Skip to main content

Darwin Approves: Berkeley Evolves Analog Design

Digital design is hard. But in the right environment, digital circuits are more forgiving than analog. That 3.3V signal coming out of the chip has to drop a lot along the way to not be a logic level at the destination. If you are trying to push the boundary then digital design has much of analog design, but mostly you get a bit of a pass on many things that plague analog designers. Berkeley’s AI research group has been experimenting with using deep learning to evolve analog IC design.

Analog ICs are plagued with noise sources and often don’t have the margins that digital circuit designers enjoy. According to the post by [Kourosh Hakhamaneshi], designers often build a few blocks and attempt to lay them out in a way that should work and meet other requirements. Then they employ simulation, make changes as required, and simulate again. Accurate simulations can be very time intensive. You can read the actual paper, too, should you want to dig into the details.

The problem is really one of optimization. An amplifier for example might need a minimum gain and bandwidth. In addition, there is always the need to take up as little space on the die as possible — just as you don’t want to waste space on a PCB. Since simulation is so expensive, the work appears to try to train a neural network to determine if one design is better than another based on the desired design criteria.

The result isn’t quantitative, but knowing which of two designs is better is very useful for selecting candidates to test in simulation. Interestingly, according to the post, it sounds as if the learning model trains itself on random permutations of the exact circuit you want to design — which makes sense. This training corpus requires simulation, of course, so that eats into the potential savings.

This technique could have applications with other optimization problems, too. We generally like simulation, even if ours don’t usually take that long. Our favorite one is pretty simple and runs in a browser.



from Hackaday https://ift.tt/2nqqUGx

Comments

Popular posts from this blog

Crypto exchange Binance prepares to add margin trading ‘soon’

Binance, the world’s most prominent crypto exchange, says it is close to adding a much-anticipated margin trading feature to its service following weeks of speculation. The company tweeted confirmation of the upcoming feature in a screenshot which subtly teases the imminent arrival of margin trading options. Binance CEO Changpeng Zhao (pictured above) first revealed that the feature was headed to Binance during a live stream following a hack earlier this month that saw Binance lose around $40 million in Bitcoin . TechCrunch understands that margin trading has been beta tested among selected users. A Binance representative declined to comment on the specifics, but did confirm that margin trading will be available on Binance.com “soon.” Dark mode or Light mode ? #Binance pic.twitter.com/pGSb1np4yp — Binance (@binance) May 24, 2019   Margin trading, which lets traders use their balance as collateral to super-size their buying power, is seen by many as an important growth...

World Economic Forum launches Global AI Council to address governance gaps

The World Economic Forum is creating a series of councils that create policy recommendations for use of things like AI, blockchain, and precision medicine. Read More from VentureBeat http://bit.ly/2EKBjD4

The hidden cost of food delivery

Noah Lichtenstein Contributor Share on Twitter Noah Lichtenstein is the founder and managing partner of Crossover , a diversified private technology fund backed by institutional investors, technology execs and professional athletes and entertainers. More posts by this contributor What Studying Students Teaches Us About Great Apps I’ll admit it: When it comes to food, I’m lazy. There are dozens of great dining options within a few blocks of my home, yet I still end up ordering food through delivery apps four or five times per week. With the growing coronavirus pandemic closing restaurants and consumers self-isolating, it is likely we will see a spike in food delivery much like the 20% jump China reported during the peak of its crisis. With the food delivery sector rocketing toward a projected $365 billion by the end of the decade, I’m clearly not the only one turning to delivery apps even before the pandemic hit. Thanks to technology (and VC funding) we can get a ri...