baysian-optimisation

LibMOON is a standard and flexible framework to study gradient-based multiobjective optimization.

Gait Recognition with 3D CNN. This project proposes a novel approach using 3D convolutional neural networks (3D CNN) to capture spatio-temporal features of gait sequences for robust recognition in an un-intrusive manner.

Time-series forecasting in Rust based on prophet

We're proud to announce that our team secured the First Prize in the BCG GAMMA Challenge for Data Science and Consulting. This GitHub repository hosts all the code, models, and analytical techniques we employed during the competition.

pipeDejavu: Hardware-aware Latency Predictable, Differentiable Search for Faster Config and Convergence of Distributed ML Pipeline Parallelism

Selected Paper from the AI-CyberSec 2021 Workshop in the 41st SGAI International Conference on Artificial Intelligence (MDPI Journal Electronics)

Molecular active learning with JAX

Aplicação Python+Streamlit para detectar automaticamente colônias bacterianas em imagens de placas de Petri. Usa visão computacional com transformada de Hough, otimização bayesiana para calibrar parâmetros e permite validação manual dos resultados com exportação de imagens. Ideal para uso laboratorial e educacional.

Dynamic control of a compliant base robotic arm to reach out of reach targets using Bayesian Optimization for chirp tuning and Deep Reinforcement Learning with energy reward shaping and curriculum learning in PyTorch.

A comparative study of Custom CNNs vs. Finetuning for garbage classification. Includes rigorous explainability analysis (Feature Maps), hyperparameter sweeping, and quantization benchmarks (FP16/INT8).

Human-in-the-Loop Bayesian optimization system that learns from ordinal human feedback (A/B/C ranks) to propose optimal experiment conditions with minimal trials.

This project is meant for my defence of the 2nd Internship for my Bachelor of Computer Science Year 4 at Cambodia Acedemy of Digital Technology.

MBTI mental health analysis with ML and Bayesian tuning

This project implements a high-performance pipeline for aerodynamic shape optimization. It uses Bayesian Optimization to discover ideal NACA 4-digit airfoils across a flight envelope and trains a Random Forest Surrogate Model to provide instantaneous aerodynamic predictions.

Automated Mixed-Precision Quantization Search for Deep Neural Networks

Hyperparameter tuning for CNN models on Fashion MNIST using KerasTuner. Includes random search and Bayesian optimization strategies to improve performance and training efficiency.

Comparing SMBO, Successive Halving and Random Search for KNN hyperparameter tuning against a Random Forest surrogate trained on learning-curve data

GUI for hyperparameter optimizer and plotting experiment results

This repository follows the work of Bidirectional Information Flow