice.
LinkedIn
BH
Blanka Horvath
Co-founder
Associate Professor in Mathematical and Computational Finance at the University of Oxford, researcher at the Alan Turing Institute, and a lead of the DataSig research group. Blanka's work sits at the intersection of stochastic analysis, rough path theory, and machine learning for finance. She is the inaugural recipient of the Risk.net Quant Rising Star Award and the 2024-25 LMS Emmy Noether Fellow in Mathematics.
LinkedIn Research
05 / Research
Selected work
Paper · In Preparation
On the Condition Number of Vortex Sheets at the Moore Singularity
Proving that κ(t) diverges with exact exponent α = ½ as the Moore singularity forms, resolving a question open since 1979 by reframing it as a condition number problem.
Mathematical Physics · Fluid Dynamics
Patent Portfolio · Filed
Cross-Domain Critical Transition Detection
A unified framework for detecting approaching system failures across clinical, cybersecurity, financial, and AI domains using confidence-entropy divergence analysis.
Clinical AI · Cybersecurity · Finance
Validation · Multi-Dataset
Clinical Early Warning System
Entropy-based patient deterioration detection validated across MIMIC-IV, eICU, HiRID, and SICdb critical care databases with AUC exceeding 0.86.
Clinical AI · Critical Care
Paper · In Preparation
Convergent miRNA Downregulation in Infrastructure Organ Cancers
Evidence that cancer lethality reflects shared regulatory degradation in high-throughput organs rather than tumour-intrinsic properties, validated across SEER, MIMIC-IV, and TCGA with a three-variable model achieving R²(LOO) = 0.913 across 20 cancer types.
Cancer Biology · Regulatory Networks
"
Every system casts a shadow. The shadow is always predictable, even when the event is not.
Founding principle, Lateral Frame
06 / Contact
Have an unsolvable
problem?

We work with researchers, organisations, and public bodies facing problems that conventional approaches have failed to solve. If your challenge has been stuck for years, it may need a different frame, not more effort.

contact@lateralframe.com