Research Overview: Toward End-to-End Quantum ApplicationsThis is an exciting time for quantum computing! With the availability of prototypes of quantum machines, especially the recently established quantum supremacy and two 53-qubit machines from both Google and IBM, it becomes possible for quantum computing researchers to go beyond purely theoretical studies and to investigate the implementation and actual performance of real-world quantum applications. A paramount goal of the QC research is to bridge the gap between theoretical studies and the limitation of real-world QC hardware to achieve end-to-end quantum applications. Research GoalsMy research aims to fill the gap between purely theoretical study of quantum applications and experimental implementation of quantum hardware and to achieve end-to-end quantum applications. In particular, I believe that research on end-to-end applications would benefit from end-to-end thinking. To that end, my research investigates a broad range of perspectives of quantum computing, including its theoretical foundation, the theory and applications of near-term quantum devices, and the software foundation (programming languages and system) of quantum computing. My research has also been published in prestigious venues of all relevant fields such as quantum information (e.g., QIP), theoretical computer science (e.g., STOC, CCC, ICALP), machine learning (e.g., ICML, NeurIPS), and programming languages (e.g., POPL, PLDI). Quantum Applications in Optimization and Machine LearningQuantum Algorithms for Optimization and Machine LearningMy recent research aims to understand the landscape of provable quantum advantage in optimization and machine learning, a major targeted domain of quantum applications. To that end, I have developed quantum algorithms with polynomial speed-ups over classical ones for semidefinite programs (SDPs), general convex optimization, training linear and kernelized classifiers, and estimating volumes of high-dimensional convex bodies. My algorithms also hint at possible exponential quantum speed-ups when using quantum data as inputs/outputs of SDPs and the principle component analysis problem. Publications:
Quantum Algorithms for Property TestingAnother well-motivated topic is the property test of quantum and classical distributions. I have closed a long-standing gap between the upper and the lower bound of the sample complexity of testing the whole information of quantum states, so-called quantum tomography, which is a fundamental step to verify the preparation of the experimental setup. I also demonstrated the quantum speed-up in estimating the Shannon and Rényi entropies of classical distributions. Publications:
Variational Quantum Methods and Quantum Neural-NetworksQuantum Neural-networks (i.e., parameterized quantum circuits) are important and promising candidates of applications of quantum machine learning. My research aims to conduct a comprehensive investigation in this regard, including functionality (e.g., representation learning, generative models), training methods (e.g., landscape characterization, efficiency of gradient and non-gradient based optimization methods), and separation between quantum and classical neural-networks. Publications:
Differentiable Quantum Programming LanguagesInspired by the emerging paradigm shift from deep learning towards differentiable programming promoted by prominent classical machine learning researchers, I have initiated the formalization of differentiable quantum programming. This project not only provides the auto- differentiation technique for quantum programs, in particular the possibility of using quantum programs to compute the gradients of another quantum program, for the scalability of gradient-based training in quantum machine learning. It also opens the possibility of designing novel quantum “neuro-symbolic” applications that combine program features/synthesis with simple neural-networks. The study of their classical counterpart, although still in its infancy, has already shown great potential and promise over conventional neural-networks. We demonstrated, for the first time, one such example for quantum machine learning.
Software Foundation of Quantum ComputingVerification of Sequence, Parallel and Concurrent Quantum ProgramsQuantum programs are error-prone and their verification is challenging due to the limit of standard software assurance techniques in the quantum setting. My research investigates the verification of quantum programs via their static analysis with the help of quantum Hoare logic. A prominent approach for program verification is to generate invariants and inductive assertions, which is already a highly non-trivial task even classically. I made the first proposal of quantum invariants and demonstrated its use in quantum program verification, which can be generated by using a semidefinite program (SDP) of dimension exponential in terms of the number of qubits. In ongoing projects, I further improve the invariant-based verification in the term of its expressive power (e.g., including ancilla in Hoare logic) and its scalability. Moving from sequential to parallel and concurrent quantum programs is a logical next step because of important quantum applications in communication and network. Specifically, we aim to develop a quantum Owicki-Gries-like logic by formulating the concept of interference-free proofs of component quantum programs. However, there is a significant difference between classical and quantum interference due to quantum correlations and casual relationship among component quantum programs. Research in quantum information on these topics (e.g., my own research on quantum correlations below) is likely to play a critical role for such development. Publications:
Reliable and Resource-optimized Quantum Program SynthesisNear-term quantum applications will likely focus on Noisy and Intermediate-Scale Quantum (NISQ) computers, where precisely controllable qubits are expensive, error-prone, and scarce. Existing work on quantum algorithms and programming languages assumes this issue will be solved by the hardware, as in classical computers, or with independently designed fault-tolerant protocols, which hence becomes less helpful for implementing near-term quantum applications. I believe that one way to attack these problems is to set aside a one-size-fits-all approach to fault tolerance and instead consider elevating the question of errors and related architecture-specific resource optimization to the level of the programming language and algorithm design. In particular, I am inspired by techniques in approximate computing that optimizes computation on unreliable classical hardware. As the first step, I've built a formal semantics and a logic for reasoning about reliability in the presence of noise in quantum computation. In continuing efforts, I am building an automatic amortized resource analysis tool for quantum programs. I am also building an automated quantum program synthesis, which takes the ideal program and reliability requirements as constraints (which could be generated in Coq via the implementation of our logic) and hardware-specific resource optimization as the objective, and outputs a optimized quantum program with possible additional error mitigation intermediate steps synthesized by a back-end optimizer. This automated tool will then be benchmarked on near-term quantum applications (e.g., variational methods) with specific hardware information. Publications:
Formally Verified Software Tool-chain for Quantum ComputingThe complexity of quantum computing and the limitations of near-term quantum devices suggest that the development of sophisticated quantum algorithms and clever optimizations is more likely to have mistakes. This calls for verifying every stage of quantum computation, from the software tools used to generate quantum circuits to the architecture and system design. I am inspired by formal methods applied in safety-critical domains to ensure correctness of code by construction, especially in the example of CompCert (a C compiler written and proved correct in Coq) and the NSF project of deep specification (a project to develop specifications of software toolchains to prove end-to-end correctness of whole systems). A verified quantum computing stack would ensure that each level of quantum computation is implemented satisfying certain specifications and the correctness of the final system, which would have wide practical impact. This approach is especially appealing to quantum computing since alternative software assurance techniques are very limited due to the substantial expense involved in the quantum setting. As the first step, I have built a proved-correct optimizing compiler for quantum algorithms to optimize the gate count, the depth of circuits, etc., while adhering to any architectural constraints. Publications:
Theoretical Studies in Quantum Information and ComputationEntanglement, Quantum Correlations, and Sum-of-SquaresResearch on this topic studies many aspects of one of the key quantum features, entanglement and non-locality. I attack this topic by exploring a surprising connection between quantum information and the sum-of-squares (SOS) proof in approximation algorithms and the famous Unique Games Conjecture (UGC). This connection allows one to leverage technical advances in one field to apply to the other. Specific problems that I am working on includes the characterization of separable (unentangled) states, the complexity of quantum Merlin-Arthur games with unentangled provers (QMA(2)), the possibility of a quantum-inspired approach to attack the UGC. Publications:
Tamper-resilient Cryptography under Physical AssumptionsDevices, classical or quantum, are subject to tampering in cryptographic settings, especially due to the proliferation of side-channel attacks. These attacks exploit the fact that de- vices leak information to the outside world not just through input-output interaction, but through physical characteristics of computation such as power consumption, timing, and electro-magnetic radiation. Research on this topic explores two possible solutions to protect cryptographic systems from (quantum) side-channel attacks.
Both device-independent and leakage-resilient cryptography can be deemed as tamper-resilient cryptography under physical assumptions. My future plan is to bring these cryptographic designs closer to practice, with better efficiency and broader functionality. Publications:
Quantum Computational ComplexityInteractive proof systems have been a central model in complexity theory with applications ranging from the PCP theorem in hardness of approximation to cryptography. It studies problems with efficiently verifiable proofs via interactions between a polynomial-time verifier and all-powerful provers, where the verifier determines the validity of the proofs. My main contribution on this topic is the development of the Equilibrium Value Method to obtain space-efficient simulations of quantum interactive proof systems, including QIP=PSPACE, QRG(2)=PSPACE. Recently, I have been working on the quantum variant of the PCP theorem in the interactive proof setting. As a concrete first step, I have obtained a parallel repetition result for entangled k-player games. Publications:
|