IEEE Transactions on Biomedical Circuits and Systems (2025/03)

電子研究所 吳易忠 (■第一作者 ) / JIF:5.1 / Rank:0.247

本研究實現了全球第一個端到端的次世代基因定序資料分析晶片，可即時分析遺傳疾病與癌症檢測。本晶片整合所有資料分析步驟，同時在各項評比(通量、能量與面積效率)中取得最高效能。本研究亦達到現有文獻中之最高準確度與靈敏度，為精準醫療開創全新的應用。  全球第一端到端基因定序資料分析晶片  現有文獻最高之準確度(99.79%)與靈敏度(99.03%)  現有平台中之最短分析時間(人類30倍基因僅16分鐘)  現有文獻中最高之能量效率與面積效率

Small (Wiley) (2025/03)

電子研究所 Ray-Hua Horng (▓ Corresponding Author ) / JIF:13 / Rank:0.0726

Selective UV-A and UV-C Synaptic Response A ZnO/ZnGa2O4 heterojunction is demonstrated as a wavelength-selective deep ultraviolet (DUV) artificial synapse, showing distinct responses to UV-A and UV-C illumination. The device exhibits persistent photoconductivity under UV-A, enabling emulation of key synaptic functions such as short-term plasticity (STP), long-term potentiation (LTP), and paired-pulse facilitation (PPF). Under UV-C, the response is rapid and decay-free, suitable for high-speed photodetection. Pyrophototronic Modulation for Synaptic Plasticity The artificial synaptic behavior is governed by the pyrophototronic effect, where thermal gradients generated under UV-A illumination induce a pyroelectric field in ZnO. This field modulates carrier dynamics at the ZnO/ZnGa2O4 interface, enhancing synaptic weight updates during learning and gradually fading during forgetting. The low thermal conductivity of ZnGa2O4 enables prolonged heat retention, crucial for mimicking biological temporal dynamics. Self-Powered, Fast Operation The device operates in a self-powered mode with sub-nanosecond switching speed under UV-C and long retention under UV-A. This dual-mode operation makes it ideal for next-generation neuromorphic computing systems and energy-efficient optical synaptic sensors operating in harsh UV

IEEE Journal of Biomedical and Health Informatics (2025/01)

電控工程研究所 吳炳飛 (■ 通訊作者) / JIF:6.7 / Rank:MEDICAL INFORMATICS 6.82% (3/44)

我們的研究開發出一種非接觸式量測心房顫動的方法，僅需具備消費級攝影鏡頭的裝置即可進行檢測，無需配戴心電圖或專業儀器。主要貢獻包括： 1) 提出輕量化AI模型，相較先前SOTA方法，參數量減少92.03%、GFLOPs下降20%、延遲縮短47.62% 2) 在不同晃動與光線變化情境下，模型平均準確率仍維持94%以上，並在我們所收錄的大量心律不整的醫院資料庫上驗證其穩定性 3) 全系統可於僅搭載CPU的中階手機與筆電上運行 4) 此論文更被選為IEEE JBHI January的Feature Article。

Engineering Applications of Artificial Intelligence (2025/03)

電控工程研究所 林顯易 (■第一作者 ) / JIF:7.5 / Rank:0.07

This paper proposes a novel 3D path planning method for robotic manipulators using an artificial potential field and a force sensor to detect contact with obstacles in narrow spaces—overcoming limitations of cameras and lasers. Unlike traditional sensors, the force sensor only gathers data upon contact, which increases the risk of local minima. To address this, the method introduces a rotating repulsive force field aligned with the tool and path directions. Experimental results show that this approach avoids local minima and enables smooth navigation through tight spaces.

Journal of Neurodevelopmental Disorders (2024/11)

電控工程研究所 柯立偉 (■ 通訊作者) / JIF:4.1 / Rank:Top 16.78% of CLINICAL NEUROLOGY

Background: A multi-method, multi-informant approach is crucial for evaluating attention-deficit/hyperactivity disorders (ADHD) in preschool children due to the diagnostic complexities and challenges at this developmental stage. However, most artificial intelligence (AI) studies on the automated detection of ADHD have relied on using a single datatype. This study aims to develop a reliable multimodal AI-detection system to facilitate the diagnosis of ADHD in young children. Methods: 78 young children were recruited, including 43 diagnosed with ADHD (mean age: 68.07 ± 6.19 months) and 35 with typical development (mean age: 67.40 ± 5.44 months). Machine learning and deep learning methods were adopted to develop three individual predictive models using electroencephalography (EEG) data recorded with a wearable wireless device, scores from the computerized attention assessment via Conners' Kiddie Continuous Performance Test Second Edition (K-CPT-2), and ratings from ADHD-related symptom scales. Finally, these models were combined to form a single ensemble model. Results: The ensemble model achieved an accuracy of 0.974. While individual modality provided the optimal classification with an accuracy rate of 0.909, 0.922, and 0.950 using the ADHD-related symptom rating scale, the K-CPT-2 score, and the EEG measure, respectively. Moreover, the findings suggest that teacher ratings, K-CPT-2 reaction time, and occipital high-frequency EEG band power values are significant features in identifying young children with ADHD. Conclusions: This study addresses three common issues in ADHD-related AI research: the utility of wearable technologies, integrating databases from diverse ADHD diagnostic instruments, and appropriately interpreting the models. This established multimodal system is potentially reliable and practical for distinguishing ADHD from TD, thus further facilitating the clinical diagnosis of ADHD in preschool young children.

IEEE Microwave and Wireless Technology Letters (2025/04)

電信工程研究所 蔡作敏 (■ 通訊作者) / JIF:2.71 / Rank:在SCImago Journal Rank（SJR）中，MWTL的SJR值為0.77，整體排名為第6,585位，並在凝聚態物理學領域被列為Q1區

這篇信件論文介紹了兩個單刀雙擲（SPDT）開關（SPDT-1 和 SPDT-2），它們採用了相變材料（PCM）為基礎的開關元件，並結合匹配網路來補償高頻寄生效應，同時利用低阻抗傳輸線作為新穎的偏壓電路，工作頻率範圍從0（直流）到200 GHz。量測結果顯示，所提出的開關在太赫茲（THz）頻段的表現優於現有的射頻（RF）開關(在200 GHz以內損耗低於2.2 dB)。據作者所知，這項研究是首次開發出可在高達200 GHz頻率下運作的PCM基礎SPDT開關。

IEEE Transactions on Information Theory (2025/03)

電信工程研究所 賴青沂 (■ 通訊作者 ■ 共同作者(作者順序 2 )) / JIF:2.2 / Rank:COMPUTER SCIENCE, 138/250

本論文針對量子穩定碼在現象型錯誤模型下的解碼問題提出創新解法。本文提出廣義量子data-syndrome (DS) code與混合字母的檢查矩陣，能同時描述量子位錯誤與量測錯誤，並建構出具有混合節點的 Tanner graph，進而設計出適用於稀疏量子碼的可靠度傳遞(BP)解碼器。模擬結果顯示：對距離為 d 的toric code，只需 d 輪量測即可達到超過 3% 的錯誤閾值。論文亦提出冗餘穩定子構造法，可支援單次量測下的錯誤更正，並證實 BP 解碼器在高測量錯誤率下仍具有高效表現。

ACS Applied Materials & Interfaces (2025/02)

光電研究所 劉柏村 (■第一作者 ■通訊作者) / JIF:8.5 / Rank:0.1548

本篇論文首次成功地利用水熱法並低於攝氏一百度的製程下製備高品質的銅?雜氧化鋅奈米柱作為元件的開關層並實現了顯著的元件間均勻性和記憶體操作穩定性的提升。奈米柱限制導電燈絲形成，防止過粗的燈絲形成。銅離子?雜適度提高氧空缺而促進銅離子的擴散。均勻?雜的銅離子參與燈絲形成，減少過程中對從電極擴散的銅離子的依賴。作為中介，這些離子使得燈絲更容易形成，並降低燈絲過粗與過細的問題，從而改善元件各項均勻性。

Applied surface science (2025/03)

光電工程學系 劉柏村 (■ 通訊作者) / JIF:6.3 / Rank:13% (3/23)

本研究提出一種以超臨界流體（Supercritical Fluid, SCF）技術處理的 InGaZnO（IGZO）光電子突觸電晶體，應用於類神經形態計算與影像辨識。相比傳統熱處理方式（FA），SCF 處理具備低溫、低能耗、製程簡便等優勢，所製成的元件展現出優異的突觸可塑性。此研究更結合卷積神經網路（CNN）進行 CIFAR-10 圖像辨識，FASCF 處理之元件在高雜訊條件下仍達 91.8% 的準確率，優於傳統元件，展現其應用於智慧感知系統與神經網路硬體的潛力。

Nanomaterials (2025/01)

光電工程學系 郭浩中 (■ 通訊作者) / JIF:4.4 / Rank:0.73

本研究首次展示以有機薄膜電晶體（OTFTs）驅動主動式Mini-LED背光系統，成功達到超過100,000 nits的高亮度，並具備極佳的電性穩定性與對比度（高達1,000,000:1）。所開發的5光罩製程OTFT具備高開關比（>10?）、低啟動電壓與優異載子遷移率（~2.15 cm?/V·s），實現數位背光驅動與PWM調光。此技術有望推進低功耗、高亮度顯示應用，特別適用於HDR顯示、穿戴裝置與未來柔性顯示平台。

ACS Photonics (2025/03)

光電工程學系 陳姿伶 (■ 通訊作者) / JIF:6.5 / Rank:19/224

在我們最近的研究中，我們突破性地證明了利用被動式光腔系統放大圓二色性（即左旋與右旋光吸收能力的不對稱）的效果，並成功開發出一種新型雷射，該雷射可產生高達1.2的旋性光純度參數??lum，相較於傳統有機薄膜雷射的0.4，實現了明顯的進步。我們設計的光學架構能夠在雷射共振腔的兩側同時產生左旋與右旋圓偏振光，這一發現預示了在量子資訊領域，尤其是光自旋角動量轉換應用中的巨大潛力。這項工作不僅展示了具非互易性圓二色性（ACD）材料在旋性光源上的應用潛力，也為下一代掌性有機分子材料在光電領域的應用提供了新啟發。

Computers & Security (2025/04)

電機工程學系碩博士班 游家牧 (■共同作者 (作者順序4)) / JIF:4.8 / Rank:46/250

這篇論文針對控制區域網路（CAN）因缺乏加密與認證，面臨汽車及工業系統安全威脅，提出以機器學習（ML）為基礎的入侵偵測系統（IDS）防禦機制。研究融合對抗訓練、集成學習及距離優化三種防禦策略，並利用模擬退火演算法優化特徵空間距離，強化模型對已知及未知對抗性攻擊的抵抗力。結果顯示，在強力攻擊（如ZOO攻擊）下，結合三種防禦方法的系統能達到最高F1分數0.97（已知攻擊）及0.61（未知攻擊），遠優於單一方法。此研究創新地將距離優化整合於電動車模擬環境，顯著提升CAN IDS在面對敵意攻擊時的穩健性與效能。