Review
Chemistry, Multidisciplinary
Cefe Lopez
Summary: Artificial intelligence is growing stronger, and materials science can benefit from it by using machine learning techniques to design and optimize new materials, systems, and processes. Understanding how machine learning can enhance the conception of advanced materials is crucial for future materials scientists. This review provides knowledge of computation and machine learning methods, and explores the impact of these techniques on the development of new advanced materials.
ADVANCED MATERIALS
(2023)
Article
Nanoscience & Nanotechnology
Karl Berggren, Qiangfei Xia, Konstantin K. Likharev, Dmitri B. Strukov, Hao Jiang, Thomas Mikolajick, Damien Querlioz, Martin Salinga, John R. Erickson, Shuang Pi, Feng Xiong, Peng Lin, Can Li, Yu Chen, Shisheng Xiong, Brian D. Hoskins, Matthew W. Daniels, Advait Madhavan, James A. Liddle, Jabez J. McClelland, Yuchao Yang, Jennifer Rupp, Stephen S. Nonnenmann, Kwang-Ting Cheng, Nanbo Gong, Miguel Angel Lastras-Montano, A. Alec Talin, Alberto Salleo, Bhavin J. Shastri, Thomas Ferreira de Lima, Paul Prucnal, Alexander N. Tait, Yichen Shen, Huaiyu Meng, Charles Roques-Carmes, Zengguang Cheng, Harish Bhaskaran, Deep Jariwala, Han Wang, Jeffrey M. Shainline, Kenneth Segall, J. Joshua Yang, Kaushik Roy, Suman Datta, Arijit Raychowdhury
Summary: Recent progress in artificial intelligence is primarily attributed to the rapid development of machine learning, but the performance and energy efficiency of hardware systems set fundamental limits on machine learning capabilities. Data-centric computing requires a revolution in hardware systems, with new hardware platforms offering hope for future computing with improved throughput and energy efficiency. However, challenges such as materials selection, device optimization, circuit fabrication, and system integration must be addressed in building such systems.
Article
Automation & Control Systems
Dovydas Joksas, AbdulAziz AlMutairi, Oscar Lee, Murat Cubukcu, Antonio Lombardo, Hidekazu Kurebayashi, Anthony J. Kenyon, Adnan Mehonic
Summary: This article discusses the gap between computing demands and existing technologies in a data-driven economy, as well as the limitations of energy efficiency. Three approaches that may play a crucial role in future computing systems are presented, and their impact on conventional computers and emerging computing paradigms is explained.
ADVANCED INTELLIGENT SYSTEMS
(2022)
Article
Chemistry, Multidisciplinary
Revannath Dnyandeo Nikam, Jongwon Lee, Wooseok Choi, Dongmin Kim, Hyunsang Hwang
Summary: This study proposes the use of monolayer graphene as a low-power heating source in O-ECRAM to enhance oxygen-ion transport for improved learning accuracy and conductance tuning, achieving better performance in artificial synapses.
Article
Chemistry, Multidisciplinary
Revannath Dnyandeo Nikam, Jongwon Lee, Wooseok Choi, Dongmin Kim, Hyunsang Hwang
Summary: By using a monolayer graphene as a heating source, the migration of ions in O-ECRAM can be increased, enabling learning and storage in neural networks. This method features long-term retention, high stability, and the ability to store analog states. These findings demonstrate the application of 2D materials as heating elements in artificial synapse chips to accelerate neuromorphic computation.
Article
Automation & Control Systems
Marlon Becker, Jan Riegelmeyer, Maximilian David Seyfried, Bart Jan Ravoo, Carsten Schuck, Benjamin Risse
Summary: This research explores the use of photoswitchable chemical compounds as activation functions in optical neural networks (ONNs). By manipulating the activation behavior through photo-induced isomerization, the nonlinearity of these compounds can be controlled. This enables the implementation of classification tasks in neural networks and opens up possibilities for explaining intelligent behavior in these networks.
ADVANCED INTELLIGENT SYSTEMS
(2023)
Article
Materials Science, Multidisciplinary
Bai Sun, Tao Guo, Guangdong Zhou, Shubham Ranjan, Yixuan Jiao, Lan Wei, Y. Norman Zhou, A. Yimin Wu
Summary: Synaptic devices, such as synaptic memristors and synaptic transistors, have the potential to revolutionize traditional data storage and computing methods by enabling high-performance super-parallel computing through neuromorphic computing. This review focuses on the applications of synaptic devices in artificial intelligence, covering topics such as circuit theory, material selection, and future applications.
MATERIALS TODAY PHYSICS
(2021)
Article
Automation & Control Systems
Meirong Huang, Zechen Li, Hongwei Zhu
Summary: This review discusses two simulation approaches in artificial intelligence, machine learning (ML) and neuromorphic computing (NC), which emulate the working principles of biological brains using artificial neurons/synapses. The authors provide a general description of ML methods and the concept of artificial synapses (ASs) derived from biological synapses, along with their relationships. The application of ML in graphene and its derivatives and composites is summarized in terms of properties prediction, structure recognition, inverse design, and task recognition. Additionally, recent progress in graphene-based synaptic transistors and memristors is briefly introduced. The article concludes by presenting the main challenges and prospects of incorporating graphene into ML and ASs.
ADVANCED INTELLIGENT SYSTEMS
(2022)
Article
Materials Science, Multidisciplinary
Sahra Afshari, Sritharini Radhakrishnan, Jing Xie, Mirembe Musisi-Nkambwe, Jian Meng, Wangxin He, Jae-sun Seo, Ivan Sanchez Esqueda
Summary: This study presents the hardware implementation of analog dot-product operation on arrays of two-dimensional (2D) hexagonal boron nitride (h-BN) memristors. It extends beyond previous work to the application of analog neural network accelerators based on 2D memristor arrays. The wafer-level fabrication of the memristor arrays is achieved through large-area transfer of CVD-grown few-layer h-BN films. The results show promising resistive switching characteristics, performance of dot-product computation, and successful demonstration of logistic regression in h-BN memristors, indicating an important step towards the integration of 2D materials for next-generation neuromorphic computing systems.
Review
Chemistry, Analytical
Alexandra Parichenko, Shirong Huang, Jinbo Pang, Bergoi Ibarlucea, Gianaurelio Cuniberti
Summary: Inspired by biological noses, e-noses imitate them by using gas sensor arrays to detect and identify surrounding gases and volatile compounds. Two-dimensional materials have shown remarkable sensitivity at room temperature, addressing energy efficiency and sensitivity issues. This review highlights advancements in the development of e-noses, including transduction mechanisms and deposition methods for two-dimensional materials. Artificial intelligence tools are discussed for smart data analysis to overcome selectivity limitations.
TRAC-TRENDS IN ANALYTICAL CHEMISTRY
(2023)
Review
Neurosciences
Dmitry Ivanov, Aleksandr Chezhegov, Mikhail Kiselev, Andrey Grunin, Denis Larionov
Summary: This article discusses the limitations of modern artificial intelligence systems based on von Neumann architecture and classical neural networks compared to the mammalian brain, and ways to overcome them through neuromorphic AI projects. It also presents the principle of classifying neuromorphic AI systems based on the brain features they use, and discusses the prospects of using new memristor element base in neuromorphic applications.
FRONTIERS IN NEUROSCIENCE
(2022)
Review
Chemistry, Multidisciplinary
Geonyeop Lee, Ji-Hwan Baek, Fan Ren, Stephen J. Pearton, Gwan-Hyoung Lee, Jihyun Kim
Summary: Neuromorphic systems, which mimic neural functionalities of a human brain using artificial synapses and neurons, have advantages of high energy efficiency and fast computing speed. 2D materials, with unique surface properties and excellent crystallinity, have emerged as promising candidates for neuromorphic computing hardware due to uncontrollable defects in bulk material-based devices.
Article
Computer Science, Theory & Methods
Haochen Hua, Yutong Li, Tonghe Wang, Nanqing Dong, Wei Li, Junwei Cao
Summary: In recent years, the widespread popularity of the Internet of Things (IoT) has greatly promoted the development of Artificial Intelligence (AI). However, the traditional cloud computing model may face difficulties in independently handling the massive data generated by IoT. In response, the new computing model of Edge Computing (EC) has gained extensive attention. Scholars have found that traditional methods have limitations in enhancing the performance of EC, leading to the exploration of AI as a solution. This article serves as a guide to explore new research ideas in optimizing EC using AI and applying AI to other fields under the EC architecture.
ACM COMPUTING SURVEYS
(2023)
Review
Physics, Multidisciplinary
Chaoran Huang, Volker J. Sorger, Mario Miscuglio, Mohammed Al-Qadasi, Avilash Mukherjee, Lutz Lampe, Mitchell Nichols, Alexander N. Tait, Thomas Ferreira de Lima, Bicky A. Marquez, Jiahui Wang, Lukas Chrostowski, Mable P. Fok, Daniel Brunner, Shanhui Fan, Sudip Shekhar, Paul R. Prucnal, Bhavin J. Shastri
Summary: Neuromorphic engineering using photonics allows for high-speed and energy-efficient artificial intelligence and neuromorphic computing applications. It mimics neurons and synapses in the brain to achieve distributed and parallel processing.
ADVANCES IN PHYSICS-X
(2022)
Article
Mathematical & Computational Biology
Nicolas Skatchkovsky, Hyeryung Jang, Osvaldo Simeone
Summary: This paper takes steps towards designing neuromorphic systems that can adapt to changing learning tasks and provide accurate uncertainty quantification estimates. By deriving online learning rules within a Bayesian continual learning framework, the proposed method updates the distribution parameters of synaptic weights based on prior knowledge and observed data. Experimental results demonstrate the advantages of Bayesian learning over frequentist learning in terms of adaptation and uncertainty quantification.
FRONTIERS IN COMPUTATIONAL NEUROSCIENCE
(2022)
Article
Chemistry, Multidisciplinary
David Lam, Dmitry Lebedev, Lidia Kuo, Vinod K. Sangwan, Beata M. Szydlowska, Filippo Ferraresi, Aljoscha Soll, Zdenek Sofer, Mark C. Hersam
Summary: In this study, alpha-RuCl3 nanosheets were successfully fabricated through liquid-phase exfoliation (LPE) and assembled into electrically conductive large-area thin films. The crystalline integrity of LPE alpha-RuCl3 nanosheets was confirmed, and the physical properties, including phase transition and magnetism, were investigated. Additionally, large-area Mott insulator photodetectors operating in the infrared region were demonstrated.
Article
Chemistry, Multidisciplinary
Lidia Kuo, Vinod K. Sangwan, Sonal Rangnekar, Ting-Ching Chu, David Lam, Zhehao Zhu, Lee J. Richter, Ruipeng Li, Beata M. Szydlowska, Julia R. Downing, Benjamin J. Luijten, Lincoln J. Lauhon, Mark C. Hersam
Summary: Aerosol-jet printing technology is utilized to achieve highly responsive photodetectors with superior optoelectronic properties, surpassing previously reported devices by three orders of magnitude. Properly designed processing and ink formulations enable the processing and application of ultrathin MoS2 nanosheets, providing a new approach for scalable manufacturing in mechanically flexible optoelectronics.
ADVANCED MATERIALS
(2022)
Article
Chemistry, Multidisciplinary
Dmitry Lebedev, Jonathan Tyler Gish, Ethan Skyler Garvey, Teodor Kosev Stanev, Junhwan Choi, Leonidas Georgopoulos, Thomas Wei Song, Hong Youl Park, Kenji Watanabe, Takashi Taniguchi, Nathaniel Patrick Stern, Vinod Kumar Sangwan, Mark Christopher Hersam
Summary: 2D magnetic materials have promising applications in quantum and spintronic devices. 2D antiferromagnetic materials are of interest due to their insensitivity to external magnetic fields and faster switching speeds compared to 2D ferromagnets. However, their lack of macroscopic magnetization hampers the detection and control of antiferromagnetic order, emphasizing the need for magneto-electrical measurements.
ADVANCED FUNCTIONAL MATERIALS
(2023)
Article
Engineering, Electrical & Electronic
Priyesh Shukla, Shamma Nasrin, Nastaran Darabi, Wilfred Gomes, Amit Ranjan Trivedi
Summary: We propose MC-CIM, a compute-in-memory framework that enables robust and low power Bayesian edge intelligence. By using the Monte Carlo Dropout approximation of Bayesian DNN, we enhance the computational efficiency of the method. We demonstrate the effectiveness of the proposed framework in applications such as MNIST character recognition and visual odometry of autonomous drones.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS
(2023)
Article
Physics, Applied
John M. Cain, Xiaodong Yan, Stephanie E. Liu, Justin H. Qian, Thomas T. Zeng, Vinod K. Sangwan, Mark C. Hersam, Stanley S. Chou, Tzu-Ming Lu
Summary: Sulfur-deficient polycrystalline two-dimensional molybdenum disulfide memtransistors demonstrate gate-tunable memristive switching for new memory operations and neuromorphic computing paradigms. The influence of grain boundaries, sulfur vacancies, and surface interactions on defect-related kinetics that produces memristive switching is studied using current transient measurements. It is observed that adsorbed water molecules alter the resistive switching kinetics by suppressing the electronic trap-filling processes.
APPLIED PHYSICS LETTERS
(2023)
Article
Chemistry, Physical
Yujin Lee, Kangsik Kim, Zonghoon Lee, Hong-Sub Lee, Han-Bo-Ram Lee, Woo-Hee Kim, Il-Kwon Oh, Hyungjun Kim
Summary: In this study, Dy incorporation was used to stabilize HfO2 films and increase the grain size, which resulted in a reduction of leakage current density and an increase in breakdown strength. The properties of Dy-doped HfO2 thin films were characterized using various analysis techniques. The phase transformation of HfO2 films from different planes to a main m(-111) plane was observed through X-ray diffraction, indicating the role of Dy in stabilizing the film. The increase in grain size due to Dy incorporation was confirmed by electron microscopy.
CHEMISTRY OF MATERIALS
(2023)
Article
Chemistry, Inorganic & Nuclear
Abhishek Rawat, Laura Clark, Chuzhong Zhang, John Cavin, Vinod K. Sangwan, Peter S. Toth, Csaba Janaky, Riddhi Ananth, Elise Goldfine, Michael J. Bedzyk, Emily A. Weiss, James M. Rondinelli, Mark C. Hersam, Efstathios I. Meletis, Krishnan Rajeshwar
Summary: Magnesium vanadate (MgV2O6) and its alloys with copper vanadate were synthesized. Experimental techniques and density functional theory (DFT) simulations were used to confirm phase purity and solid solution formation. The alloy composition showed systematic variation in optical bandgap modification and band alignment, indicating n-type semiconductor behavior.
INORGANIC CHEMISTRY
(2023)
Article
Energy & Fuels
Chaeseon Hong, Minjae Kim, Jin-Gyu Lee, Qingyi Shao, Hong-Sub Lee, Hyung-Ho Park
Summary: Si-ZnO thin-film transistors were fabricated via atomic layer deposition (ALD) and then annealed in oxygen, resulting in improved electrical characteristics and lower power consumption.
INTERNATIONAL JOURNAL OF ENERGY RESEARCH
(2023)
Article
Chemistry, Multidisciplinary
David W. Burke, Raghunath R. Dasari, Vinod K. Sangwan, Alexander K. Oanta, Zoheb Hirani, Chloe E. Pelkowski, Yongjian Tang, Ruofan Li, Daniel C. Ralph, Mark C. Hersam, Stephen Barlow, Seth R. Marder, William R. Dichtel
Summary: Researchers synthesized a new controllable oriented two-dimensional covalent organic framework (COF), named OTPA-BDT, via transimination reactions between benzophenone-imine-protected azatriangulenes (OTPA) and benzodithiophene dialdehydes (BDT). The material exhibits tunable semiconductor properties with a Dirac-cone-like band structure, making it an ideal candidate for next-generation flexible electronics. After orientation and doping, the OTPA-BDT COF films demonstrate high electrical conductivities of up to 1.2 x 10(-1) S cm(-1), the highest reported for imine-linked 2D COFs to date.
JOURNAL OF THE AMERICAN CHEMICAL SOCIETY
(2023)
Article
Engineering, Electrical & Electronic
Seung-Hyeon Kang, Seonguk Yang, Donghyun Lee, Sungkyu Kim, Joonki Suh, Hong-Sub Lee
Summary: This study presents a highly tunable synaptic weight update based on a multiterminal memtransistor device as a solution for nonlinear synaptic operations and crosstalk issues in CAA memristors. The memtransistor device exhibits a significant tunable weight update property at the gate knob and can function as a selector in the CAA and improve the linearity of the potentiation and depression curves. The tunable synaptic weight update functions provide advantageous features for accuracy and crosstalk issues in neuromorphic hardware.
ACS APPLIED ELECTRONIC MATERIALS
(2023)
Article
Chemistry, Multidisciplinary
Sonal V. V. Rangnekar, Vinod K. K. Sangwan, Mengru Jin, Maryam Khalaj, Beata M. M. Szydlowska, Anushka Dasgupta, Lidia Kuo, Heather E. E. Kurtz, Tobin J. J. Marks, Mark C. C. Hersam
Summary: Electroluminescence from molybdenum disulfide (MoS2) nanosheet films has been demonstrated by using a monolayer-rich MoS2 ink produced by electrochemical intercalation and megasonic exfoliation. The megasonicated MoS2 films retain their direct bandgap character in electrically percolating thin films even following multistep solution processing. This work establishes megasonicated MoS2 inks as an additive manufacturing platform for flexible, patterned, and miniaturized light sources.
Article
Energy & Fuels
Chaeseon Hong, Minjae Kim, Byeong-Min Lim, Sunil Moon, Keonwook Kang, Ioannis Kymissis, Hong-Sub Lee, Hyung-Ho Park
Summary: In this study, cation and anion codoped ZnO thin films were prepared through atomic layer deposition, and the effects of different anion doping concentrations on the film properties were investigated. The codoped ZnO films showed reduced oxygen vacancy due to the codoping effect. However, the electrical and optical characteristics of the films exhibited contrasting tendencies depending on the type of anion element. These differences were attributed to the correlation effects between Si doping, reduced interfacial trap density, and anion doping, oxygen vacancy passivation. The behavior of each anion also varied in terms of charge carrier concentration.
INTERNATIONAL JOURNAL OF ENERGY RESEARCH
(2023)
Article
Materials Science, Multidisciplinary
Elisa Trippodo, Vincenzo Campisciano, Liang-Wen Feng, Yao Chen, Wei Huang, Joaquin M. Alzola, Ding Zheng, Vinod K. Sangwan, Mark C. Hersam, Michael R. Wasielewski, Bruno Pignataro, Francesco Giacalone, Tobin J. Marks, Antonio Facchetti
Summary: Organic solar cells based on donor-acceptor blends have shown great potential in efficiency improvement, but performance degradation is still a major obstacle for commercialization. Ternary solar cells with fullerene acceptors as the third component exhibit enhanced stability with over 90% retention of initial power conversion efficiency even after 6 months of storage. This improved stability is attributed to a more robust blend morphology that reduces charge recombination in the aging process.
JOURNAL OF MATERIALS CHEMISTRY C
(2023)
Article
Chemistry, Multidisciplinary
Nathan P. Bradshaw, Zoheb Hirani, Lidia Kuo, Siyang Li, Nicholas X. Williams, Vinod K. Sangwan, Lindsay E. Chaney, Austin M. Evans, William R. Dichtel, Mark C. Hersam
Summary: This paper presents a method of aerosol jet printing of COFs with micron-scale resolution using a pre-synthesized colloidal ink. The ink formulation enables the integration of COFs with other colloidal nanomaterials to form printable nanocomposite films, resulting in high-sensitivity temperature sensors.
ADVANCED MATERIALS
(2023)
Article
Computer Science, Information Systems
Shamma Nasrin, Ahish Shylendra, Nastaran Darabi, Theja Tulabandhula, Wilfred Gomes, Ankush Chakrabarty, Amit Ranjan Trivedi
Summary: This work proposes a novel ENOS approach to address the energy-accuracy trade-offs of a DNN accelerator. ENOS allows for optimal layer-wise integration of inference operators with optimal precision to maintain high prediction accuracy and energy efficiency. Experimental results show that ENOS can improve accuracy by 10-20% under the same energy budget.