High-level energy investigations of 3D Digital Compute-in-memory design

Novel 2.5D/3D IC design.

Analyzed physical characteristics of low-dimensional Tellurium.

Proposed Gate-all-around Field-effect-transistor (GAA-FET) using Tellurium Nanowire.

Compared with Carbon Nanotube FET with similar configurations.

Designed bandpass filter (1~100Hz, IIR Butterworth) & notch filter (59~61Hz, IIR Butterworth)

Acquire multi-channel brain waves invoked by thinking a vowel (without speaking out)

Transform the waves into Time-Frequency Representation (TFR)

Train a convolutional neural network for classification of brain waves using TFR

Proposed reinforcement learning method for autonomous chemical detection using drones.

Developed an environment emulating the Gaussian gas dispersion in urban areas.

Applied deep Q network (DQN) to train the optimal policy

Validated through the real flight experiment while spreading DMMP gas, a gas simulant of Sarin.

Observed peak distortion of Sulfur Hexafluoride (SF6) gas at high concentration using low resolution Fourier-Transform Infrared (FT-IR) spectroscopy.

Developed a distortion library of SF6 to enhance detection capability.

The performance increased by 15% compared to the single library approach.

Validated through the outdoor experiment.

It is extremely difficult to conduct the outdoor measurement of toxic gases.

Proposed deep generative model to augment the spectral data while controlling the attributes.

Compared the generated data with the linear model.

Investigate chemical sensing channel materials.

Analyzed edge-selectively oxidized graphene with HFIPPH functionals using DFT simulator.

Implemented gas experiments.

Voluntary, free lectures for freshmen to help them cope with exams.

Delivered physics 101 classes for 12 times in total. (1 ~ 2 hours per class)

Source: Fundamentals of Physics (Halliday et al.)