SLStella Lau

Stella Lau

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Stella Lau has a total of 14 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Specification and Verification of Strong Timing Isolation of Hardware Enclaves

    by Stella Lau, Thomas Bourgeat, Clément Pit-Claudel, Adam Chlipala

    The process isolation enforceable by commodity hardware and operating systems is too weak to protect secrets from malicious code running on the same machine: attacks exploit timing side channels derived from contention on shared microarchitectural resources to extract secrets. With appropriate hardware support, however, we can construct isolated enclaves and safeguard independent processes from interference through timing side channels, a step towards confidentiality and integrity guarantees. In this paper, we describe our work on formally specifying and verifying that a synthesizable hardware architecture implements strong timing isolation for enclaves. We reason about the cycle-accurate semantics of circuits with respect to a trustworthy formulation of strong isolation based on "air-gapped machines" and develop a modular proof strategy that sidesteps the need to prove functional correctness of processors. We apply our method on a synthesizable, multicore, pipelined RISC-V design formalized in Coq.


  • Co-Archiving a Ballet Repertory Course during the COVID-19 Pandemic

    by Louise Lee, Xianliang Dong, Jenny Yee, Stella Lau

    Archiving is a process of organizing and creating content for sustainable, long-term access. However, the “archivists” actively shape the content in the realm of the performing arts, which often involves adding a new layer of knowledge to the original artform. This article explores the archiving of a ballet repertory course conducted in a higher education institution in Hong Kong during the pandemic, including the archival product and the collaborative process between a multidisciplinary team of people from dance, education, and library science. While teaching repertory online has its challenges, being “online” also offered unique learning experiences. Aiming to illustrate the topics related to hybrid pedagogy and individual creativity in use during the pandemic, the team conducted interviews with some participants for the purpose of creating a series of short educational films. These materials were collected and stored in the institution’s digital repository as the first archive for a specific course.

  • Open Access

    Effective simulation and debugging for a high-level hardware language using software compilers

    by Clément Pit-Claudel, Thomas Bourgeat, Stella Lau, Arvind, Adam Chlipala

    Rule-based hardware-design languages (RHDLs) promise to enhance developer productivity by offering convenient abstractions. Advanced compiler technology keeps the cost of these abstractions low, generating circuits with excellent area and timing properties. Unfortunately, comparatively little effort has been spent on building simulators and debuggers for these languages, so users often simulate and debug their designs at the RTL level. This is problematic because generated circuits typically suffer from poor readability, as compiler optimizations can break high-level abstractions. Worse, optimizations that operate under the assumption that concurrency is essentially free yield faster circuits but often actively hurt simulation performance on platforms with limited concurrency, like desktop computers or servers. This paper demonstrates the benefits of completely separating the simulation and synthesis pipelines. We propose a new approach, yielding the first compiler designed for effective simulation and debugging of a language in the Bluespec family. We generate cycle-accurate C++ models that are readable, compatible with a wide range of traditional software-debugging tools, and fast (often two to three times faster than circuit-level simulation). We achieve these results by optimizing for sequential performance and using static analysis to minimize redundant work. The result is a vastly improved hardware-design experience, which we demonstrate on embedded processor designs and DSP building blocks using performance benchmarks and debugging case studies.


  • High-Level Human Adenosine Deaminase Expression in Dog Skin Fibroblasts Is Not Sustained Following Transplantation

    by N. Ramesh, Stella Lau, Theo D. Palmer, Rainer Storb, William R. A. Osborne

    Primary skin fibroblasts are an attractive target tissue for retroviral-mediated gene therapy; however, transient expression of therapeutic genes has been a recurrent problem in several rodent models. The gradual decrease in gene expression could be tissue or species specific. To investigate the phenomenon further, human adenosine deaminase (ADA) expression was monitored in genetically modified skin fibroblasts transplanted in beagle dogs. In culture, transduced canine fibroblasts expressed high levels of human ADA activity (33.6 μmoles adenosine metabolized per hour per milligram of soluble protein) in comparison to canine ADA in untreated control cells (1.3 μmole/hr · mg protein) and for 2 weeks following transplantation, the graft contained up to four-fold more enzyme activity from human ADA than the endogenous canine enzyme. However, by 10 weeks, human ADA expression was undetectable. At the time when human ADA expression was greatly reduced, DNA analysis showed the presence of vector sequences. These results directly parallel those observed in rodent models and suggest retroviral vector inactivation is a tissue-specific rather than species-specific mechanism. A major question in human gene therapy protocols is how long the transferred gene will remain active in the target cells. Unfortunately, many constructs that appear to maintain stable gene expression in cultured cells do not do so in vivo in mice. Ramesh et al. have studied this phenomenon in primary skin fibroblasts transplanted into dogs using the adenosine deaminase gene. They report that the same loss of gene expression occurs in the canine in vivo as occurs in rodents in vivo.