TCEG

Differential Diagnosis for AI Failures

Distinguishing hallucinations from infrastructure stress. When AI fails, knowing why determines what to do. High CEG means the model is lying. Low CEG means the system is struggling.

UK Patents: GB2600522.3 • GB2600558.7 • GB2600711.2 • GB2600835.9 • GB2601070.2 • GB2601234.1

The Core Discovery

Hallucinations and infrastructure stress produce opposite signatures. TCEG's patented analysis distinguishes between them — enabling the right response to the right problem.

Patented Gradient Analysis View methodology in patent portfolio →

✓ Normal Operation

System healthy

⚠ Hallucination

Safety intervention needed

⚙ Infrastructure Stress

IT intervention needed

The key insight: Different failures require different responses. TCEG's patented methodology distinguishes between them automatically — it doesn't just detect problems, it diagnoses them.

Why this matters: Hallucination requires output suppression or human review. Infrastructure stress requires load balancing or scaling. Wrong diagnosis = wrong response = continued failure.

Validated Performance

Independently validated across multiple domains with consistent results.

0.87
AUC on TruthfulQA
LLM Hallucination Detection
99.4%
AUC Threat Detection
Cybersecurity SIEM
13×
Better Zero-Day Detection
vs Baseline Classifiers
10×
Faster Than Multi-Pass
Single-Pass Detection

Technical Comparison

Method Detection Speed Compute Explainability
TCEG 0.87 AUC Real-time ~10% High
Semantic Entropy ~0.80 AUC 5-10× slower High Medium
Static Thresholds Variable Fast Low High
ML Classifiers Training-dependent Medium Medium Low

TCEG achieves competitive accuracy with dramatically lower computational overhead through single-pass temporal gradient analysis.

Industry Applications

One framework, multiple domains: TCEG for temporal streams, P-CEG for static images.

🤖

AI Safety & LLMs

Real-time hallucination detection for ChatGPT, Claude, GPT-4, and enterprise LLMs. Single-pass detection at 10× lower compute cost than multi-sample methods.

💊

Pharma & Biotech

Predict drug sensitivity in hours vs 24-72hr conventional endpoints. Temporal entropy gradients detect cellular response trajectories in real-time screening.

🏥

Medical Monitoring

Early detection of sepsis, heart failure decompensation, and transplant rejection. Trajectory analysis catches critical transitions before threshold crossings.

🔒

Cybersecurity & SIEM

13× better zero-day detection. Filter false positives by entropy trajectory: true threats drift upward, noise spikes and falls. Reduce SOC alert fatigue.

📊

Financial Markets

Flash crash prediction and market manipulation detection. Differential diagnosis distinguishes algorithmic spoofing from genuine market stress.

🎭

Deepfake Detection

P-CEG authenticates images via perturbation stability. Three-class discrimination: authentic photographs, AI-enhanced, or fully AI-generated deepfakes.

Patent Portfolio

Complete IP protection: TCEG for temporal streams, P-CEG for static images. Filed January 2026.

GB2600522.3

09 Jan 2026

Core TCEG Framework

Foundational TCEG methodology for hallucination detection via temporal confidence-entropy gradients. Real-time streaming AI analysis.

GB2600558.7

12 Jan 2026

Medical Diagnostics

AI-assisted diagnostic systems with uncertainty quantification. Identifies "confident guessing" requiring human review.

GB2600711.2

13 Jan 2026

Cybersecurity & Finance

Threat detection in SIEM systems and black swan events in financial markets. Entropy trajectory analysis.

GB2600835.9

14 Jan 2026

P-CEG: Deepfake Detection

Perturbation stability for static image authentication. Three-class: authentic, AI-enhanced, or AI-generated deepfakes.

GB2601070.2

19 Jan 2026

Cellular Drug Response

Temporal entropy gradients predict drug sensitivity in vitro. Classification in hours vs 24-72hr conventional endpoints.

GB2601234.1

24 Jan 2026

Incremental Computation

O(1) bounded-memory temporal gradient computation. Circular buffer architecture for real-time streaming anomaly detection across all modalities.

Complete IP Coverage

Six UK patents protect complementary frameworks. TCEG (temporal) for streaming AI, medical monitoring, cybersecurity, and cellular drug response. P-CEG (perturbation) for static deepfakes. Incremental Computation for O(1) bounded-memory processing. Complete coverage from LLM safety to pharmaceutical screening.

Evolution Timeline

09 Jan
Core TCEG
AI Hallucination
12 Jan
Medical
Patient Monitoring
13 Jan
Cyber/Finance
SIEM & Markets
14 Jan
P-CEG
Deepfake Detection
19 Jan
Cellular
Drug Response
24 Jan
Incremental
O(1) Computation

From AI safety to pharmaceutical screening in 15 days: same core framework, domain-specific calibration.

Interactive Demo

See the differential diagnosis in action: watch CEG distinguish hallucination from infrastructure stress in real-time.

Demo Launching Soon

About

KW

Kevin Wharram

Independent Researcher

TCEG emerged from a cross-disciplinary observation: Trajectory matters more than absolute value. After 25 years in cybersecurity and 18 years with SIEM systems, I noticed that benign threats spike and fall, while real threats drift upward.

The differential diagnosis discovery came from asking the wrong question in the right way: "Can CEG detect infrastructure stress?" The answer was no, but the reason why revealed something more valuable. Infrastructure stress and hallucination produce opposite signatures, making CEG a diagnostic tool, not just a detector.

25+
Years Cybersec
18
Years SIEM
6
Patents Filed

Get in Touch

kevin@tceg.ai