Giwon Shin
Hey there 👋
I'm Giwon Shin
Machine Learning Engineer & Researcher
This is my homepage where I share my projects and works on machine learning, research, and more. I'm passionate about building things that make a difference and love to learn new technologies. Feel free to explore and connect with me!
Publication highlights
MuSe: Packing-Aware Causal Sequence Modeling under Fully Homomorphic Encryption

MuSe: Packing-Aware Causal Sequence Modeling under Fully Homomorphic Encryption
Pinned

This master's thesis turns MuSe from an alternative multilinear sequence model into a packing-aware architecture for pure-FHE causal inference under bootstrapped CKKS, with both outer- and inner-based packing and cache-enabled generation.

Thesis Homomorphic Encryption Sequence Modeling Tensor operators
Project highlights
Inducing and Analyzing Hallucination in Diffusion LLMs

Inducing and Analyzing Hallucination in Diffusion LLMs
Pinned

We studied how factual hallucinations can be induced in diffusion-based language models—which iteratively denoise entire sequences rather than generate tokens autoregressively—to understand what these failures reveal about their internal behavior.

Hallucination Diffusion LLM
Exploring Beyond VisRAG

Exploring Beyond VisRAG
Pinned

We evaluated multimodal RAG on visually simple but text-heavy documents to analyze where VisRAG breaks down, how far text-only retrieval can go, and whether combining the two modalities produces a stronger retrieval system.

RAG VLM Multimodal
Multilinear Operators for Arithmetic Tasks

Multilinear Operators for Arithmetic Tasks
Pinned

This project explored whether sequence-to-sequence models could be built entirely from multilinear operations, without softmax attention, nonlinear activations, or positional encodings, and resulted in MuSe, a multilinear sequence model designed for arithmetic tasks such as addition, multiplication, and parity.

Polynomial networks Tensor operators
© 2026 Giwon Shin