CDD Research & Development
Research Topics, Methodologies & Facilities
● Countering cyberattacks against AI/ML (e.g. Model Extraction, Bias Introduction, Reflexive Adaptation & Data Repudiation [RADaR])
● Advanced machine augmented learning architectures (e.g. SqP/SyP) that are resistent to cyberattack
● Neuromorphic computation applied to the cyber security arena
● Applied Machine Learning - Hybrid cybersecurity for blended attacks
● Augmented Identification and Authentication - Similarity heuristics and hylomorphic analytics
● Alternative Quantum data structures supportive of Quantum Machine Learning (QML)
● Enhanced, blended cybersecurity modeling using formal frameworks
(E.g. NIST CSF - ISO27K - NCSC-CAF)
● Commercial Quantum Readiness Programme
● Methodology Interoperation (e.g.CDD[CMM]/NIST-CSF/MITREAtt&ck)
● Strict and full PRINCE-II Development Disciplines (Internal to CDD)
The CDD R&D arena includes:
Development
- General Development - XTools, Swift, C & C++, Python, Jupyter
- 14000+ cores using 48GB RAM (total) GPU architecture
- 40GBPS dedicated closed N/W
- High Performance Parallelism and Multi-processing Multi-threaded Python / OpenCL methods etc. for GPU-architectures
- High Speed Demand (Assembler) - NASM / YASM / CUDA-ROCm
- Custom LLVM to implement bespoke Dynamic Compilation Infrastructure designed to fully integrate Intel AVX-512 specific instruction set capabilities.
Research
- Augmented Human Congition ('Generative AI') - KERAS, TensorFlow, PYTorch(Tensors/Transformers/LLM/Datamining), Scikit-Learn, MATLAB, Numpy, (+ Bitarray(large) GIMP_ML Open-CV etc.)
- Digital Twin Technology for the Security Arena) - to mitigate unpredictable, undesirable emergent behavior in complex cyber security environs.
- Quantum Computation - CIRQ, IBM Qiskit, TensorFlow-Quantum