Friday, April 20th 2018 at 16:30-18:00
Room 202, South Bldg, Academic Center for Computing and Media Studies
S-143H Remote Lecture Room
YRP Mobile Lab
Note: If you would like to attend from Yokosuka, let me know it in advance.
Dr. Keisuke Fujii
Physics Department, Graduate School of Science, Kyoto University
Recent theoretical and experimental progress on quantum computing
I will provide a pedagogical introduction of quantum computing, specifically its definition, how it is different from conventional probabilistic computation, and how it can solve certain problems much faster than conventional “classical” computers. Then, I will review recent progress on quantum computing from both theoretical and experimental aspects.
Friday, May 25th 2018 at 16:30-18:00
Room 202, South Bldg, Academic Center for Computing and Media Studies
S-143H Remote Lecture Room
YRP Mobile Lab
Note: If you would like to attend from Yokosuka, let me know it in advance.
Prof. Takehiro Sato
Department of Communications and Computer Engineering, Graduate School of Informatics, Kyoto University
Optical access network: History and recent topics
A passive optical network (PON), which is an access network system using optical fiber cables and splitters, has pushed forward the popularization of fiber-to-the-home (FTTH). Due to its cost-effectiveness, not only residential Internet access but also other services such as 5G mobile backhauls are expected to be provided in the next-generation PON. This presentation first introduces the history and basic mechanism of optical access networks. Then the recent research topics in optical access networks, such as flexible utilization of network resources by applying virtualization techniques, are presented.
Friday, June 15th 2018 at 16:30-18:00
Room 202, South Bldg, Academic Center for Computing and Media Studies
S-143H Remote Lecture Room
YRP Mobile Lab
Note: If you would like to attend from Yokosuka, let me know it in advance.
Dr. Shigeaki Okumura
FURUNO ELECTRIC CO., LTD.
Signal processing techniques for non-invasive radar and ultrasound-based human measurement and their applications
Recently, several non-invasive measurement techniques that measure outside and inside of the human body using radar and ultrasound-based devices have been reported. In this talk, we introduce the signal processing techniques and applications. Especially, we focus on adaptive signal processing techniques that suppress the interferences and extract the desired signals.
Friday, July 13th 2018 at 16:30-18:00
Room 202, South Bldg, Academic Center for Computing and Media Studies
S-143H Remote Lecture Room
YRP Mobile Lab
Note: If you would like to attend from Yokosuka, let me know it in advance.
Prof. Priyank Kalla
Electrical and Computer Engineering University of Utah, Salt Lake City, USA
Verification of Arithmetic Circuits Using Algebraic Geometry and Symbolic Computation
Until a few years ago, automatic formal verification of arithmetic datapath circuits was considered an unsolvable problem. It was indeed impossible to formally verify custom-designed arithmetic circuits, automatically, beyond even 16-bit datapath word-lengths. Computational techniques quickly encountered exponential space and time explosion of complexity. In the past few years, the state-of-the-art in this area has witnessed a significant leap in verification capacity, facilitating verification for up to 500-bit wordlengths. Algebraic Geometry has played a fundamental role in this success – helping us understand the nature of the problem, and enabling algorithmic implementations to exploit domain-specific knowledge for efficiency and scalability.
In this talk, I will describe our work on formal verification of datapath designs using techniques that lie at the cross-roads of commutative algebra, algebraic geometry and electronic design automation (EDA) techniques. We particularly exploit the theory and technology of Gröbner bases to verify datapath circuits. Algorithms in computational algebraic geometry exhibit very high complexity. However, datapath designs exhibit some form of structure and symmetry in their functions and implementations. Our work has shown that the powerful Gröbner basis reasoning helps us discover this “structure and symmetry,” which can be exploited to simplify the algorithms, data-structures and implementations – thus enabling scalability. In this talk, I will describe the verification context, the problem formulations, our discoveries, and results. I will conclude the talk with discussions on other significant problems that can be solved by analyzing word-level abstractions of hardware designs using algebraic geometry. The talk should be accessible to electrical and computer engineers, (theoretical) computer scientists as well as algebraists.
Friday, October 19th 2018 at 16:30-18:00
Room 202, South Bldg, Academic Center for Computing and Media Studies
S-143H Remote Lecture Room
YRP Mobile Lab
Note: If you would like to attend from Yokosuka, let me know it in advance.
Mr. Masamichi Kawarabayashi
Vice President, Strategy & Planning Division, Industrial Solution Business Unit, Renesas Electronics Corporation
Semiconductor industry direction and solution technology
Overview the semiconductor industry and the leading solution technologies, such as e-AI (embedded Artificial Intelligence) and DRP (Dynamic Reconfigurable Processor).
Friday, November 16th 2018 at 16:30-18:00
Room 202, South Bldg, Academic Center for Computing and Media Studies
S-143H Remote Lecture Room
YRP Mobile Lab
Note: If you would like to attend from Yokosuka, let me know it in advance.
Prof. Anwar Hasan
Dept. of Electrical and Computer Engineering, University of Waterloo, Canada
Low cost hardware/software algorithms for arithmetic in extended finite fields
Finite fields (a.k.a. Galois fields) play an important role in areas such as cryptography, error control coding and the testing of integrated circuits. Some cryptosystems use very large finite fields. Performance of such cryptosystems depends on the efficiency of arithmetic operations over the finite fields used. Arithmetic operations over large finite fields are however not well supported by today’s general purpose processors. This talk critically looks into a number of low cost hardware/software algorithms for arithmetic over extended finite fields and brings up an unconventional trade-off between sequential multiplier and divider.
Friday, December 21st 2018 at 16:30-18:00
Room 202, South Bldg, Academic Center for Computing and Media Studies
S-143H Remote Lecture Room
YRP Mobile Lab
Note: If you would like to attend from Yokosuka, let me know it in advance.
Prof. Tatsuhiro Yokoyama
Research Institute for Sustainable Humanosphere, Kyoto University, Associate Professor
Earth’s ionosphere and space weather prediction
Earth’s upper atmosphere exists as partially ionized plasma, the so-called ionosphere. The phenomena originating from the Sun to the near-Earth space is called “space weather”. Radiowave propagating in the ionospheric plasma is often interfered and cause severe errors in communication and navigation systems. Brief introduction about space weather prediction and recent research topics on ionospheric observations and simulations is presented.
Friday, January 18th 2019 at 16:30-18:00
Room 202, South Bldg, Academic Center for Computing and Media Studies
S-143H Remote Lecture Room
YRP Mobile Lab
Note: If you would like to attend from Yokosuka, let me know it in advance.
Prof. Jaijeet Roychowdhury
University of California, Berkeley
Computing with Oscillators
In the 1950s, Eiichi Goto and John von Neumann showed how Boolean computation could be performed if logic states are encoded in the phase of oscillatory signals. However, the AC-pumped circuit realizations they proposed were not well suited for scaling and miniaturization, hence their scheme could not compete with the level-based logic now ubiquitous in IC implementations. We show how DC-powered self-sustaining nonlinear oscillators of practically any type can function as phase-logic latches. Phase-based Boolean computation therefore becomes possible using a wide variety of natural and engineered oscillators (including CMOS realizations) as substrates. We indicate how phase-encoded logic has inherent noise immunity advantages over level-based logic, and can potentially perform logical operations in a single cycle with low energy consumption. We also show how self-sustaining oscillator networks can be used to implement Ising machines, and outline their considerable promise for solving hard (NP-complete) problems rapidly in hardware.