Fault Tolerance For Quantum Computing Errors


Scientists working on quantum computers—dream machines that might solve problems that would exceed any supercomputer—are learning to identify and fix their errors like a kid learns arithmetic. 

In the most recent stage, a team showed a method for detecting mistakes in the setup of a quantum bit, or qubit, that is guaranteed not to exacerbate the problem. 

Such "fault tolerance" is a crucial step toward the ultimate aim of keeping fussy qubits alive long enough to be controlled. 

“It seems to be a genuine watershed moment,” says Scott Aaronson, a theoretical computer scientist at the University of Texas at Austin who wasn't involved in the research. 

“We knew it was just a matter of time until someone did something like this.” 

However, John Martinis, an experimental physicist at the University of California, Santa Barbara, wonders whether the latest study's authors are exaggerating their findings. 

He describes it as a "really good step. But it's just a first step.” 

A traditional computer uses small electrical switches, or bits, that can be set to 0 or 1, while a quantum computer uses qubits that can be set to 0 and 1 at the same time. 

A qubit may be a single ion spinning one way, the other, or both directions at the same time, or a small circuit of superconducting metal with two distinct energy states. 

  • A quantum computer can encode all of the possible solutions to particular problems as quantum waves sloshing across the qubits thanks to such both-ways-at-once states. 
  • Interference cancels out the incorrect answers, allowing the correct solution to emerge. 
  • Such methods would allow a big quantum computer to rapidly factor enormous numbers, which is difficult for a regular computer to do, and therefore defeat encryption systems used to secure data on the internet. 

However, even the tiniest disturbance may destabilize a qubit's fragile condition. 

  • If a qubit were like a regular bit, researchers could simply duplicate it and count the majority to keep it in the correct condition. 
  • If a duplicate does flip, adding up several subsets of the bits (so-called parity tests) will disclose which one it is. 

Quantum theory, on the other hand, prohibits the copying of one qubit's state onto another. 

  • Worse, every effort to test a qubit to determine if it is in the proper state causes it to collapse to one of two states: 0 or 1. 
  • Researchers circumvent these issues by using entanglement, a quantum link that enables them to distribute the state of an initial "logical" qubit—the object that will ultimately execute the required operation—across many physical qubits. 
  • A 0-and-1 state of one qubit, for example, may be extended to three qubits in a condition in which all three are 0 at the same time. 
  • Researchers may then entangle additional ancillary qubits with the group and measure the ancillary qubits to identify faults in the main qubits—without ever touching them—in a quantum version of parity checks. 

In fact, the method is considerably more complex since developers must avoid two kinds of errors: 

  1. bit flips 
  2. and phase flips. 

Despite this, scientists have made progress. 

In June, Google researchers using superconducting qubits demonstrated that spreading a logical qubit over as many as 11 physical qubits with 10 ancillas may decrease the incidence of one kind of mistake but not both at the same time. 

Now, physicists Laird Egan and Christopher Monroe of the University of Maryland (UMD) in College Park, together with others, have shown a method that simultaneously corrects both kinds of flips—and therefore any mistake. 

Individual ytterbium ions are trapped in an electromagnetic field on the chip's surface to form qubits. 

  • The researchers utilized nine ions to encode a single logical qubit, as well as four more ions to keep track of the primary ones. 
  • Most importantly, in certain respects, the encoded logical qubit outperformed the physical ones on which it is based. 
  • The researchers, for example, were able to prepare either the logical 0 or logical 1 state 99.67 percent of the time, which is higher than the 99.54 percent for individual qubits. 

Monroe, creator of IonQ, a firm creating ion-based quantum computers, says, “This is truly the first time where the quality of the [logical] qubit is greater than the components that encode it.”

However, that the encoded qubit did not outperform the individual ions in every manner. 

Instead, the true breakthrough is proving fault tolerance, which implies that the error-correcting mechanism does not create more mistakes than it corrects. 

Fault tolerance is the design concept that prevents mistakes from spreading.

Martinis, on the other hand, has reservations about the term's usage. 

Researchers must also accomplish two additional things, in order to claim genuine fault-tolerant mistake correction. 

  • They must demonstrate that as the number of physical qubits grows, the mistakes in a logical qubit become exponentially less. 
  • They must also demonstrate that they can measure the auxiliary qubits frequently in order to keep the logical qubit stable, he adds. 

Those are the apparent next steps for the UMD and IonQ teams.

He points out that in order for the encoded logical qubit to outperform the underlying physical qubits in every manner, the latter must first have a low enough error rate. 

~ Jai Krishna Ponnappan

You may also want to read more about Quantum Computing here.

What Is Artificial General Intelligence?

Artificial General Intelligence (AGI) is defined as the software representation of generalized human cognitive capacities that enables the ...