Advanced Materials (2024/03)

電子研究所 曾銘綸 (通訊作者) / JIF:29 / Rank:#15/211

此研究介紹了一個新型半導體共振梯度超穎介面(Resonance gradient metasurface)，其中通過局部高品質因子共振來實現所需的光譜選擇性，同時通過沿著平面結構逐漸調整單元尺寸實現寬頻之工作波長範圍。梯度超穎介面的高度可設計性可用於高精密分析特定生醫分子或材料之光學吸收並同時保持輕薄短小的元件面積。研究中利用這種超穎介面於新穎之生醫光譜檢測，包括混合聚合物之解析、生物分子間動態反應之監測，及振動強耦合(Vibrational strong coupling)。研究展示了梯度超穎介面的新穎物理機制及對於未來生醫檢測之應用潛力。

Nano letters (2024/01)

光電工程 黃耀緯 (通訊作者) / JIF:10.8 / Rank:0.103

本論文結合超穎介面與光子晶體面射型雷射之特性,使結構光投影儀的光學架構大幅簡化,且與市售產品如Apple相比,面積與功耗能夠大幅縮小的同時,還能達到相同的3D感測表現。除此之外,利用砷化鎵基板製作的超穎介面於未來能夠與面射型雷射進一步整合為單石整合元件,從而達到縮小整體體積的目的。

IEEE Transactions on Neural Networks and Learning Systems (2024/02)

電機工程學系 簡仁宗 (第一作者、通訊作者) / JIF:10.4 / Rank:3/54 (Computer Science, Hardware & Architecture)

流模型是利用一系列的可逆轉換直接估計出模型的生成分布，並且在訓練過程中，利用潛在變量精確地推論出對數相似度函數。本論文建構新穎性流模型，既平滑且擁有連續流形且透過潛在表示可以真實表達資料的豐富資訊。本研究在潛在空間中的解纏表示對於辨別獨立因素和解釋自然圖像或文本的分解特徵具有關鍵性貢獻，每一個特徵的解纏維度提供了一項特殊屬性，此屬性可被用來控制資料合成。本論文是生成式機器學習的代表著作之一。

Robotics and Computer-Integrated Manufacturing (2024/02)

電控所 林顯易 (第一作者、通訊作者) / JIF:10.4 / Rank:3/42, Category: Robotics, Q1

對於目前全世界淨零碳排在製造生產的貢獻上，本論文首先收提出基於人工智慧深度學習的工業機器人機械手能耗數位鑾生建模方法，並得到不錯成效。本論文所提出的模型在公開資料及 UR3e 和 UR10e 資料集上取得了優異的效能，均方根（RMS）誤差分別提高了 52.76% 和 52.4%。除此之外，本論文方法在工業機器手臂能量預測優於其他常見的機器學習模型，例如 Linear Regression (LR)、Regression Trees (RTs)、 Ensembles Tress (ETs)、Support Vector Machine (SVR)、feed-forward neural network以及CNN-GRU。

Chemosphere (2024/02)

電子所 荊鳳德 (通訊作者) / JIF:8.8 / Rank:0.11

粉狀微米或奈米顆粒光催化劑，會對環境造成二次污染並對動物有害。 為了解決這些問題，這項工作報告了在低成本基板上，形成 Cu2O-SnO2異質結構薄膜。 實驗表明，透過使用新的光催化劑方法，可以實現淨化有機物｢亞甲基藍和甲基橙｣，目前最佳的效果，而無二次污染。

ACS Sustainable Chemistry & Engineering (2024/01)

生醫所 李博仁 (通訊作者) / JIF:8.4 / Rank:91.2

 Novel method transforms SiC byproduct into efficient photocatalytic plates.  Nanogold-coated SiC plates exhibit superior photocatalytic activity, accelerating pollutant degradation.  Rapid synthesis of Au/SiC plates in 30 minutes contributes to sustainability.

IEEE Transactions on Communications (2024/02)

電信所 黃昱智 (通訊作者) / JIF:8.3 / Rank:26/275 = 0.094

此研究探討在分散式計算系統中進行批次大維度矩陣乘法時，利用稀疏編碼解決拖延者（stragglers）問題以加速運算完成時間。我們提出了一種新的編碼方案，稱為廣義稀疏編碼（Generalized Sparse Code，GSC），其中將矩陣分割成較小的子矩陣，這些子矩陣會被重新分組並由各自的稀疏編碼進行編碼。我們嚴謹的分析了GSC的預期完成時間，並基於此提出了與GSC相關的反向注水任務分配方案。通過這一原則可以開發出一個效率高的工作者分配算法。模擬結果驗證了所提出的GSC在各種稀疏程度下皆優於現有最佳之方案。

Future Generation Computer Systems (2024/03)

電機工程學系 温宏斌 (共同作者(作者順序 4 )) / JIF:7.5 / Rank:10/111=9%

 在商用可程式化交換機上實作出不需SDN控制器之IPS系統，其準確度高達99.357%  在具備輕量型CPU之交換機上實作基於機器學習之辨識模型，搭配Tofino ASIC之硬體封包特徵萃取能力，相較於SDN-based IPS可提升Throughput 183倍，並減少IPS之反應時間高達99.99%。

Journal of Materials Chemistry B (2024/01)

電機系所 黃薇蓁 (通訊作者) / JIF:7 / Rank:10/45

本論文利用電交聯製程成功開發了攜帶藥物的植入式蠶絲/明膠/石墨烯/抗發炎藥物複合電極。此電極可因不同含水條件下呈現極端的軟硬度（楊氏模量）變化，進而展現適宜植入式電極的電特性和可變的機械性。此外，本論文所開發的電極可經由電化學控制降解並進而釋放抗發炎藥物以中斷發炎星狀細胞的生長，減少發炎反應，也對神經元表現出無毒性。有別於傳統電極僅提供訊號傳輸，此電極綜合之生物相容性、可降解及載藥物特性，此多功能將提高植入式神經電極的利用價值。

IEEE Transactions on Neural Systems and Rehabilitation Engineering (2024/01)

電控工程研究所 柯立偉 (通訊作者) / JIF:4.9 / Rank:4/68 = top 5.88% (Q1) of REHABILITATION

Attention deficit hyperactivity disorder (ADHD) is a chronic neurological and psychiatric disorder that affects children during their development. To find neural patterns for ADHD and provide subjective features as decision references to assist specialists and physicians. Many studies have been devoted to investigating the neural dynamics of the brain through resting-state or continuous performance tests (CPT) with EEG or functional magnetic resonance imaging (fMRI). The present study used coherence, which is one of the functional connectivity (FC) methods, to analyze the neural patterns of children and adolescents (8-16 years old) under CPT and continuous auditory test of attention (CATA) task. In the meantime, electroencephalography (EEG) oscillations were recorded by a wireless brain-computer interface (BCI). 72 children were enrolled, of which 53 participants were diagnosed with ADHD and 19 presented to be typical developing (TD). The experimental results exhibited a higher difference in alpha and theta bands between the TD group and the ADHD group. While the differences between the TD group and the ADHD group in all four frequency domains were greater than under CPT conditions. Statistically significant differences (p < 0.05) were observed between the ADHD and TD groups in the alpha rhythm during the CATA task in the short-range of coherence. For the temporal lobe FC during the CATA task, the TD group exhibited statistically significantly FC (p < 0.05) in the alpha rhythm compared to the ADHD group. These findings offering new possibilities for more techniques and diagnostic methods in finding more ADHD features. The differences in alpha and beta frequencies were more pronounced in the ADHD group during the CPT task compared to the CATA task. Additionally, the disparities in brain activity were more evident across delta, theta, alpha and beta frequency domains when the task given was a CATA as opposed to a CPT. The findings presented the underlying mechanisms of the FC differences between children and adolescents with ADHD. Moreover, these findings should extend to use machine learning approaches to assist the ADHD classification and diagnosis.

Microchemical Journal (2024/01)

生醫所 李博仁 (通訊作者) / JIF:4.8 / Rank:80.8

 The electrochemical cell with transparent ITO heater was fabricated.  The cell was used in spectroelectrochemical studies of ferricyanide and cytochrome c.  The transparent ITO heater in the cell provides a rapid heating process within 7 mins.

IEEE Sensors Journal (2024/03)

電機系/電控所 戴立嘉 (共同作者(作者順序 3 )) / JIF:4.3 / Rank:Q1 (前 25%)

The Nernst equation defines the maximum sensitivity of metal oxide pH sensors, which has a theoretical value of 59.2 mV/pH. However, the phenomenon of super-Nernstian pH response of a metal oxide sensing membrane is related to the hydration of the sensing film. In this work, we demonstrated the DI-pretreated RuOx sample, with a super-Nernstian pH response of 64.13 mV/pH, with a linearity of 0.997. Additionally, the DI-pretreated RuO x EGFET sensor demonstrated several advantages, including a low hysteresis voltage (~2 mV), a minimal drift rate (0.25 mV/h), good flexibility, long-term stability (120 days), and excellent repeatability (with 97.95 % retention).

Optics Express (2024/01)

光電工程學系 陳皇銘 (通訊作者) / JIF:3.8 / Rank:30/100

本論文利用矽基液晶空間光調製器 (LCoS-SLM) 產生的932 nm結構光做動態影像投影在目標上，並以一台相機作為資料接收擷取並收集一系列移動光帶影像進行數據分析。 此方法解決了三角測量法中單一物體距離確定的限制，以及相機與光源之間角度要求的限制。在距離2公尺範圍內，平均量測誤差約 3%。這種基於單相機動態全像系統的多物體距離量測方法，適合應用於室內空間。

IEEE Transactions on Electron Devices (2024/01)

電信 李義明 (通訊作者、共同作者(作者順序2 ) ) / JIF:3.1 / Rank:41.25% (PHYSICS, APPLIED; Q2) 44% (ENGINEERING, ELECTRICAL & ELECTRONIC; Q2)

CFET為次一奈米積體電路技術節點的重要解決方案，李義明教授的研究團隊運用物理知情學習演算法針對矽基CFET以及奈米片電晶體進行人工智慧建模，發表一系列的長篇技術論文，團隊投稿的三篇文章同時被電子元件領域最頂尖的權威期刊: IEEE Transactions on Electron Devices接受，刊登於該期刊今年一月第七十一卷第一期的電路與系統設計之半導體元件模型特刊 (Special Issue on Semiconductor Device Modeling for Circuit and System Design) 中。根據統計，李義明教授的實驗室是我國過去五年(2020~2024)在同一卷(Volume)同一期(Issue)的IEEE Transactions on Electron Devices期刊中以同一實驗室、同一通訊作者發表最多論文的團隊。此研究不僅正持續深化，且其技術強度吸引半導體業界重視，正洽談技術合作中。

IEEE Microwave and Wireless Technology Letters (2024/02)

電信所 蔡作敏 (通訊作者) / JIF:3 / Rank:125/275

This letter presents the design and measured performance of a full Ka -band stacked-FET power amplifier (PA) fabricated using the 0.15- μ m Gallium arsenide (GaAs) pHEMT process. A stacked FET configuration is adopted in the power stage of the PA to achieve wideband power performance and linearity improvement at high input power. Measurements indicate that the proposed PA has a small-signal gain of 17 dB, a saturated output power ( PSAT ) of 24.5 dBm, an output 1-dB compression point ( OP1dB ) of 24.3 dBm, and a power-added efficiency (PAE) of 33.3%. The bandwidth of the PA ranges from 26 to 40 GHz, and thus covers the entire Ka -band. The proposed PA exhibits high performance, which makes it suitable for the applications of 5G wireless systems.