QFit is an analysis software for physicists to characterize superconducting quantum computing devices. The advancement of superconducting quantum computing depends on creating scalable and reliable hardware devices.

Characterizing properties of a manufactured device is a cumbersome step for hardware development, which involves identifying features in experimental data that reveal quantum behavior of the device, and developing a matching physical model.

After importing a physical model to QFit with a few lines of code, users can effortlessly visualize and calibrate their experimental data, extract and label data features, adjust the model interactively for an optimal fit, and refine it with advanced algorithms through simple clicks. This transforms the intricate characterization process into a step-by-step workflow, reduces the cognitive burden for physicists, and reduces 90% pipeline preparation time. It therefore provides an intuitive and entirely new experience that's incomparable to anything else in the market for physicists working in this cutting-edge field.

INNOVATIONS

QFit revolutionizes how physicists perform device characterization in superconducting quantum computing.

Traditionally, physicists would write extensive code to create a personalized, non-GUI pipeline for their characterization work. This is often inefficient and prone to errors, typically taking one an afternoon, or even a few days, to complete the task. By providing a pre-established digital pipeline with GUI, QFit cuts the setup work by 90%.

Moreover, QFit addresses the limitations of existing tools that fall short in assisting with parameter fitting. Characterizing a superconducting circuit involves: (1) extracting features showing quantum behaviors from two-tone spectroscopy data; (2) creating and coding physical models to simulate these behaviors; (3) tuning model parameters for experimental data alignment. While there are tools that only deal with (1) or (2), QFit fills the missing gap. It overlays experimental data and model prediction for comparison and offers interactive sliders for parameter adjustment. Users can interact directly with model parameters, iterate through cycles of experimenting and receive immediate feedback, and finalize on model parameters that fit most precisely.

Contrary to traditional scientific tools, which often limit model selection, target narrow audiences, and are rarely designed to be user friendly. QFit supports customizable models for diverse needs and appeals to a wider audience. Designed with user convenience in mind, we innovated the experience to improved the learnability through tasks breakdown, a series of user tests with iterations, and adjustment around information hierarchy.

USER-CENTERED DESIGNS

The navigation of QFit is based on 4 tabs, each representing a single step of device characterization. It sets up one goal for each step, helps users to focus on one thing at a time and therefore reduce their cognitive burden. The navigation tabs also help users to understand the sequence and have less frustration by knowing where they are in the entire process. QFit guides users throughout the experience with the use of call to actions.

QFit visualizes model predictions with experimental data, so that users receive real-time feedback for every edit they make, therefore fostering faster iteration. It also helps users to manage their experimental data and extracted features.

In its settings menu, QFit enables users to customize the coloring for their visualized data in order to identify features in different situations and for improved accessibility.

With all these approaches and the design and technology discussed above, QFit condenses what used to take hours into a matter of minutes, thereby accelerating quantum hardware development.

VISUAL ASPECTS

The design aesthetic of QFit centers around user experience, refined through iterations based on feedback and the frustrations identified during user tests. The software offers various color settings for data visualization to enhance accessibility. We employ color and weight distinctions to clearly separate primary from secondary elements and to differentiate between active and inactive user interfaces. Data table columns are color-coded to aid in distinguishing between them. To minimize noise, non-essential and additional information is collapsed by default. Additionally, QFit adopts a dark mode interface and is designed with figurative contrast tests to alleviate eye strain and facilitate focus. This consideration is crucial because, although QFit significantly streamlines the device characterization process, it remains a task that demands considerable time and attention.