Eraraya Ricardo Muten
Hi! I’m Edo! π
I’m currently writing my master’s thesis in quantum optimization as part of my study at TUM while working part-time at Infineon Technologies. This year, I am part of TUM Venture Lab Quantum Fellowship Program, a one-year entrepreneurship education program focusing on quantum applications.
I’ve been to several cities, mostly for academic, research & software dev work. I’ve previously worked in Japan (Tokyo & Yokohama), Indonesia (Bandung & Jakarta), Germany (Munich), and the US (New York). Meeting new people from different cultural backgrounds has always been a very exciting experience :)
- Quantum Machine Learning, Quantum Optimization, Quantum Simulation
- Classical Optimization, Machine Learning, Deep Learning
- Product Development, Entrepreneurship
M.Sc. in Quantum Science & Technology, Present
Technical University of Munich
B.S. in Engineering Physics (First Class Honours), 2021
Bandung Institute of Technology
AIMS Programme Exchange Student, Applied Physics, Sept 2018 - Jan 2019
Tokyo University of Agriculture and Technology
Experience
Research & Working
Studying algorithms to enhance QAOA performance, taking advantage of qubit measurement correlation from an optimized QAOA circuit, e.g., the quantum relax-and-round algorithm and variations of recursive QAOA, in solving binary optimization problems.
β’ Enhanced quantum computing research & education framework for Cleveland Clinic (CCF) by making Quantum Repository, a compilation & documentation of IBM-CCF research and code-base.
β’ Developed quantum computing education module for healthcare & life sciences research for CCF. Made quantum computing poster for touchscreen interactive panel as info board on the on-site CCF quantum computer.
β’ Part of the Quantum Intern Advisory Board: Arranged and initiated 10+ events, panel discussions, and talks, for all 2023 interns.
Developed internal WMI software suites for real-time visualizations of quantum computing hardware experimental data with Plotly Dash.
Developed a novel Monte Carlo simulation within epistemically-restricted phase-space formulation of quantum mechanics for quantum many-body systems. This novel method allows for a better calculation time at a large sampling with reasonable accuracy. Used TensorFlow to code the simulation and for automatic optimization.
Investigated the Quantum Generative Adversarial Networks algorithm to simulate the $t\bar{t}H(b\bar{b})$ production processes in the LHC experiment. Benchmarked the results with the classical models, studied how the quantum model affects the performance. Used TensorFlow Quantum to build and train the model.
Conducted research on the potential of Quantum Convolutional Neural Networks in classifying images of particles from high-energy physics events. Benchmarked the results with the classical models, studied how the quantum model affects the performance. The results and codes used for this project are publicly available as an open-source project. The official project page is here.
β Studied the Quantum Graph Recurrent Neural Networks for finding the Hamiltonian parameters of transverse-field Ising model (QGRNN, Verdon, et al.) and translated the algorithm to Qiskit code. Presented a code demo at the Qiskit Advocate Monthly Meetup.
β‘ Developing Qiskit’s Circuit Library Python module for data re-uploading classifier (PΓ©rez-Salinas, et al.) and Notebook tutorial on how to train the circuit as a quantum machine learning model for the Qiskit Textbook.
Designed variational quantum circuits that work similarly to convolution filters in CNN to classify images. The proposed architectures achieved higher testing accuracy compared to some previous related works. Click here for the project page.
Designed a real-time face tracking and blemish removal system to create a webcam filter application. A numerical threshold in HSV color space and elliptical kernel dilations was applied to the image to detect the skin. Blemishes were detected by utilizing CLAHE and blob detection. Trained a YOLO model to track the face using PyTorch. Coded algorithms for blemish removal using OpenCV. Achieved 85-90% of blemishes removal.
Conducted research in utilizing Error-state Kalman Filter and Diagonal Recurrent Neural Network & LSTM to make the localization of an autonomous car more reliable in the absence of GPS signal, reducing 70% of localization errors. Gathered the training data using CARLA Simulator. Trained and tested the model using Keras and TensorFlow.
Did research in using Convolutional Neural Networks to classify types of touch interaction (poke, scratch, etc.) from humans by learning the data pattern from a force sensor (ShokacChip TS). Trained the model using Keras and TensorFlow. The model reached 88% real-time accuracy. Coded a robotic arm’s servos using inverse kinematics in MATLAB to make it moves according to the type of touch being predicted by the network as a response.
Experience
Teaching
Delivered academic and handsβon tutorials (software, programming languages, practicum kits). Provided students with assistance on exam preparations, laboratory activities, assessed quizzes, and homework. Subjects covered:
- Engineering Drawing (Aug β Dec 2017)
- Introduction to Information Technology (Jan - May 2018)
- Fluid Mechanics (Jan β May 2019)
- Electric Circuits and Electronics (Aug β Dec 2019)
- Wave Phenomena (Aug - Dec 2020)
Projects
Quantum Convolutional Neural Networks for High-Energy Physics Analysis at the LHC
A Google Summer of Code 2021 Project Repository. This project aims to demonstrate quantum machine learning’s potential, specifically Quantum Convolutional Neural Network (QCNN), in HEP events classification from particle image data. The code used in the research is wrapped as an open-source package to ease future research in this field.
Event Classification with Layerwise Learning for Data Re-uploading Classifier in High-Energy Physics
This project aims to use modified layerwise learning on data re-uploading classifier to classify events in HEP. The project won second place at Xanadu’s QHack Quantum Machine Learning Open Hackathon 2021.
Quantum Image Classifier Design with Data Re-uploading Quantum Convolution and Data Re-uploading Classifier Scheme
Undergraduate Thesis Project.
This research proposed a modification scheme of the VQA-based Data Reuploading Classifier (DRC) and a DRC-based quantum convolution scheme for MNIST classification. The proposals achieved an improvement in accuracy compared to the previous related works.
Nightmode Photography: Exposure Fusion
Implementation of night mode photography algorithm utilizing histogram equalization, image registration, and Mertens exposure fusion.
Quantum Distance-based Classifier
This is a code implementation of distance-based classifier (similar to k-nearest neighbour algorithm) using quantum computer adapted from research paper "Implementing a distance-based classifier with a quantum interference circuit" by Maria Schuld, Mark Fingerhuth, and Francesco Petruccione.
Learning Human Touch Interaction with Convolutional Neural Networks
This project is the works done during my research internship at Gentiane Venture’s Laboratory, Tokyo Univ. of Agriculture and Technology, Japan. The project is about using CNN to classify several types of touch interaction from humans by learning the data pattern from a force sensor.
Recent Publications
Recent Talks
Achievements & Honors
Open Hackathon project title: Enhanced Quantum Autoencoders for Anomaly Detection.
Our team presented an extensive study of Quantum Autoencoders (QAE) from the application and the theoretical side. The project is awarded in three categories:
β’ 1st in Open Hackathon Experiment on Amazon Braket Simulators
β’ 1st in Hybrid Algorithms Challenge
β’ 3rd in Quantum Finance Challenge
The repository of the project can be accessed here. The official announcement about the competition and the winners can be read here.
β’ Translated more than 15000 words.
β’ Passed the Qiskit Advocate test and interview.
β’ Proposed a machine learning algorithm to solve supply chain problems in Indonesia to decrease unemployment and poverty.
