Pub Date : 2024-08-02DOI: 10.1038/s44172-024-00254-9
Chen Wang, Haoyang Cui, Qinghao Zhang, Paul Calle, Yuyang Yan, Feng Yan, Kar-Ming Fung, Sanjay G. Patel, Zhongxin Yu, Sean Duguay, William Vanlandingham, Ajay Jain, Chongle Pan, Qinggong Tang
Percutaneous renal biopsy is commonly used for kidney cancer diagnosis. However, the biopsy procedure remains challenging in sampling accuracy. Here we introduce a forward-viewing optical coherence tomography probe for differentiating tumor and normal tissues, aiming at precise biopsy guidance. Totally, ten human kidney samples, nine of which had malignant renal carcinoma and one had benign oncocytoma, were used for system evaluation. Based on their distinct imaging features, carcinoma could be efficiently distinguished from normal renal tissues. Additionally, oncocytoma could be differentiated from carcinoma. We developed convolutional neural networks for tissue recognition. Compared to the conventional attenuation coefficient method, convolutional neural network models provided more accurate carcinoma predictions. These models reached a tissue recognition accuracy of 99.1% on a hold-out set of four kidney samples. Furthermore, they could efficiently distinguish oncocytoma from carcinoma. In conclusion, our convolutional neural network-aided endoscopic imaging platform could enhance carcinoma diagnosis during percutaneous renal biopsy procedures. Chen Wang and colleagues develop a forward-viewing optical coherence tomography endoscope for differentiating tumor tissues in renal biopsy. In conjunction with a convolutional neural network developed by the team, tissue recognition rates of over 99% were achieved.
{"title":"Automatic renal carcinoma biopsy guidance using forward-viewing endoscopic optical coherence tomography and deep learning","authors":"Chen Wang, Haoyang Cui, Qinghao Zhang, Paul Calle, Yuyang Yan, Feng Yan, Kar-Ming Fung, Sanjay G. Patel, Zhongxin Yu, Sean Duguay, William Vanlandingham, Ajay Jain, Chongle Pan, Qinggong Tang","doi":"10.1038/s44172-024-00254-9","DOIUrl":"10.1038/s44172-024-00254-9","url":null,"abstract":"Percutaneous renal biopsy is commonly used for kidney cancer diagnosis. However, the biopsy procedure remains challenging in sampling accuracy. Here we introduce a forward-viewing optical coherence tomography probe for differentiating tumor and normal tissues, aiming at precise biopsy guidance. Totally, ten human kidney samples, nine of which had malignant renal carcinoma and one had benign oncocytoma, were used for system evaluation. Based on their distinct imaging features, carcinoma could be efficiently distinguished from normal renal tissues. Additionally, oncocytoma could be differentiated from carcinoma. We developed convolutional neural networks for tissue recognition. Compared to the conventional attenuation coefficient method, convolutional neural network models provided more accurate carcinoma predictions. These models reached a tissue recognition accuracy of 99.1% on a hold-out set of four kidney samples. Furthermore, they could efficiently distinguish oncocytoma from carcinoma. In conclusion, our convolutional neural network-aided endoscopic imaging platform could enhance carcinoma diagnosis during percutaneous renal biopsy procedures. Chen Wang and colleagues develop a forward-viewing optical coherence tomography endoscope for differentiating tumor tissues in renal biopsy. In conjunction with a convolutional neural network developed by the team, tissue recognition rates of over 99% were achieved.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11297278/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141879879","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
For steel bridges, corrosion has historically led to bridge failures, resulting in fatalities and injuries. To enhance public safety and prevent such incidents, authorities mandate in-situ evaluation and reporting of corroded members. The current inspection and evaluation protocol is characterized by intense labor, traffic delays, and poor capacity predictions. Here we combine full-scale experimental testing of a decommissioned girder, 3D laser scanning, and convolutional neural networks (CNNs) to introduce a continuous inspection and evaluation framework. Classification and regression CNNs are trained on a databank of 1,421 naturally inspired corrosion scenarios, generated computationally based on point clouds of three corroded girders collected in lab conditions. Results indicate low errors of up to 2.0% and 3.3%, respectively. The methodology is validated on eight real corroded ends and implemented for the evaluation of an in-service bridge. This framework promises significant advancements in assessing aging bridge infrastructure with higher accuracy and efficiency compared to analytical or semi-analytical approaches. Dr Georgios Tzortzinis and colleagues use a combination of experimental testing and 3D laser scanning to describe the corrosion profile of bridge girders. Their results demonstrate how laser scanners and convolutional neural networks can provide accurate predictions on the structural capacity of ageing steel bridges.
{"title":"Structural integrity of aging steel bridges by 3D laser scanning and convolutional neural networks","authors":"Georgios Tzortzinis, Angelos Filippatos, Jan Wittig, Maik Gude, Aidan Provost, Chengbo Ai, Simos Gerasimidis","doi":"10.1038/s44172-024-00255-8","DOIUrl":"10.1038/s44172-024-00255-8","url":null,"abstract":"For steel bridges, corrosion has historically led to bridge failures, resulting in fatalities and injuries. To enhance public safety and prevent such incidents, authorities mandate in-situ evaluation and reporting of corroded members. The current inspection and evaluation protocol is characterized by intense labor, traffic delays, and poor capacity predictions. Here we combine full-scale experimental testing of a decommissioned girder, 3D laser scanning, and convolutional neural networks (CNNs) to introduce a continuous inspection and evaluation framework. Classification and regression CNNs are trained on a databank of 1,421 naturally inspired corrosion scenarios, generated computationally based on point clouds of three corroded girders collected in lab conditions. Results indicate low errors of up to 2.0% and 3.3%, respectively. The methodology is validated on eight real corroded ends and implemented for the evaluation of an in-service bridge. This framework promises significant advancements in assessing aging bridge infrastructure with higher accuracy and efficiency compared to analytical or semi-analytical approaches. Dr Georgios Tzortzinis and colleagues use a combination of experimental testing and 3D laser scanning to describe the corrosion profile of bridge girders. Their results demonstrate how laser scanners and convolutional neural networks can provide accurate predictions on the structural capacity of ageing steel bridges.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11294330/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141876863","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-31DOI: 10.1038/s44172-024-00249-6
Pedro Rolo, João V. Vidal, Andrei L. Kholkin, Marco P. Soares dos Santos
Triboelectric and piezoelectric energy harvesters can hardly power most microelectronic systems. Rotational electromagnetic harvesters are very promising alternatives, but their performance is highly dependent on the varying mechanical sources. This study presents an innovative approach to significantly increase the performance of rotational harvesters, based on dynamic coil switching strategies for optimization of the coil connection architecture during energy generation. Both analytical and experimental validations of the concept of self-adaptive rotational harvester were carried out. The adaptive harvester was able to provide an average power increase of 63.3% and 79.5% when compared to a non-adaptive 16-coil harvester for harmonic translation and harmonic swaying excitations, respectively, and 83.5% and 87.2% when compared to a non-adaptive 8-coil harvester. The estimated energy conversion efficiency was also enhanced from ~80% to 90%. This study unravels an emerging technological approach to power a wide range of applications that cannot be powered by other vibrationally driven harvesters. Pedro Rolo and colleagues designed a rotational electromagnetic harvester using an adaptive coil-switching architecture. This adaptability significantly enhances the energy efficiency and power density, as demonstrated analytically and experimentally.
{"title":"Self-adaptive rotational electromagnetic energy generation as an alternative to triboelectric and piezoelectric transductions","authors":"Pedro Rolo, João V. Vidal, Andrei L. Kholkin, Marco P. Soares dos Santos","doi":"10.1038/s44172-024-00249-6","DOIUrl":"10.1038/s44172-024-00249-6","url":null,"abstract":"Triboelectric and piezoelectric energy harvesters can hardly power most microelectronic systems. Rotational electromagnetic harvesters are very promising alternatives, but their performance is highly dependent on the varying mechanical sources. This study presents an innovative approach to significantly increase the performance of rotational harvesters, based on dynamic coil switching strategies for optimization of the coil connection architecture during energy generation. Both analytical and experimental validations of the concept of self-adaptive rotational harvester were carried out. The adaptive harvester was able to provide an average power increase of 63.3% and 79.5% when compared to a non-adaptive 16-coil harvester for harmonic translation and harmonic swaying excitations, respectively, and 83.5% and 87.2% when compared to a non-adaptive 8-coil harvester. The estimated energy conversion efficiency was also enhanced from ~80% to 90%. This study unravels an emerging technological approach to power a wide range of applications that cannot be powered by other vibrationally driven harvesters. Pedro Rolo and colleagues designed a rotational electromagnetic harvester using an adaptive coil-switching architecture. This adaptability significantly enhances the energy efficiency and power density, as demonstrated analytically and experimentally.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11291956/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141861841","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-25DOI: 10.1038/s44172-024-00246-9
Rashindrie Perera, Peter Savas, Damith Senanayake, Roberto Salgado, Heikki Joensuu, Sandra O’Toole, Jason Li, Sherene Loi, Saman Halgamuge
Tumour-Infiltrating Lymphocytes (TILs) are pivotal in the immune response against cancer cells. Existing deep learning methods for TIL analysis in whole-slide images (WSIs) demand extensive patch-level annotations, often requiring labour-intensive specialist input. To address this, we propose a framework named annotation-efficient segmentation and attention-based classifier (ANSAC). ANSAC requires only slide-level labels to classify WSIs as having high vs. low TIL scores, with the binary classes divided by an expert-defined threshold. ANSAC automatically segments tumour and stroma regions relevant to TIL assessment, eliminating extensive manual annotations. Furthermore, it uses an attention model to generate a map that highlights the most pertinent regions for classification. Evaluating ANSAC on four breast cancer datasets, we demonstrate substantial improvements over three baseline methods in identifying TIL-relevant regions, with up to 8% classification improvement on a held-out test dataset. Additionally, we propose a pre-processing modification to a well-known method, enhancing its performance up to 6%. Perera et al. developed a machine-learning approach for classifying whole-slide images into binary categories of tumour-infiltrating lymphocytes. They have benchmarked it against two established models and made the entire processing pipeline available as open source.
肿瘤浸润淋巴细胞(TIL)在针对癌细胞的免疫反应中起着关键作用。现有的全切片图像(WSI)TIL分析深度学习方法需要大量的斑块级注释,通常需要劳动密集型的专家输入。为了解决这个问题,我们提出了一个名为 "高效注释分割和基于注意力的分类器(ANSAC)"的框架。ANSAC 只需要玻片级标签就能将 WSI 划分为 TIL 分数高与低,二元类别由专家定义的阈值划分。ANSAC 自动分割与 TIL 评估相关的肿瘤和基质区域,省去了大量的人工标注。此外,它还利用注意力模型生成地图,突出显示最相关的分类区域。我们在四个乳腺癌数据集上对 ANSAC 进行了评估,结果表明,与三种基线方法相比,ANSAC 在识别 TIL 相关区域方面有了实质性的改进,在保留的测试数据集上,分类改进率高达 8%。此外,我们还对一种著名的方法提出了预处理修改建议,将其性能提高了 6%。Perera 等人开发了一种机器学习方法,用于将整张切片图像划分为肿瘤浸润淋巴细胞的二元类别。他们将该方法与两个成熟的模型进行了比对,并将整个处理管道作为开放源代码提供。
{"title":"Annotation-efficient deep learning for breast cancer whole-slide image classification using tumour infiltrating lymphocytes and slide-level labels","authors":"Rashindrie Perera, Peter Savas, Damith Senanayake, Roberto Salgado, Heikki Joensuu, Sandra O’Toole, Jason Li, Sherene Loi, Saman Halgamuge","doi":"10.1038/s44172-024-00246-9","DOIUrl":"10.1038/s44172-024-00246-9","url":null,"abstract":"Tumour-Infiltrating Lymphocytes (TILs) are pivotal in the immune response against cancer cells. Existing deep learning methods for TIL analysis in whole-slide images (WSIs) demand extensive patch-level annotations, often requiring labour-intensive specialist input. To address this, we propose a framework named annotation-efficient segmentation and attention-based classifier (ANSAC). ANSAC requires only slide-level labels to classify WSIs as having high vs. low TIL scores, with the binary classes divided by an expert-defined threshold. ANSAC automatically segments tumour and stroma regions relevant to TIL assessment, eliminating extensive manual annotations. Furthermore, it uses an attention model to generate a map that highlights the most pertinent regions for classification. Evaluating ANSAC on four breast cancer datasets, we demonstrate substantial improvements over three baseline methods in identifying TIL-relevant regions, with up to 8% classification improvement on a held-out test dataset. Additionally, we propose a pre-processing modification to a well-known method, enhancing its performance up to 6%. Perera et al. developed a machine-learning approach for classifying whole-slide images into binary categories of tumour-infiltrating lymphocytes. They have benchmarked it against two established models and made the entire processing pipeline available as open source.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s44172-024-00246-9.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141803619","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-24DOI: 10.1038/s44172-024-00250-z
Sebastian Nilsson, David Sanned, Adrian Roth, Jinguo Sun, Edouard Berrocal, Mattias Richter, Andreas Ehn
Gliding arc plasmas, a versatile form of non-thermal plasma discharges, hold great promise for sustainable chemical conversion in electrified industrial applications. Their relatively high temperatures compared to other non-thermal plasmas, reactive species generation, and efficient energy transfer make them ideal for an energy-efficient society. However, plasma discharges are transient and complex 3D entities influenced by gas pressure, mixture, and power, posing challenges for in-situ measurements of chemical species and spatial dynamics. Here we demonstrate a combination of innovative approaches, providing a comprehensive view of discharges and their chemical surroundings by combining fluorescence lifetime imaging of hydroxyl (OH) radicals with optical emission 3D tomography. This reveals variations in OH radical distributions under different conditions and local variations in fluorescence quantum yield with high spatial resolution from a single laser shot. Our results and methodology offer a multidimensional platform for interdisciplinary research in plasma physics and chemistry. Sebastian Nilsson and colleagues capture the 3D structure of a gliding arc plasma and simultaneously apply a single shot fluorescence lifetime imaging from which the physical properties linked to the chemistry and molecular collision dynamics can be extracted.
滑动电弧等离子体是一种多用途的非热等离子体放电形式,在电气化工业应用的可持续化学转换方面大有可为。与其他非热等离子体相比,滑弧式等离子体具有相对较高的温度、反应性物种生成和高效的能量传递等特点,是实现高能效社会的理想选择。然而,等离子体放电是受气体压力、混合物和功率影响的瞬态复杂三维实体,给化学物种和空间动态的现场测量带来了挑战。在这里,我们展示了一种创新方法的组合,通过将羟基(OH)自由基的荧光寿命成像与光发射 3D 层析成像相结合,提供了放电及其周围化学环境的全面视图。这揭示了羟基自由基在不同条件下的分布变化以及荧光量子产率的局部变化,单次激光发射即可实现高空间分辨率。我们的成果和方法为等离子体物理和化学的跨学科研究提供了一个多维平台。Sebastian Nilsson 及其同事捕捉了滑弧等离子体的三维结构,并同时应用了单次荧光寿命成像,从中提取了与化学和分子碰撞动力学相关的物理特性。
{"title":"Holistic analysis of a gliding arc discharge using 3D tomography and single-shot fluorescence lifetime imaging","authors":"Sebastian Nilsson, David Sanned, Adrian Roth, Jinguo Sun, Edouard Berrocal, Mattias Richter, Andreas Ehn","doi":"10.1038/s44172-024-00250-z","DOIUrl":"10.1038/s44172-024-00250-z","url":null,"abstract":"Gliding arc plasmas, a versatile form of non-thermal plasma discharges, hold great promise for sustainable chemical conversion in electrified industrial applications. Their relatively high temperatures compared to other non-thermal plasmas, reactive species generation, and efficient energy transfer make them ideal for an energy-efficient society. However, plasma discharges are transient and complex 3D entities influenced by gas pressure, mixture, and power, posing challenges for in-situ measurements of chemical species and spatial dynamics. Here we demonstrate a combination of innovative approaches, providing a comprehensive view of discharges and their chemical surroundings by combining fluorescence lifetime imaging of hydroxyl (OH) radicals with optical emission 3D tomography. This reveals variations in OH radical distributions under different conditions and local variations in fluorescence quantum yield with high spatial resolution from a single laser shot. Our results and methodology offer a multidimensional platform for interdisciplinary research in plasma physics and chemistry. Sebastian Nilsson and colleagues capture the 3D structure of a gliding arc plasma and simultaneously apply a single shot fluorescence lifetime imaging from which the physical properties linked to the chemistry and molecular collision dynamics can be extracted.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s44172-024-00250-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141809719","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-23DOI: 10.1038/s44172-024-00235-y
Dmytro Vovchuk, Gilad Uziel, Andrey Machnev, Jurgis Porins, Vjaceslavs Bobrovs, Pavel Ginzburg
High-gain antennas are essential hardware devices, powering numerous daily applications, including distant point-to-point communications, safety radars, and many others. While a common approach to elevate gain is to enlarge an antenna aperture, highly resonant subwavelength structures can potentially grant high gain performances. The Chu-Harrington limit is a standard criterion to assess electrically small structures and those surpassing it are called superdirective. Supergain is obtained in a case when internal losses are mitigated, and an antenna is matched to radiation, though typically in a very narrow frequency band. Here we develop a concept of a spectrally overlapping resonant cascading, where tailored multipole hierarchy grants both high gain and sufficient operational bandwidth. Our architecture is based on a near-field coupled wire bundle. Genetic optimization, constraining both gain and bandwidth, is applied on a 24-dimensional space and predicts 8.81 dBi realized gain within a half-wavelength in a cube volume. The experimental gain is 8.22 dBi with 13% fractional bandwidth. The developed approach can be applied across other frequency bands, where miniaturization of wireless devices is highly demanded. Vovchuk et al. utilized generic algorithms to demonstrate a multi-wire superdirective antenna. The small device not only surpasses common directivity bounds but also demonstrates a superbandwidth, thus making it useful for wireless communications.
{"title":"Genetically synthesized supergain broadband wire-bundle antenna","authors":"Dmytro Vovchuk, Gilad Uziel, Andrey Machnev, Jurgis Porins, Vjaceslavs Bobrovs, Pavel Ginzburg","doi":"10.1038/s44172-024-00235-y","DOIUrl":"10.1038/s44172-024-00235-y","url":null,"abstract":"High-gain antennas are essential hardware devices, powering numerous daily applications, including distant point-to-point communications, safety radars, and many others. While a common approach to elevate gain is to enlarge an antenna aperture, highly resonant subwavelength structures can potentially grant high gain performances. The Chu-Harrington limit is a standard criterion to assess electrically small structures and those surpassing it are called superdirective. Supergain is obtained in a case when internal losses are mitigated, and an antenna is matched to radiation, though typically in a very narrow frequency band. Here we develop a concept of a spectrally overlapping resonant cascading, where tailored multipole hierarchy grants both high gain and sufficient operational bandwidth. Our architecture is based on a near-field coupled wire bundle. Genetic optimization, constraining both gain and bandwidth, is applied on a 24-dimensional space and predicts 8.81 dBi realized gain within a half-wavelength in a cube volume. The experimental gain is 8.22 dBi with 13% fractional bandwidth. The developed approach can be applied across other frequency bands, where miniaturization of wireless devices is highly demanded. Vovchuk et al. utilized generic algorithms to demonstrate a multi-wire superdirective antenna. The small device not only surpasses common directivity bounds but also demonstrates a superbandwidth, thus making it useful for wireless communications.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s44172-024-00235-y.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141968498","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Spiking neural network algorithms require fine-tuned neuromorphic hardware to increase their effectiveness. Such hardware, mainly digital, is typically built on mature silicon nodes. Future artificial intelligence applications will demand the execution of tasks with increasing complexity and over timescales spanning several decades. The multi-timescale requirements for certain tasks cannot be attained effectively enough through the existing silicon-based solutions. Indium-Gallium-Zinc-Oxide thin-film transistors can alleviate the timescale-related shortcomings of silicon platforms thanks to their bellow atto-ampere leakage currents. These small currents enable wide timescale ranges, far beyond what has been feasible through various emerging technologies. Here we have estimated and exploited these low leakage currents to create a multi-timescale neuron that integrates information spanning a range of 7 orders of magnitude and assessed its advantages in larger networks. The multi-timescale ability of this neuron can be utilized together with silicon to create hybrid spiking neural networks capable of effectively executing more complex tasks than their single-technology counterparts. Mauricio Velazquez Lopez and colleagues fabricate a neuromorphic node with a response time that spans a range of 7 orders of magnitude. Their technology is compatible with complementary metal-oxide semiconductors, which makes it suitable for a variety of machine learning tasks.
{"title":"A tunable multi-timescale Indium-Gallium-Zinc-Oxide thin-film transistor neuron towards hybrid solutions for spiking neuromorphic applications","authors":"Mauricio Velazquez Lopez, Bernabe Linares-Barranco, Jua Lee, Hamidreza Erfanijazi, Alberto Patino-Saucedo, Manolis Sifalakis, Francky Catthoor, Kris Myny","doi":"10.1038/s44172-024-00248-7","DOIUrl":"10.1038/s44172-024-00248-7","url":null,"abstract":"Spiking neural network algorithms require fine-tuned neuromorphic hardware to increase their effectiveness. Such hardware, mainly digital, is typically built on mature silicon nodes. Future artificial intelligence applications will demand the execution of tasks with increasing complexity and over timescales spanning several decades. The multi-timescale requirements for certain tasks cannot be attained effectively enough through the existing silicon-based solutions. Indium-Gallium-Zinc-Oxide thin-film transistors can alleviate the timescale-related shortcomings of silicon platforms thanks to their bellow atto-ampere leakage currents. These small currents enable wide timescale ranges, far beyond what has been feasible through various emerging technologies. Here we have estimated and exploited these low leakage currents to create a multi-timescale neuron that integrates information spanning a range of 7 orders of magnitude and assessed its advantages in larger networks. The multi-timescale ability of this neuron can be utilized together with silicon to create hybrid spiking neural networks capable of effectively executing more complex tasks than their single-technology counterparts. Mauricio Velazquez Lopez and colleagues fabricate a neuromorphic node with a response time that spans a range of 7 orders of magnitude. Their technology is compatible with complementary metal-oxide semiconductors, which makes it suitable for a variety of machine learning tasks.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s44172-024-00248-7.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141811471","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-19DOI: 10.1038/s44172-024-00247-8
Priya Paulachan, René Hammer, Joerg Siegert, Ingo Wiesler, Roland Brunner
More than Moore technology is driving semiconductor devices towards higher complexity and further miniaturization. Device miniaturization strongly impacts failure analysis (FA), since it triggers the need for non-destructive approaches with high resolution in combination with cost and time efficient execution. Conventional scanning acoustic microscopy (SAM) is an indispensable tool for failure analysis in the semiconductor industry, however resolution and penetration capabilities are strongly limited by the transducer frequency. In this work, we conduct an acoustic interferometry approach, based on a SAM-setup utilizing 100 MHz lenses and enabling not only sufficient penetration depth but also high resolution for efficient in-line FA of Through Silicon Vias (TSVs). Accompanied elastodynamic finite integration technique-based simulations, provide an in-depth understanding concerning the acoustic wave excitation and propagation. We show that the controlled excitation of surface acoustic waves extends the contingency towards the detection of nm-sized cracks, an essential accomplishment for modern FA of 3D-integration technologies. Dr Roland Brunner and colleagues demonstrate how acoustic interferometry can be used to conduct a non-destructive and high-resolution failure analysis of through-silicon vias. They analyse the detection of nanometre-scale cracks and discuss how the opening angle of the acoustic lens impacts on performance.
超越摩尔定律的技术正推动半导体器件向着更高的复杂性和进一步微型化的方向发展。器件微型化对故障分析(FA)产生了强烈的影响,因为它引发了对高分辨率、低成本、高效率的非破坏性方法的需求。传统的扫描声学显微镜(SAM)是半导体行业故障分析不可或缺的工具,但其分辨率和穿透能力受到换能器频率的严重限制。在这项工作中,我们基于利用 100 MHz 镜头的 SAM 设置,采用声学干涉测量方法,不仅能获得足够的穿透深度,还能获得高分辨率,从而对硅通孔(TSV)进行有效的在线故障分析。伴随着基于弹性动力学有限积分技术的模拟,我们对声波的激发和传播有了深入的了解。我们的研究表明,表面声波的受控激发扩展了对纳米级裂纹的检测能力,这是现代三维集成技术 FA 的一项重要成就。罗兰-布鲁纳(Roland Brunner)博士及其同事展示了如何利用声学干涉测量法对硅通孔进行非破坏性和高分辨率的故障分析。他们分析了纳米级裂纹的检测,并讨论了声透镜的开口角度对性能的影响。
{"title":"Fast in-line failure analysis of sub-micron-sized cracks in 3D interconnect technologies utilizing acoustic interferometry","authors":"Priya Paulachan, René Hammer, Joerg Siegert, Ingo Wiesler, Roland Brunner","doi":"10.1038/s44172-024-00247-8","DOIUrl":"10.1038/s44172-024-00247-8","url":null,"abstract":"More than Moore technology is driving semiconductor devices towards higher complexity and further miniaturization. Device miniaturization strongly impacts failure analysis (FA), since it triggers the need for non-destructive approaches with high resolution in combination with cost and time efficient execution. Conventional scanning acoustic microscopy (SAM) is an indispensable tool for failure analysis in the semiconductor industry, however resolution and penetration capabilities are strongly limited by the transducer frequency. In this work, we conduct an acoustic interferometry approach, based on a SAM-setup utilizing 100 MHz lenses and enabling not only sufficient penetration depth but also high resolution for efficient in-line FA of Through Silicon Vias (TSVs). Accompanied elastodynamic finite integration technique-based simulations, provide an in-depth understanding concerning the acoustic wave excitation and propagation. We show that the controlled excitation of surface acoustic waves extends the contingency towards the detection of nm-sized cracks, an essential accomplishment for modern FA of 3D-integration technologies. Dr Roland Brunner and colleagues demonstrate how acoustic interferometry can be used to conduct a non-destructive and high-resolution failure analysis of through-silicon vias. They analyse the detection of nanometre-scale cracks and discuss how the opening angle of the acoustic lens impacts on performance.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s44172-024-00247-8.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141736853","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-15DOI: 10.1038/s44172-024-00242-z
Bahadır Utku Kesgin, Uğur Teğin
Machine learning studies need colossal power to process massive datasets and train neural networks to reach high accuracies, which have become gradually unsustainable. Limited by the von Neumann bottleneck, current computing architectures and methods fuel this high power consumption. Here, we present an analog computing method that harnesses chaotic nonlinear attractors to perform machine learning tasks with low power consumption. Inspired by neuromorphic computing, our model is a programmable, versatile, and generalized platform for machine learning tasks. Our mode provides exceptional performance in clustering by utilizing chaotic attractors’ nonlinear mapping and sensitivity to initial conditions. When deployed as a simple analog device, it only requires milliwatt-scale power levels while being on par with current machine learning techniques. We demonstrate low errors and high accuracies with our model for regression and classification-based learning tasks. Bahadır Utku Kesgin and Uğur Teğin propose using a Lorenz attractor as a nonlinear transfer function for neural network nodes. They design a power-efficient electrical circuit and use them for regression and classification test tasks.
机器学习研究需要巨大的能量来处理海量数据集和训练神经网络,以达到较高的精确度,这已逐渐变得难以为继。受限于冯-诺依曼瓶颈,当前的计算架构和方法助长了这种高能耗。在此,我们提出一种模拟计算方法,利用混沌非线性吸引子以低功耗执行机器学习任务。受神经形态计算的启发,我们的模式是一个可编程、多功能和通用的机器学习任务平台。我们的模式利用混沌吸引子的非线性映射和对初始条件的敏感性,在聚类方面提供了卓越的性能。当作为一个简单的模拟设备部署时,它只需要毫瓦级的功率,同时与当前的机器学习技术相当。我们展示了我们的模型在回归和分类学习任务中的低误差和高准确度。Bahadır Utku Kesgin 和 Uğur Teğin 建议使用洛伦兹吸引子作为神经网络节点的非线性传递函数。他们设计了一种高能效电路,并将其用于回归和分类测试任务。
{"title":"Implementing the analogous neural network using chaotic strange attractors","authors":"Bahadır Utku Kesgin, Uğur Teğin","doi":"10.1038/s44172-024-00242-z","DOIUrl":"10.1038/s44172-024-00242-z","url":null,"abstract":"Machine learning studies need colossal power to process massive datasets and train neural networks to reach high accuracies, which have become gradually unsustainable. Limited by the von Neumann bottleneck, current computing architectures and methods fuel this high power consumption. Here, we present an analog computing method that harnesses chaotic nonlinear attractors to perform machine learning tasks with low power consumption. Inspired by neuromorphic computing, our model is a programmable, versatile, and generalized platform for machine learning tasks. Our mode provides exceptional performance in clustering by utilizing chaotic attractors’ nonlinear mapping and sensitivity to initial conditions. When deployed as a simple analog device, it only requires milliwatt-scale power levels while being on par with current machine learning techniques. We demonstrate low errors and high accuracies with our model for regression and classification-based learning tasks. Bahadır Utku Kesgin and Uğur Teğin propose using a Lorenz attractor as a nonlinear transfer function for neural network nodes. They design a power-efficient electrical circuit and use them for regression and classification test tasks.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s44172-024-00242-z.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141624601","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-07-12DOI: 10.1038/s44172-024-00241-0
Mohammad Askari, Michele Benciolini, Hoang-Vu Phan, William Stewart, Auke J. Ijspeert, Dario Floreano
Perching with winged Unmanned Aerial Vehicles has often been solved by means of complex control or intricate appendages. Here, we present a method that relies on passive wing morphing for crash-landing on trees and other types of vertical poles. Inspired by the adaptability of animals’ and bats’ limbs in gripping and holding onto trees, we design dual-purpose wings that enable both aerial gliding and perching on poles. With an upturned nose design, the robot can passively reorient from horizontal flight to vertical upon a head-on crash with a pole, followed by hugging with its wings to perch. We characterize the performance of reorientation and perching in terms of impact speed and angle, pole material, and size. The robot robustly reorients at impact angles above 15° and speeds of 3 m ⋅ s−1 to 9 m ⋅ s−1, and can hold onto various pole types larger than 28% of its wingspan in diameter. We demonstrate crash-perching on tree trunks with an overall success rate of 73%. The method opens up new possibilities for the use of aerial robots in applications such as inspection, maintenance, and biodiversity conservation. Mohammad Askari and colleagues report a strategy for Unmanned Aerial Vehicles to perch on vertical poles and trees upon crash landing. An upturned nose passively reorients the robot, while dual-purpose wings secure the robot using an enveloping grasp, not unlike a hug.
{"title":"Crash-perching on vertical poles with a hugging-wing robot","authors":"Mohammad Askari, Michele Benciolini, Hoang-Vu Phan, William Stewart, Auke J. Ijspeert, Dario Floreano","doi":"10.1038/s44172-024-00241-0","DOIUrl":"10.1038/s44172-024-00241-0","url":null,"abstract":"Perching with winged Unmanned Aerial Vehicles has often been solved by means of complex control or intricate appendages. Here, we present a method that relies on passive wing morphing for crash-landing on trees and other types of vertical poles. Inspired by the adaptability of animals’ and bats’ limbs in gripping and holding onto trees, we design dual-purpose wings that enable both aerial gliding and perching on poles. With an upturned nose design, the robot can passively reorient from horizontal flight to vertical upon a head-on crash with a pole, followed by hugging with its wings to perch. We characterize the performance of reorientation and perching in terms of impact speed and angle, pole material, and size. The robot robustly reorients at impact angles above 15° and speeds of 3 m ⋅ s−1 to 9 m ⋅ s−1, and can hold onto various pole types larger than 28% of its wingspan in diameter. We demonstrate crash-perching on tree trunks with an overall success rate of 73%. The method opens up new possibilities for the use of aerial robots in applications such as inspection, maintenance, and biodiversity conservation. Mohammad Askari and colleagues report a strategy for Unmanned Aerial Vehicles to perch on vertical poles and trees upon crash landing. An upturned nose passively reorients the robot, while dual-purpose wings secure the robot using an enveloping grasp, not unlike a hug.","PeriodicalId":72644,"journal":{"name":"Communications engineering","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.nature.com/articles/s44172-024-00241-0.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141624598","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}