A5038937805- Advanced Memory and Neural Computing
- Ferroelectric and Negative Capacitance Devices
- Neural Networks and Reservoir Computing
- Microwave Engineering and Waveguides
- Electromagnetic Compatibility and Noise Suppression
- Stochastic Gradient Optimization Techniques
- Modular Robots and Swarm Intelligence
- Simulation Techniques and Applications
- Advanced Database Systems and Queries
- Advanced Antenna and Metasurface Technologies
University of California, Berkeley
2017-2023
Indian Institute of Technology Kanpur
2013
We outline a model of computing with high-dimensional (HD) vectors-where the dimensionality is in thousands. It built on ideas from traditional (symbolic) and artificial neural nets/deep learning, complements them probability theory, statistics, abstract algebra. Key properties HD include well-defined set arithmetic operations vectors, generality, scalability, robustness, fast ubiquitous parallel operation, making it possible to develop efficient algorithms for large-scale real-world tasks....
Brain-inspired hyperdimensional (HD) computing models cognition by exploiting properties of high dimensional statistics-high-dimensional vectors, instead working with numeric values used in contemporary processors. A fundamental weakness existing HD algorithms is that they require to use floating point order provide acceptable accuracy on realistic classification problems. However, significantly increases the computation cost. To address this issue, we proposed QuantHD, a novel framework for...
Hyper-dimensional Computing (HDC), a bio-inspired paradigm defined on random high-dimensional vectors, has emerged as promising IoT paradigm. It is known to provide competitive accuracy sequential prediction tasks with much smaller model size and training time compared conventional ML, well-suited for human-centric IoT. In the post-Moore scaling era, where increasing variability challenged traditional designers, its novel computing method based randomness can be leveraged continued...
In this paper, design principles for implementation of broadband T-type and Y-type power divider based on Substrate-Integrated Waveguide (SIW) technology has been presented. Different curves have developed simulation results are used to implement 1×2 dividers T-type. Finally, by cascading stages a 1×4 Power equal division is designed simulated. The proposed shows broad bandwidth (-20 dB) 18.3% insertion-loss Bandwidth (1%) 17.86%.
Summary form only given. Some of most compelling application domains the IoT and Swarm concepts relate to how humans interact with world around it cyberworld beyond. While proliferation communication data processing devices has profoundly altered our interaction patterns, little been changed in way we process inputs (sensory) outputs (actuation). The combination (Swarms) wearable offers potential for changing all this, opening door true human augmentation. epitome this would be a direct...
Hyper-Dimensional Computing (HDC), a nanoscalable learning paradigm for low-energy predictions and lightweight models, has seen surge in interest from the hardware accelerator community. Its statistical distributed data representation leads to highly-efficient classifiers with inherent robustness errors. A digital, 28nm CMOS chip, representing first programmable HDC biosignal processor, achieves 25.6 nJ/pred. on leading EMG gesture recognition dataset. Measurements confirm high of HDC: 47%...
Hyper-Dimensional Computing (HDC), a promising nano-scalable paradigm for low-energy predictions and lightweight learned models, has seen surge of interest from the hardware accelerator community. However, classical single-bit per vector element approach HDC seldom achieves higher classification accuracy than multi-bit alternatives, is inadequate to support rapidly growing application space. A great challenge negotiate enormous increase in logic vis-a-vis hardware. Key minimizing this cost...