Day: July 9, 2021

Artificial intelligence is now helping to design computer chips—including the very ones needed to run the most powerful AI code.

Sketching out a computer chip is both complex and intricate, requiring designers to arrange billions of components on a surface smaller than a fingernail. Decisions at each step can affect a chip’s eventual performance and reliability, so the best chip designers rely on years of experience and hard-won know-how to lay out circuits that squeeze the best performance and power efficiency from nanoscopic devices. Previous efforts to automate chip design over several decades have come to little.

But recent advances in AI have made it possible for algorithms to learn some of the dark arts involved in chip design. This should help companies draw up more powerful and efficient blueprints in much less time. Importantly, the approach may also help engineers co-design AI software, experimenting with different tweaks to the code along with different circuit layouts to find the optimal configuration of both.

At the same time, the rise of AI has sparked new interest in all sorts of novel chip designs. Cutting-edge chips are increasingly important to just about all corners of the economy, from cars to medical devices to scientific research.

Chipmakers, including Nvidia, Google, and IBM, are all testing AI tools that help arrange components and wiring on complex chips. The approach may shake up the chip industry, but it could also introduce new engineering complexities, because the type of algorithms being deployed can sometimes behave in unpredictable ways.

At Nvidia, principal research scientist Haoxing “Mark” Ren is testing how an AI concept known as reinforcement learning can help arrange components on a chip and how to wire them together. The approach, which lets a machine learn from experience and experimentation, has been key to some major advances in AI.

The AI tools Ren is testing explore different chip designs in simulation, training a large artificial neural network to recognize which decisions ultimately produce a high-performing chip. Ren says the approach should cut the engineering effort needed to produce a chip in half while producing a chip that matches or exceeds the performance of a human-designed one.

“You can design chips more efficiently,” Ren says. “Also, it gives you the opportunity to explore more design space, which means you can make better chips.”

Nvidia started out making graphics cards for gamers but quickly saw the potential of the same chips for running powerful machine-learning algorithms, and it is now a leading maker of high-end AI chips. Ren says Nvidia plans to bring chips to market that have been crafted using AI but declined to say how soon. In the more distant future, he says, “you will probably see a major part of the chips that are designed with AI.”

Reinforcement learning was used most famously to train computers to play complex games, including the board game Go, with superhuman skill, without any explicit instruction regarding a game’s rules or principles of good play. It shows promise for various practical applications, including training robots to grasp new objects, flying fighter jets, and algorithmic stock trading.

Song Han, an assistant professor of electrical engineering and computer science at MIT, says reinforcement learning shows significant potential for improving the design of chips, because, as with a game like Go, it can be difficult to predict good decisions without years of experience and practice.

His research group recently developed a tool that uses reinforcement learning to identify the optimal size for different transistors on a computer chip, by exploring different chip designs in simulation. Importantly, it can also transfer what it has learned from one type of chip to another, which promises to lower the cost of automating the process. In experiments, the AI tool produced circuit designs that were 2.3 times more energy-efficient while generating one-fifth as much interference as ones designed by human engineers. The MIT researchers are working on AI algorithms at the same time as novel chip designs to make the most of both.

Other industry players—especially those that are heavily invested in developing and using AI—also are looking to adopt AI as a tool for chip design.

Lobbyists and trade groups for big tech companies and equipment manufacturers have long argued that giving consumers more access to the tools required to fix products, whether a smartphone or a car, poses safety and security risks. The debate has gotten especially heated as more products become internet-connected, adding a software element to repairs that in the past might have just required swapping parts.

The links to news reports in the White House’s fact sheet that back up its claims of stymied competition specifically point to the issues around cell phone repair, but the language of the order itself urges the FTC to broaden the right to repair by restricting “tech and other” companies from discouraging DIY tinkering. Such language indicates that the FTC’s regulatory target will be much bigger than the device in your pocket.

In an emailed response to the executive order, a spokesperson from John Deere claims the company “leads our industry in providing repair tools, spare parts, information guides, training videos and manuals needed to work on our machines.” But the spokesperson also says that while less than two percent of tractor repairs require a software update, the company still does not support the right to modify embedded software “due to risks associated with the safe operation of the equipment.”

Turn of the Screw

Proctor, of US PIRG, notes that it could still be awhile before the FTC starts enforcing new repair laws, saying that the rule-making process is “not always an expeditious one.” He cites as an example the FTC’s finalization of a rule around “Made in the USA” labels that are falsely applied to products not manufactured in the US. (Congress first enacted legislation around “Made in the USA” claims back in 1994, but for years there was bipartisan consensus that this kind of fraud shouldn’t be subject to tough penalties. Just last week, the FTC codified the rules in such a way that violators would be penalized.)

“Right to Repair is even more complex than that case, and if this is just a directive towards rule-making, it might kick off another long process,” Proctor says. “Still, I’m hopeful that this is a mechanism that gets us to where we need to go a little faster.”

Sheehan from iFixit is more optimistic that the FTC might act quickly around Right to Repair, partly because the agency recently introduced a series of changes designed to streamline rule-making procedures—and partly because the order is coming directly from the White House. “Obviously we want the agency to move quickly on this, and pressure from the Biden administration does make that more possible,” Sheehan says.

A spokeswoman for the FTC declined to comment directly on the matter, instead pointing towards the White House’s statement and the report that the Commission already released in May.

In that report, the FTC concluded that products have, in fact, become harder to fix and maintain, and that “repair restrictions have … steered consumers into manufacturers’ repair networks to replace products before the end of their useful lives.” The FTC also noted that repair restrictions may also “place a greater financial burden on communities of color and lower-income Americans.”

But the FTC also warned in the May report that Right to Repair is a complicated issue, and that expanding consumers’ repair options, whether through industry initiatives or through legislation, “raises numerous issues that will warrant examination.”

Ultimately, the Right to Repair fight will likely still continue at the state level, and advocates plan to continue to lobby Congress for changes as well.

“I think, depending on the scope of the FTC rules, this may not be a substitute for what Congress can do and what states can do,” Sheehan says. As many as 25 states have considered Right to Repair legislation this year, but that, of course, doesn’t mean the bills in those states will be signed into law. A few states have what Sheehan calls “repair-related laws,” including California, Rhode Island, and Indiana. Right now, Massachusetts is the only state with an official Right to Repair law for automobiles, which won the vote by a large margin in 2012 and again in 2020, despite vocal opposition from a coalition of big automakers.

“Whatever rule the FTC passes, it will be up to the FTC to enforce,” Sheehan says. “Whereas state legislation can be enforced by state attorneys general, and occasionally they have more leeway or more resources to focus on these things than the FTC might in the context of all its other many priorities.”

---

More Great WIRED Stories