πŸ„β€β™€οΈ About Me

  • Highly motivated Computer Engineering graduate student with expertise in Software Development and Artificial Intelligence.
  • Skilled in Java, Python, Golang, C# and JavaScript.
  • I currently working under the supervision of Prof. Jianbo Jiao on Representation Learning and Application on Computer Vision tasks.
  • I have had solid research background in backdoor learning, feature representation in computer vision, and medical image segmentation. I have worked closely with Dr.Erick Purwanto and Prof.Jie Zhang during my undergraduate study.
  • Experienced in full-stack software development and interned at Wensi Haihui Information Technology (Pactera) Co., Ltd.
  • Also a published researcher with strong time management and teamwork skills.
  • Fluent in English, Mandarin, and Japanese.

My recent research interest includes robust machine learning and representation learning in computer vision.

πŸ”₯ News

  • 2023.12: Β πŸŽ‰πŸŽ‰ Glad to join MIx group as a research intern.
  • 2023.08: Β πŸŽ‰πŸŽ‰ Started my Master degree in UIUC ECE.
  • 2023.07: Β πŸŽ‰πŸŽ‰ Got my distinction bachelor degree at UOL and XJTLU.

πŸ“ Publications

  • SPSS: A Salience-based Poisoning Selection Strategy for Selecting Backdoor Attack Victims, Zihan Lyu1, Dongheng Lin1, Jie Zhang†, Ruiming Zhang2 IJCNN 2024
    • Designed an algorithm uses Salience Metric to evaluate sample feature significance towards backdoor learning process.
    • With its assistance, we managed to realized a more data-efficient backdoor attack to DNN models achieved the same attack success rate to vanilla backdoor attack with only 38.44% of poisoned samples.
CyberC 2022 (IEEE)
sym

Conditional Metadata Embedding Data Preprocessing Method for Semantic Segmentation

Juntuo Wang1, Qiaochu Zhao1, Dongheng Lin1, Erick Purwanto†, Ka Lok Man2

  • In this paper, we propose a conditional data preprocessing strategy, i.e., Conditional Metadata Embedding (CME) data preprocessing strategy. The CME data preprocessing method will embed conditional information to the training data, which can assist the model to better overcome the differences in the datasets and extract useful feature information in the images.

πŸŽ– Honors and Awards

  • 2022.10 XJTLU 2022 Summer Undergraduate Research Fellowship Poster Group Winner
  • 2022.07 University Academic Excellence Award β€” Scholarship for top 5% students

πŸ“– Education

πŸ’» Internships

  • 2023.11 - Now, Research Intern at MiX group at UofB, UK (Remote).
  • 2021.06 - 2021.10, Backend Developer Intern at Pactera, China.

πŸ–¨ Projects

sym

Arxiv Explorer: An efficient paper recommendation system we developed on a full dataset of all available Arxiv Papers (until Feb 2024).

  • All papers are compressed into paper embeddings using category, abstract, and title, stored in a FAISS indexing system.
  • We modeled a coauthorship map to conduct authority/hub based HITS reranking of the paper recommendation results.
sym

MapleJuice: A light-weight counterpart of Hadoop supported with SQL-like query

  • The distributed system is built upon a self-implemented file system similar to GFS with corresponding NameNode and DataNode.
  • We also implemented an efficient Gossip-style failure detection protocol to maintain all the node status using UDP packets, a Bully-algorithm based re-election ensures new leader will be available in case of any failures on master nodes.
  • The task scheduling mechanism is similar to MapReduce, ensures the parallelism among nodes. We tested the system against Hadoop within a cluster with 10 VMs. The MapleJuice is generally 25% faster than Hadoop when dealing with small clusters.
sym

My undergraduate thesis: StyleDiffuser: Cartoon-Style Image Creation with Diffusion Model and GAN Fusion

  • In this work, I have introduced a novel approach that fuses Generative Adversarial Networks (GANs) with the Stable Diffusion Model for creating cartoon-style images.
  • Utilizes StyleGAN2 generated feature maps and corresponding metadata to constrain the Stable Diffusion Network.
  • Reduces the number of diffusion steps required for the model to converge to a final image, streamlining the image generation process.
  • The method reduces the reliance on verbose prompts for controlling the output, making the generation process more straightforward.

πŸŽͺ Miscellaneous

sym
  • FUJICHROME Velvia 100, taken at my hometown.

πŸ¦‰ Site Visits