Bio
𝒯𝒾𝒶𝓃 𝒯𝒶𝓃 𝐵𝓊𝒹𝒹𝒽𝒶, 𝐻𝑜𝓃𝑔 𝒦𝑜𝓃𝑔
I am a Ph.D. student in Computer Engineering at Boston University, under the supervision of Prof. Vijaya Kolachalama. Prior to BU, I received my M.Phil. degree (2023) in Computer Science from City University of Hong Kong, supervised by Prof. Shiqi Wang, and B.Eng. degree (2020) in Intelligence Science and Technology from Northeast Electric Power University, supervised by Prof. Yimin Hou, Prof. Jinglei Lv, and Prof. Yang Li. In 2017, I attended a summer school at University of California, Irvine.
𝐏𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧 - Artificial Intelligence (AI) 𝐏𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧 - Graph Neural Network (GNN) 𝐏𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧 - Natural Language Processing (NLP) 𝐏𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧 - Computer Vision (CV) 𝐏𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧 - Tutorials and Useful Coding Scripts 𝐄𝐱𝐜𝐞𝐥𝐥𝐞𝐧𝐭 𝐏𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧 𝐓𝐞𝐜𝐡𝐧𝐢𝐜𝐚𝐥 𝐁𝐨𝐨𝐤𝐬 𝐀𝐜𝐚𝐝𝐞𝐦𝐢𝐜 𝐑𝐞𝐬𝐨𝐮𝐫𝐜𝐞𝐬 𝐑𝐞𝐜𝐨𝐦𝐦𝐞𝐧𝐝𝐞𝐝 𝐑𝐞𝐚𝐝𝐢𝐧𝐠𝐬
𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗜𝗻𝘁𝗲𝗿𝗲𝘀𝘁𝘀: I put all my effort into Multimodal Foundation Models, Reasoning and Problem Solving with Foundation Models, e.g., equip Language Models with strong mathematical reasoning capabilities, Generative Models, and AI for Medicine and Healthcare.
Research Focus Illustration
Research Presentations and Resources
𝐖𝐡𝐚𝐭 𝐢𝐬 𝐭𝐡𝐞 𝐦𝐨𝐬𝐭 𝐬𝐮𝐫𝐩𝐫𝐢𝐬𝐢𝐧𝐠 𝐟𝐢𝐧𝐝𝐢𝐧𝐠 𝐢𝐧 𝐲𝐨𝐮𝐫 𝐫𝐞𝐬𝐞𝐚𝐫𝐜𝐡?
Email: brucejia@bu.edu
Strava
Scholar
🤗 HuggingFace
GitHub
News
- Jan 2024 🔥 We are releasing 🤗 GSM8K-Consistency, a benchmark database for analyzing the consistency of Arithmetic Reasoning on GSM8K.
- Dec 2023 Open Source 🔨 PromptCraft and its published PyPI Package, a prompt perturbation toolkit from the character, word, and sentence levels for prompt robustness analysis .
- Oct 2023 Open Source Awesome Semantic Textual Similarity, a curated list of Semantic/Sentence Textual Similarity (STS) in Large Language Models (LLMs) and Natural Language Processing (NLP) .
- Oct 2023 A Presentation of Sentence Textual Similarity: Model Evolution Overview.
- Oct 2023 Open Source Awesome LLM Self-Consistency, a curated paper and presentation list of self-consistency in Large Language Models (LLMs) .
- Oct 2023 A Presentation of Prompt Perturbation and Robustness Evaluation.
- Sept 2023 A Presentation of Prompt-based Learning and Robustness Evaluation.
- Sept 2023 A Presentation of Self-Consistency Benefits Large Language Models.
- Sept 2023 Our paper Deep Transfer Learning is accepted by IEEE T-IM.
- May 2023 𝗠.𝗣𝗵𝗶𝗹.𝗧𝗵𝗲𝘀𝗶𝘀: No-reference Image Quality Assessment via Non-local Modeling and 𝗗𝗲𝗳𝗲𝗻𝘀𝗲 𝗦𝗹𝗶𝗱𝗲𝘀: Image Quality Assessment and Perceptual Optimization: A Non-local Modeling Approach.
- Mar 2023 A Presentation of Foundation Models for Sequential Decision Making.
- Jan 2023 A Presentation of IQA Regression and EEG Classification.
- Nov 2022 A Research Proposal of Video Panoptic Segmentation (VPS).
- Aug 2022 Our paper GCNs-Net is accepted by IEEE T-NNLS .
- Aug 2022 Our paper NLNet is accepted by IEEE MMSP. Source codes are available on GitHub , and the trained models are available on 🤗 HuggingFace for real-life IQA inference.
- Dec 2021 Our paper BiLSTM-GCNs is accepted by Frontiers in Bioengineering and Biotechnology.
- Apr 2020 Open Source EEG-DL, a Deep Learning (DL) library written by TensorFlow for EEG Signals Classification .
- Feb 2020 Our paper ESI-CNNs is accepted by Journal of Neural Engineering .
Publications
-
No-reference Image Quality Assessment via Non-local Dependency Modeling
Shuyue Jia, Dingquan Li, Shiqi Wang
IEEE 24th International Workshop on Multimedia Signal Processing (IEEE MMSP'22)See More
-
A no-reference image quality assessment method based on non-local features learned by a graph neural network (GNN). The proposed quality assessment framework is rooted in the view that the human visual system perceives image quality with long-dependency constructed among different regions, inspiring us to explore the non-local interactions in quality prediction.
-
GCNs-Net: A Graph Convolutional Neural Network Approach for Decoding Time-resolved EEG Motor Imagery SignalsShuyue Jia, Yimin Hou, Xiangmin Lun, Yan Shi, Yang Li, Rui Zeng, Jinglei Lv
IEEE Transactions on Neural Networks and Learning Systems (IEEE T-NNLS)
See More
-
Traditional works classify EEG signals without considering the topological relationship among electrodes. Thus, a graph convolutional neural network is presented while cooperating with the functional topological relationship of electrodes.
-
Deep Feature Mining via Attention-based BiLSTM-GCN for Human Motor Imagery Recognition
Yimin Hou, Shuyue Jia (Corresponding Author), Xiangmin Lun, Shu Zhang, Jinglei Lv
Frontiers in Bioengineering and BiotechnologySee More
-
This paper presents a novel deep learning approach designed toward both remarkably accurate and responsive motor imagery (MI) recognition based on scalp EEG. Bidirectional long short-term memory (BiLSTM) with the attention mechanism is employed, and the graph convolutional neural network (GCN) promotes the decoding performance by cooperating with the topological structure of features.
-
A Novel Approach of Decoding EEG Four-Class Motor Imagery Tasks via Scout ESI and CNN
Yimin Hou, Lu Zhou, Shuyue Jia, Xiangmin Lun
Journal of Neural EngineeringSee More
-
We presented a novel approach that could potentially improve the current stroke rehabilitation strategies by implementing a deep learning approach for an Electroencephalogram (EEG) based on MI Brain-Computer Interface System.
- Constructed 6 convolutional layers, 2 max-pooling layers, and 3 FC layers CNNs for four-class motor imagery classification through TensorFlow, with 50% dropout (spatial dropout after every Conv layer and regular dropout for FC layers) – 11.44% accuracy improvement, batch normalization (BN) – 10.15% improvement, and Short-cut Connection – 1.76% improvement to prevent overfitting, and achieved SOTA results: 94.50% accuracy on scout R5, 94.54% at subject level, and 96% for left fist prediction.
- Took charge of DNNs design, including methods comparisons, such as MLPs, CNNs, RNNs, and LSTMs, classification results calculations, and programming. 10 and 14 subjects’ data were utilized (19,320 and 27,048 samples in the experiments)
- Benchmark Dataset: EEG Motor Movement/Imagery Dataset.
Academic Services
- Reviewer of IEEE T-MM, IEEE T-CSVT, and IEEE Journal of Biomedical and Health Informatics
- Student Member of IEEE, ACM, ACL, and AAAI
Selected Awards
-
CityU Top 5 Runner, City University of Hong Kong
Athlete
-
Outstanding Athlete, Northeast Electric Power University
Athlete
-
3000-meter Steeplechase, The 45th Northeast Electric Power University Games
The 7th Place in college
-
2015 National High School Math League, China
Second Prize
-
The 32nd Chinese Physics Olympiad (CPhO), China
Third Prize
