gayanku / nbc-softmax Goto Github PK

About Me · Email · LinkedIn · Facebook

AI, ML Scientist / Researcher (PhD) with CSIRO and Microsoft awards, Data Engineer (MSc) with 20+ years of software engineering (BSc), including the London Stock Exchange and Central Bank (IT, finance, regulatory, AML). A chartered management accountant (UK/USA) and a senior tech manager (10+ years). Compulsive learner, Geek, mentor, consultant, founder, startup enthusiast, food critic and travel addict.

Practicale software design 📲, solid architecture 👷‍♀️, best practices 🧰, and documentation 📖.

• 🔭 Currently keen in LLM (Large Language Models) and optimization of LLM for private use (private GPT).
• 🌱 Passionate about building effcient AI and Machine Learning (ML) models in Graph, Vision and Language domains
• 💬 Ask me about GPT, LLM, Python, Dart, Flutter, C++.
• 📫 How to reach me: [email protected]
• 😄 Pronouns: He/Him/His
• ⚡ Fun fact: A huge sci-fi fan, since I read my first book at the age of 12

FDGATII
Fast Dynamic Graph Attention with Initial Residual and Identity Mapping, adding 3 main enhancements on Graph Attention (GAT). AJCAI22/CSIRO Best paper

SGGC
Self-Supervised Contrastive Graph Clustering & Influence Augmented Contrastive (IAC) loss, a more effective, novel, Influence Augmented Contrastive (IAC) loss to fuse richer structural information, and half the original model parameters. SCGC(*) is faster with simple linear units, completely eliminate convolutions and attention of traditional GNNs, yet efficiently incorporates structure. It is impervious to layer depth and robust to over-smoothing, incorrect edges and heterophily. It is scalable by batching, a limitation in many prior GNN models, and trivially parallelizable.

PAMC
Proxy approximated meta-node Contrastive (PAMC) loss, a meta-node based approximation technique that is (a) simple, (b) canproxy all negative combinations (c) in quadratic cluster size time complexity, (d) at graph level, not node level, and (e) exploit graph sparsity. By replacing node-pairs with additive cluster-pairs, we compute the negatives in cluster-time at graph level. The resulting Proxy approximated meta-node Contrastive (PamC) loss, based on simple optimized GPU operations, captures the full set of negatives, yet is efficient with a linear time complexity. By avoiding sampling, we effectively eliminate sample bias.

NBC-Softmax
NBC-Softmax : Darkweb Author fingerprinting and migration tracking, a contrastive loss based clustering technique for softmax loss, which is more intuitive and able to achieve superior performance. Our technique meets the criterion for larger number of samples, thus achieving block contrastiveness, which is proven to outperform pair-wise losses. It uses mini-batch sampling effectively and is scalable.

Ugly Ducklings or Swans
A Tiered Quadruplet Network with Patient-Specific Mining for Improved Skin Lesion Classification, - a deep metric learning network to learn lesion features at two tiers - patient-level and lesion-level. We introduce a patient-specific quadruplet mining approach together with a tiered quadruplet network, to drive the network to learn more contextual information both globally and locally between the two tiers.

FlowTransformer: A Transformer Framework for Flow-based Network Intrusion Detection Systems
XG-BoT: An Explainable Deep Graph Neural Network for Botnet Detection and Forensics
Inspection-L: Self-Supervised GNN Node Embeddings for Money Laundering Detection in Bitcoin
DOC-NAD: A Hybrid Deep One-class Classifier for Network Anomaly Detection