Chi-Sheng Chen

Chi-Sheng Chen

Auto-generated profile, based on publicly available information

Chi-Sheng Chen

Graduate Institute of Biomedical Electronics and Bioinformatics, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 10617

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 December 2025 in Journal of Signal Processing Systems
Verified
Journal Article
Open access
CCChi-Sheng Chen
Chi-Sheng Chen
SCSamuel Yen-Chi Chen
Samuel Yen-Chi Chen
HTHuan-Hsin Tseng
Huan-Hsin Tseng

8 August 2022 in Physical Review Research
Verified
Journal Article
Open access
PDF file
DTDaisuke Takegami
Daisuke Takegami
LNLaurent Nicolaï
Laurent Nicolaï
YUYuki Utsumi
Yuki Utsumi
AMAnna Meléndez-Sans
Anna Meléndez-Sans
DBDaria A. Balatsky
Daria A. Balatsky
LTLiu Hao Tjeng
Liu Hao Tjeng
It was hypothesized already more than 40 years ago that photoelectron spectroscopy should in principle be able to image atomic orbitals. If this can be made to work for orbitals in crystalline solids, one would have literally a different view on the electronic structure of a wide range of quantum materials. Here, we demonstrate how hard x-ray photoelectron spectroscopy can make direct images of the orbitals making up the band structure of our model system, ReO3. The images are energy specific and enable us to unveil the role of each of those orbitals for the chemical bonding and the Fermi surface topology. The orbital image information is complementary to that from angle-resolved photoemission and thus completes the determination of the electronic structure of materials.

25 June 2025 in JMIR Medical Informatics
Verified
Journal Article
Open access
CCChi-Sheng Chen
Chi-Sheng Chen
SCShao-Hsuan Chang
Shao-Hsuan Chang
CLChe Wei Liu
Che Wei Liu
Tung-Ming PanTung-Ming Pan
Tung-Ming Pan
Background: Electroencephalography (EEG) has been widely used to measure brain activity, but its potential to generate accurate images from neural signals remains a challenge. Most EEG-decoding research has focused on tasks such as motor imagery, emotion recognition, and brain wave classification, which involve EEG signal analysis and classification. Some studies have explored the correlation between EEG and images, primarily focusing on EEG-image pair classification or transformation. However, EEG-based image generation remains underexplored. Objective: The primary goal of this study was to extend EEG-based classification to image generation, addressing the limitations of previous methods and unlocking the full potential of EEG for image synthesis. To achieve more meaningful EEG-to-image generation, we developed a novel framework, Neural-Cognitive Multimodal EEG-Informed Image (NECOMIMI), which was specifically designed to generate images directly from EEG signals. Methods: We developed a 2-stage NECOMIMI method, which integrated the novel Neural Encoding Representation Vectorizer (NERV) EEG encoder that we designed with a diffusion-based generative model. The Category-Based Assessment Table (CAT) score was introduced to evaluate the semantic quality of EEG-generated images. In addition, the ThingsEEG dataset was used to validate and benchmark the CAT score, providing a standardized measure for assessing EEG-to-image generation performance. Results: The NERV EEG encoder achieved state-of-the-art performance in several zero-shot classification tasks, with an average accuracy of 94.8% (SD 1.7%) in the 2-way task and 86.8% (SD 3.4%) in the 4-way task, outperforming models such as Natural Image Contrast EEG, Multimodal Similarity-Keeping Contrastive Learning, and Adaptive Thinking Mapper ShallowNet. This highlighted its superiority as a feature extraction tool for EEG signals. In a 1-stage image generation framework, EEG embeddings often resulted in abstract or generalized images such as landscapes instead of specific objects. Our proposed 2-stage NECOMIMI architecture effectively extracted semantic information from noisy EEG signals, showing its ability to capture and represent underlying concepts derived from brain wave activity. We further conducted a perturbation study to test whether the model overly depended on visual cortex EEG signals for scene-based image generation. The perturbation of visual cortex EEG channels led to a notable increase in Fréchet inception distance scores, suggesting that our model relied heavily on posterior brain signals to generate semantically coherent images. Conclusions: NECOMIMI demonstrated the potential of EEG-to-image generation, revealing the challenges of translating noisy EEG data into accurate visual representations. The novel NERV EEG encoder for multimodal contrastive learning reached state-of-the-art performance both on n-way zero-shot and EEG-informed image generation. The introduction of the CAT score provided a new evaluation metric, paving the way for future research to refine generative models. In addition, this study highlighted the significant clinical potential of EEG-to-image generation, particularly in enhancing brain-machine interface systems and improving quality of life for individuals with motor impairments.

29 July 2026 in Quantum Science and Technology
Verified
Journal Article
Open access
CCChi-Sheng Chen
Chi-Sheng Chen
EKEn-Jui Kuo
En-Jui Kuo
A recurring weakness in quantum machine learning (QML) is that reported ‘quantum advantages’ are seldom tested against a capacity-matched classical control, leaving it unclear whether a gain comes from the quantum substrate or from the architectural change that accompanies it. Our primary contribution is methodological: a protocol for attributing such gains honestly—a capacity-matched classical bottleneck of identical parameter budget, transparent reporting of where quantum does not help, and validation on real quantum hardware—which we develop and apply through a concrete case study. That case study is Quantum Adaptive Self-Attention (QASA), a hybrid Transformer that replaces the value projection of a single encoder layer with a 36-parameter parameterized quantum circuit (PQC), keeping all other layers classical. Across nine synthetic benchmarks and the real-world ETTh1 dataset, QASA improves on a full-capacity classical Transformer for chaotic and trend-dominated signals. To ask whether this is a genuinely quantum effect, we introduce a control rarely applied in QML—a capacity-matched classical bottleneck with the same parameter budget—and find that it matches the PQC on the error metrics. The gain is therefore attributable to the low-rank value-projection bottleneck (an architectural parsimony principle), not to quantumness; adding further quantum layers only degrades performance and trainability. We accordingly position the quantum layer not as a source of accuracy advantage but as a competitive instantiation of this principle: its low-rank compression onto the signal’s intrinsic dimensionality is matched by a classical bottleneck, so the gain is architectural rather than quantum. The quantum layer’s distinguishing features are physical—a high circuit entanglement (Meyer–Wallach Q = 0.981 with only 27 CNOTs, which we report as a circuit-level property) and, most concretely, noisy intermediate-scale quantum deployability, which we verify by executing the trained model on a real IBM Quantum processor (one-step prediction within 7.5 % of noiseless simulation on the quantum-favored task and 25 % , with overlapping error bars, on a classical-favored control; no error mitigation). Relative to other quantum sequence models (quantum long short-term memory (QLSTM), QnnFormer), QASA is moreover the most resource-efficient of the three: it reaches competitive accuracy with the fewest quantum parameters (36 vs 90–128) and the strongest entanglement-per-gate ( Q = 0.981 with 27 CNOTs vs QLSTM’s 56), and retains trainable gradients that deeper circuits lose. We argue that capacity-matched baselines and honest reporting of where quantum does not help are prerequisites for credible quantum-machine-learning claims.

28 October 2024 in Preprints.org
Verified
Preprint
Open access
PDF file
CCChi-Sheng Chen
Chi-Sheng Chen
YYYu-Hsuan Yang
Yu-Hsuan Yang
GCGuan-Ying Chen
Guan-Ying Chen
SCShao-Hsuan Chang
Shao-Hsuan Chang
Background: Food properties can directly influence individual dietary intake. Therefore, computer vision-based food recognition could be used to estimate meal contents for patients with metabolic diseases. Food recognition based on deep learning can create opportunities for breakthroughs in dietary interventions for personal health management and have rapidly emerged as dietary assessment strategies. Methods: This study proposed a methodology for automatic food recognition based on fine-grained visual classification, the High-temperaturE Refinement and Background Suppression network (HERBS). A technical investigation was conducted involved in the HERBS model on CNFOOD241 benchmark to verify the model effectiveness. The visualization analysis of HERBS was compared with the VGG16 and RepViT on CNFOOD241. Results: The system achieved classification accuracy of 82.72 % and 97.19 % in Top-1 accuracy and Top-5 accuracy respectively. Data showed that HERBS structure widened the attention area and outperformed VGG16 and RepViT on feature maps. Conclusions: Our findings elucidated that the proposed methodology using HERBS model establishes a new benchmark for SOTA performance in food recognition on CNFOOD 241 dataset. This study proved the feasibility of the proposed approach in a challenging food dataset. We further described a vision-language model structure consisting of a multi-modal vision model, language encoder and a multilayer perceptron classifier for food composition recognition.

29 January 2026 in arXiv
Verified
Preprint
Open access
PDF file
CCChi-Sheng Chen
Chi-Sheng Chen
EKEn-Jui Kuo
En-Jui Kuo
GCGuan-Ying Chen
Guan-Ying Chen
XZXinyu Zhang
Xinyu Zhang
FZFan Zhang
Fan Zhang
Spatial covariance matrices of EEG signals are Symmetric Positive Definite (SPD) and lie on a Riemannian manifold, yet the theoretical connection between embedding geometry and optimization dynamics remains unexplored. We provide a formal analysis linking embedding choice to gradient conditioning and numerical stability for SPD manifolds, establishing three theoretical results: (1) BWSPD's $\sqrtκ$ gradient conditioning (vs $κ$ for Log-Euclidean) via Daleckii-Kreĭn matrices provides better gradient conditioning on high-dimensional inputs ($d \geq 22$), with this advantage reducing on low-dimensional inputs ($d \leq 8$) where eigendecomposition overhead dominates; (2) Embedding-Space Batch Normalization (BN-Embed) approximates Riemannian normalization up to $O(\varepsilon^2)$ error, yielding $+26\%$ accuracy on 56-channel ERP data but negligible effect on 8-channel SSVEP data, matching the channel-count-dependent prediction; (3) bi-Lipschitz bounds prove BWSPD tokens preserve manifold distances with distortion governed solely by the condition ratio $κ$. We validate these predictions via a unified Transformer framework comparing BWSPD, Log-Euclidean, and Euclidean embeddings within identical architecture across 1,500+ runs on three EEG paradigms (motor imagery, ERP, SSVEP; 36 subjects). Our Log-Euclidean Transformer achieves state-of-the-art performance on all datasets, substantially outperforming classical Riemannian classifiers and recent SPD baselines, while BWSPD offers competitive accuracy with similar training time.