A5040105498- Advanced Memory and Neural Computing
- Ferroelectric and Negative Capacitance Devices
- Semiconductor materials and devices
- Neuroscience and Neural Engineering
- Neural dynamics and brain function
- CCD and CMOS Imaging Sensors
- Neural Networks and Reservoir Computing
- Transition Metal Oxide Nanomaterials
- Physical Unclonable Functions (PUFs) and Hardware Security
- Magnetic properties of thin films
- Photoreceptor and optogenetics research
- Electronic and Structural Properties of Oxides
- Neural Networks and Applications
- Multiferroics and related materials
- Magnetic and transport properties of perovskites and related materials
- Machine Learning and ELM
- Graphene research and applications
- Photonic and Optical Devices
- Advancements in Semiconductor Devices and Circuit Design
- Perovskite Materials and Applications
- Conducting polymers and applications
- Quantum and electron transport phenomena
- Advanced Sensor and Energy Harvesting Materials
- Gas Sensing Nanomaterials and Sensors
- 2D Materials and Applications
Tsinghua University
2016-2025
Institute of Microelectronics
2015-2024
Center for Information Technology
2024
Shanghai University
2024
Beijing Information Science & Technology University
2024
Beijing Advanced Sciences and Innovation Center
2023
Collaborative Innovation Center of Advanced Microstructures
2023
Nanjing University
2023
University of Glasgow
2022
Beijing University of Posts and Telecommunications
2020
Abstract Conventional hardware platforms consume huge amount of energy for cognitive learning due to the data movement between processor and off-chip memory. Brain-inspired device technologies using analogue weight storage allow complete tasks more efficiently. Here we present an non-volatile resistive memory (an electronic synapse) with foundry friendly materials. The shows bidirectional continuous modulation behaviour. Grey-scale face classification is experimentally demonstrated...
Abstract Resistive switching (RS) is an interesting property shown by some materials systems that, especially during the last decade, has gained a lot of interest for fabrication electronic devices, with nonvolatile memories being those that have received most attention. The presence and quality RS phenomenon in system can be studied using different prototype cells, performing experiments, displaying figures merit, developing computational analyses. Therefore, real usefulness impact findings...
Abstract Realizing increasingly complex artificial intelligence (AI) functionalities directly on edge devices calls for unprecedented energy efficiency of hardware. Compute-in-memory (CIM) based resistive random-access memory (RRAM) 1 promises to meet such demand by storing AI model weights in dense, analogue and non-volatile RRAM devices, performing computation within RRAM, thus eliminating power-hungry data movement between separate compute 2–5 . Although recent studies have demonstrated...
A nociceptor is a critical and special receptor of sensory neuron that able to detect noxious stimulus provide rapid warning the central nervous system start motor response in human body humanoid robotics. It differs from other common receptors with its key features functions, including "no adaptation" "sensitization" phenomena. In this study, we propose experimentally demonstrate an artificial based on diffusive memristor dynamics for first time. Using nociceptor, further built alarm...
Abstract Reservoir computing is a highly efficient network for processing temporal signals due to its low training cost compared standard recurrent neural networks, and generating rich reservoir states critical in the hardware implementation. In this work, we report parallel dynamic memristor-based system by applying controllable mask process, which parameters, including state richness, feedback strength input scaling, can be tuned changing length range of signal. Our achieves word error...
Abstract Owing to their attractive application potentials in both non-volatile memory and unconventional computing, memristive devices have drawn substantial research attention the last decade. However, major roadblocks still remain device performance, especially concerning relatively large parameter variability limited cycling endurance. The response of active region within between switching cycles plays dominating role, yet microscopic details elusive. This Review summarizes recent...
Abstract Threshold switches with Ag or Cu active metal species are volatile memristors (also termed diffusive memristors) featuring spontaneous rupture of conduction channels. The temporal dynamics the conductance evolution is closely related to electrochemical and metals which could be modulated by electric field strength, biasing duration, temperature, so on. Microscopic pictures electron microscopy quantitative thermodynamics modeling examined give insights into underlying physics...
As artificial intelligence calls for novel energy-efficient hardware, neuromorphic computing systems based on analog resistive switching memory (RSM) devices have drawn great attention recently. Different from the well-studied binary RSMs, RSMs are featured by a continuous and controllable conductance-tuning ability thus capable of combining data storage at device level. Although significant research achievements been accomplished, there few works demonstrating large-scale systems. A major...
Abstract Memristors as electronic artificial synapses have attracted increasing attention in neuromorphic computing. Emulation of both “learning” and “forgetting” processes requires a bidirectional progressive adjustment memristor conductance, which is challenge for cutting‐edge intelligence. In this work, device with structure Ag/Zr 0.5 Hf O 2 :graphene oxide quantum dots/Ag presented the feature conductance tuning. The proposed adjusted through voltage pulse number, amplitude, width. A...
Experimental demonstration of resistive neural networks has been the recent focus hardware implementation neuromorphic computing. Capacitive networks, which call for novel building blocks, provide an alternative physical embodiment featuring a lower static power and better emulation functionalities. Here, we develop neuro-transistors by integrating dynamic pseudo-memcapacitors as gates transistors to produce electronic analogs soma axon neuron, with "leaky integrate-and-fire" dynamics...
Abstract Memory cells have always been an important element of information technology. With emerging technologies like big data and cloud computing, the scale complexity storage has reached unprecedented peak with a much higher requirement for memory As is well known, better mostly achieved by miniaturization. However, as size device reduced, series problems, such drain gate‐induced leakage, greatly hinder performance units. To meet increasing demands technology, novel high‐performance...
Non-volatile memory (NVM) based computing-in-memory (CIM) shows significant advantages in handling deep learning tasks for artificial intelligence (AI) applications. To overcome the decreasing cost effectiveness of transistor scaling and intrinsic inefficiency data-shuttling von-Neumann architecture, CIM is proposed to realize high-speed low-power system with parallel multiplication accumulation (MAC) computing [1] [2]. However, current demonstrations are mainly on single macro present...
On-chip implementation of large-scale neural networks with emerging synaptic devices is attractive but challenging, primarily due to the pre-mature analog properties today's resistive memory technologies. This work aims realize a network using available binary RRAM for image recognition. We propose methodology binarize parameters goal reducing precision weights and neurons 1-bit classification <;8-bit online training. experimentally demonstrate (BNN) on Tsinghua's 16 Mb macro chip fabricated...
The crossbar array architecture with resistive synaptic devices is attractive for on-chip implementation of weighted sum and weight update in the neuro-inspired learning algorithms. This paper discusses design challenges on scaling up size due to non-ideal device properties parasitics. Circuit-level mitigation strategies have been proposed minimize accuracy loss a large array. also peripheral circuits considerations architecture. Finally, circuit-level macro simulator developed explore...