國立臺灣大學｜電子工程學研究所
Graduate Institute of Electronics Engineering
National Taiwan University
National Taiwan University
We often enter our careers with a script of what we will and won’t do, only to discover that life has its own plans; every never declared is a gamble as uncertainty becomes the norm. Today’s engineering challenges compound this unpredictability with system-level complexity, as tightly-coupled constraints evolve at an accelerating pace, constantly challenging existing assumptions. Thus, an open, proactive mindset becomes a critical asset. This talk draws on personal experience to invite the audience to move from static objectives to adventures grounded in integrity, resilience, curiosity, and grit. Our goal is not to avoid missteps perfectly, but to remain adaptable enough to stay relevant in the field — long enough to grow, to contribute, and to actually enjoy the ride.
In this talk, we will review our recent work about high-speed APD and VCSEL arrays to overcome the fundamental limitations of transceiver modules for free-space optical communication. State-of-the-art brightness, EO/OE bandwidth, and large active window size can be achieved simultaneously.
The first part of the talk will describe the common challenges that occur during comparative
evaluations of the performance of formal-methods tools (solvers, verifiers, etc.). We will also
describe the tool kit BenchExec, which makes it easy for benchmarking engineers to use Linux kernel
features like cgroups, name spaces, and overlay file systems.
The second part of the talk describes the competition on software verification,
a yearly comparative evaluation of tools for software verification.
[TALK INFORMATION]
Analog Design Experiments with AI
Speaker：Prof. Behzad Razavi(University of California, Los Angeles )
Time：2025/09/24(Web) PM4:00~5:00
Venue：NTU EE2-105
Registration：https://reurl.cc/E6V8KR
Abstrate﹕
With the rapid rise of AI in various fields, we naturally wonder whether it can help or even replace analog designers. As a first step in our investigation, we evaluate the present capabilities of ChatGPT and probe its conceptual understanding beyond what it can find online. We limit our study to undergraduate analog CMOS circuits, pose questions to ChatGPT, and analyze the answers that it provides. We report answers of which many are surprising, either because they are not obvious but correct or because they are obvious but incorrect. Whether or not AI will replace analog designers in the near future will be clear by the end of the presentation.
Bio﹕
Behzad Razavi is Professor of Electrical Engineering at UCLA, where he conducts research on analog and RF integrated circuits. Prof. Razavi has served as an IEEE Distinguished Lecturer and published more than 200 papers and nine books. He has received nine IEEE best paper awards and six teaching and education awards, and his books have been published in seven languages. He received the IEEE Pederson Award in Solid-State Circuits and was recognized as a top author in the 50-year and 75-year histories of the IEEE International Solid-State Circuits Conference. He is a member of the US National Academy of Engineering and a fellow of the US National Academy of Inventors.
