*An interesting fact: Bobojon's research has received 39,603 reads

on ResearchGate as of December 2024, demonstrating

significant interest from researchers in his work.

Industry Collaborated Projects

PhD: Variational Autoencoder and Diffusion Models for Plasmonic Design (Accelerated Design)

PhD, Chapter 9: Developed an AI-powered pipeline for plasmonic NFT design, generating 2000+/h optimized structures - automating a process that traditionally requires years of expertise. Achieved 20% higher power efficiency vs expert-designed structures, with results validated through advanced electromagnetic simulations.

PhD: Machine Learning Applications in Magnetic Domain Pattern Images

To extract parameters from and/or generate new images of magnetization patterns using micromagnetic simulation data, the images of simulated magnetization patterns will be input into Convolutional Neural Networks (CNNs) and Generative Adversarial Networks (GANs) for training.

Benefit: This is an opportunity to create domains with out-of-range parameters (which you can’t execute in micromagnetic simulation).

BS: Machine Learning Applications in Wireless Communication Networks

Goal: The project aims to detect user anomalies and inference patterns in wireless networks, with the goal of improving communication systems' data rates.

Senior design project: Machine Learning Applications in Wireless Communication Networks: Interference Detection and Authentication Aspects.

BS: Electric Field Effects on the Chocolate Flow

The electric field was applied along the direction of the flow of chocolate, and the effects of the electric field on temperature, gradient, and pressure were investigated. PID was used to control and set temperature. Solid-state relays heated the system, and nitrogen gas pressed it. Cu plates created an electric field. The graph of chocolate weight versus time was graphed in real time.

In conclusion, companies can use Electric Field instead of fat to make the chocolate production ready.

Other Ph.D. Research

Tunable Spike-Timing-Dependent Plasticity (STDP) in Magnetic Skyrmion Manipulation Chambers

STDP adjusts synaptic weights based on the timing of pre- and post-synaptic spikes. The proposed three-chamber design encodes synaptic weight in the number of skyrmions in the center chamber, with left and right chambers for pre-synaptic and post-synaptic spikes. The device exhibits adaptability and learning capabilities by manipulating chamber parameters, mimicking Hebbian and dendritic location-based plasticity.

Unconventional Computing using Magnetic Skyrmions

Skyrmion stability and movement under different temperatures, current densities, and physical constraints.

Novel Materials:

Magnetic Skyrmions

Sample

Sputter Deposition

CoFeB multilayer: hosts skyrmions

Novel Materials: Dynamics of Hopfions

Study of resonant spin dynamics of topological spin textures.

Using micromagnetic simulations, e.g., MuMax3 and OOMMF to find resonant spin dynamics of a three-dimensional topological spin texture hopfion in different chiral magnets and identifying the ground state spin configuration of hopfions, effects of anisotropies, geometric confinements, and demagnetizing fields. Calculate the resonance frequencies and spin-wave modes of spin precession dynamics under multiple magnetic fields.

Novel Materials: Antiferromagnetic Material Switching

NiO Antiferromagnetic Switching:

Heterostructure growth using a sputtering system
Hall bar measurement device fabrication using photolithography
A temperature effect on the antiferromagnetic domain switching
Laser-induced dynamics
THz resonance dynamics

6-gun AJA for magnetic material deposition

Photolithography process overview

Photolithography of 8 contact Hall bar measurement device used for experimental studies.

Testing setup

Labview script for the experimental work on the switching of Antiferromagnet material.

Switching result:

Miniature Magnetometry

AutoCAD design and 3D printing.

Analytical calculation.

Applied current to generate magnetic field and measure generated field.

Arduino and amplifiers.

Other Projects

Brain-Computer Interfaces: Decoding Motor imagery-based EEG Signals using Machine Learning

Brain-computer interfaces represent a life-changing possibility for disabled patients to regain control of their lives, for example through highly functional prosthetics or movement devices. However, everyday movements, like walking, and the increased brain activity it causes pose significant difficulty for BCI decoders.

This paper proposes a method to clean EEG data and create a decoder resistant to walking-induced artifacts to classify left and right-hand motor imagery. We first perform Artifact Subspace Reconstruction to remove major artifacts. Bandpass in the alpha and mu bands to find physiologically relevant frequencies, Common Average Reference Filter to improve spatial resolution, extract trials and samples, find the mean percent difference in power spectrum density between resting and motor imagery to identify event-related desynchronization best, perform fisher score feature selection to reduce dimensionality, and finally train and cross-validate our model offline and test online with evidence accumulation.

The paper is attached under publications.

Some of the code for this work is available at https://github.com/zkhodzhaev/ML_brain_computer_interface

EEG data before and after ASR processing. The upper subplot shows the raw signal, and the lower illustrates the filtered signal without the spike.

The topological plot of the grand averaged mu power from the last 0.5s of the task trials, with and without spatial filtering.

Feature Extraction/ Selection

Fischer scores measuring statistical significance for the first experimental run 1 to 4 in session one for subject 41. The y-axis shows Fischer score values ranging from 8 to 30 Hz with ∆2. The x-axis shows the number of channels from one to fifty-five. The color represents the discriminant power of the Fischer score, with higher scores indicating features that better differentiate the conditions/classes:

Feature extraction for testing the ML model

Rethinking Fair and Stable Representation Learning

Machine Learning, particularly in the context of Graph Neural Networks (GNNs), is being deployed across various applications, including safety-critical domains.

While the accuracy of these systems remains a critical consideration, it is equally important to examine them through the lenses of fairness and stability. Prior research has introduced a unified framework to learn stable and fair node representations.

In this work, we revisit and refine this framework, presenting a simplified version demonstrating superior performance in accuracy and fairness/stability metrics.

Furthermore, we introduce a set of straightforward baseline methods that prove effective and competitively perform, often surpassing the performance of previous state-of-the-art (SOTA) approaches.

The paper is attached under publications.

The code for our work is available at https://tinyurl.com/AMLFinalCode

and specifically to my contribution (MLP): https://github.com/zkhodzhaev/representation_learning

Advanced Reinforcement Learning Agent for Pong Using TensorFlow and Keras

Keras implementation of a blog post (Deep Reinforcement Learning: Pong from Pixels) that initially used Python's numpy library for neural network operations.

Training process.

Implemented the project by integrating TensorFlow and Keras, showcasing the ability to adapt and modernize legacy code with current deep learning frameworks.
Constructed a neural network model using Keras with a hidden layer of 200 neurons, demonstrating skills in neural network architecture design.
Employed Policy Gradient methods for intelligent decision-making, reflecting an understanding of advanced reinforcement learning techniques.
Implemented custom preprocessing and a tailored loss function to optimize the agent's performance, displaying proficiency in data manipulation and algorithm customization.
Demonstrated the agent's gameplay in real-time through OpenAI Gym's environment, showcasing the practical application of machine learning in a dynamic setting.
Successfully trained, saved, and reloaded the model for demonstration purposes, illustrating the project's adaptability and reproducibility.
GitHub code can be found here: https://github.com/zkhodzhaev/reinforcement_learning

During testing, the trained model was executed using pure Python in Jupyter Notebook.