Hi, I'm Yuxi Yang.
Chasing the stars through
code & cognition.

A comprehensive study comparing convolutional and attention-based architectures on the CIFAR-10 benchmark, exploring the fundamental trade-offs between inductive bias and data-driven learning.

Approach

Implemented a ResNet model with 2.8M parameters from scratch using PyTorch. The network was trained with standard augmentation techniques including random cropping and horizontal flipping, achieving 93% test accuracy. In parallel, trained a Vision Transformer (ViT) to understand how self-attention mechanisms compare with convolutional feature extraction on relatively small-scale datasets.

Key Results

Conducted systematic convergence analysis comparing both architectures. Applied cosine learning rate scheduling and mixed precision (FP16) training to accelerate convergence. Documented the effect of patch size, depth, and attention heads on ViT performance, revealing the data-efficiency gap between CNNs and transformers at the CIFAR-10 scale.

An end-to-end model optimization project that took YOLOv5 from a research prototype to a deployment-ready inference engine, published at ACM GAIIS 2024.

Pipeline

Designed a complete deployment pipeline: PyTorch → ONNX → TensorRT. Exported the trained YOLOv5 model to ONNX format with dynamic batch support, then compiled to TensorRT with INT8 post-training quantization using a calibration dataset for optimal range estimation.

Key Results

Achieved a 4.17× throughput improvement (789 vs. 189 QPS) with only 0.6 mAP drop and 70% model size reduction. The quantized model maintained real-time performance on edge GPU devices, demonstrating the viability of INT8 inference for vehicle detection in autonomous driving scenarios.

Publication

Proceedings of the ACM, May 2024

A rigorous benchmarking study of classical denoising algorithms, bridging variational methods and local filtering approaches under controlled experimental conditions.

Methods Evaluated

Implemented and compared 6 algorithms: MAP-L2, ROF (Total Variation), TGV2 (Total Generalized Variation), Non-Local Means, Bilateral Filter, and a classical baseline. Each method was tested across 50 COCO images corrupted with Gaussian, salt-and-pepper, and Poisson noise at varying intensities.

Technical Depth

Developed proximal gradient and Chambolle–Pock solvers for the variational models. Performed statistical performance analysis using PSNR, SSIM, and perceptual quality metrics. The study revealed that TGV2 consistently outperformed simpler TV models by preserving smooth gradients while maintaining sharp edges.

An exploration of neural radiance fields for photorealistic 3D reconstruction, comparing implicit volumetric representations with the emerging Gaussian Splatting paradigm.

Data & Pipeline

Captured a custom 59-view dataset and processed it through COLMAP for structure-from-motion and camera pose estimation. Trained a Nerfacto model using Nerfstudio, optimizing occupancy grids and appearance embeddings for high-fidelity novel view synthesis.

Comparative Analysis

Benchmarked volumetric NeRF against 3D Gaussian Splatting across three dimensions: rendering quality (PSNR, LPIPS), inference speed (FPS at various resolutions), and memory footprint. Gaussian Splatting showed significant speed advantages while NeRF maintained superior geometric consistency in fine-detail regions.