Search

The diversity of California's teaching force continues to lag behind its student population. While students of Color make up 78% of the state's K-12 population, Teachers of Color and Indigenous Teachers (TOCIT) comprise just 34% of the teaching workforce (California Department of Education, 2018), a statistic that has dominated the teacher shortage narrative. Although there is promise in the fact that teacher education program (TEP) enrollment is more diverse than the state's current educator workforce, it's still 27% less diverse than the state's K-12 students. As demand for greater representation increases, so have the initiatives to recruit and retain more racially diverse teachers. In the past few years, California has spent billions in an effort to diversify its teacher workforce, including Assembly Bill 520, which allocated $15 million to be distributed to school districts to develop and implement programs that diversify teaching staffs, and Assembly Bill 130, which appropriated $350 million over the next five years to create or expand Teacher Residency Programs, a pathway that has been shown to recruit and retain higher numbers of TOCIT (California Legislative Information, 2023). This study sets out to (1) explore how current policies, structures, practices, attitudes, and ideologies across the pipeline and profession impede the recruitment and retention of TOCIT, and (2) better understand the racialized experiences and perspectives of pre-service and in-service Teachers of Color and Indigenous Teachers. The following paper is based on a mix of qualitative and quantitative data collected from system leaders in teacher preparation, pre-service teachers, in-service teachers, and former teachers over the course of approximately one year: spring 2021 to summer 2022.

This work explores the challenge of enhancing Automatic Speech Recognition (ASR) model performance across various user-specific domains while preserving user data privacy. We employ federated learning and parameter-efficient domain adaptation methods to solve the (1) massive data requirement of ASR models from user-specific scenarios and (2) the substantial communication cost between servers and clients during federated learning. We demonstrate that when equipped with proper adapters, ASR models under federated tuning can achieve similar performance compared with centralized tuning ones, thus providing a potential direction for future privacy-preserved ASR services. Besides, we investigate the efficiency of different adapters and adapter incorporation strategies under the federated learning setting.

The Nab experiment will measure the electron-neutrino correlation and Fierz interference term in free neutron beta decay to test the Standard Model and probe Beyond the Standard Model Physics. Using National Instrument's PXIe-5171 Reconfigurable Oscilloscope module, we have developed a data acquisition system that is not only capable of meeting Nab's specifications, but flexible enough to be adapted in situ as the experimental environment dictates. The L1 and L2 trigger logic can be reconfigured to optimize the system for coincidence event detection at runtime through configuration files and LabVIEW controls. This system is capable of identifying L1 triggers at at least $1$ MHz, while reading out a peak signal rate of approximately $2$ GB/s. During commissioning, the system ran at a sustained readout rate of $400$ MB/s of signal data originating from roughly $6$ kHz L2 triggers, well within the peak performance of the system., Comment: 11 pages, 6 figures

Diffusion models (DPMs) have demonstrated remarkable performance in image generation, often times outperforming other generative models. Since their introduction, the powerful noise-to-image denoising pipeline has been extended to various discriminative tasks, including image segmentation. In case of medical imaging, often times the images are large 3D scans, where segmenting one image using DPMs become extremely inefficient due to large memory consumption and time consuming iterative sampling process. In this work, we propose a novel conditional generative modeling framework (LDSeg) that performs diffusion in latent space for medical image segmentation. Our proposed framework leverages the learned inherent low-dimensional latent distribution of the target object shapes and source image embeddings. The conditional diffusion in latent space not only ensures accurate n-D image segmentation for multi-label objects, but also mitigates the major underlying problems of the traditional DPM based segmentation: (1) large memory consumption, (2) time consuming sampling process and (3) unnatural noise injection in forward/reverse process. LDSeg achieved state-of-the-art segmentation accuracy on three medical image datasets with different imaging modalities. Furthermore, we show that our proposed model is significantly more robust to noises, compared to the traditional deterministic segmentation models, which can be potential in solving the domain shift problems in the medical imaging domain. Codes are available at: https://github.com/LDSeg/LDSeg., Comment: 9 pages, 7 figures

This paper examines the intersection of goals and values within grassroots organizations operating in the realm of quantum technologies (QT) education. It delineates a fundamental distinction between the objective to provide education and the drive to democratize learning through principles of inclusivity, accessibility, and diversity. The analysis reveals how these organizations navigate their nascent stages, grappling with the dual challenge of adhering to their foundational values while aspiring for sustainable growth and development in the highly specialized field of QT. The study uncovers the strategic approaches adopted by these entities, including efforts to create educational ecosystems and foster community engagement. The research underscores the potential vulnerabilities of these grassroots organizations, particularly in relation to the longevity and evolution of their initiatives as members transition into professional roles within the quantum sector. Through this investigation, the paper contributes to a nuanced understanding of how emerging educational organizations in the QT field balance their ideological commitments with practical growth considerations, highlighting the critical factors that influence their trajectory and impact., Comment: 21 pages, 6 figures

We analyze ionized gas emission lines in deep rest-frame optical spectra of 16 quiescent galaxies at redshift $1.7

We present the first comparison of waveforms evaluated using the effective-one-body (EOB) approach and gravitational self-force (GSF) theory for inspiralling black hole binaries with a non-spinning primary and a spinning secondary. This paper belongs to a series of papers comparing the EOB model TEOBResumS to GSF results, where the latter are used to benchmark the EOB analytical choices in the large-mass-ratio regime. In this work, we explore the performance of two gauge choices for the gyro-gravitomagnetic functions GS, GS* entering the spin-orbit sector within the EOB dynamics. In particular, we consider the usual gauge of TEOBResumS, where GS and GS* only depend on the inverse radius and the radial momentum, and a different gauge where these functions also depend on the azimuthal momentum. The latter choice allows us to exploit as prefactor in GS* the complete expression GKS* for a spinning particle on Kerr. As done previously, we employ both waveform alignments in the time domain and a gauge-invariant frequency-domain analysis to gain a more complete understanding of the impact of the new analytical choice. The frequency-domain analysis is particularly useful in confirming that the gyro-gravitomagnetic functions in the new chosen gauge bring the EOB spin contribution at 1st post-adiabatic order closer to the GSF one. We finally implement the improved functions within the public code for TEOBResumS-Dal\'i, which already incorporates eccentricity. In this way, we upgrade the EOB model for extreme-mass-ratio inspirals presented in our previous work., Comment: 15 pages, 6 figures

The nucleation of gas hydrates is of great interest in flow assurance, global energy demand, and carbon capture and storage. A complex molecular understanding is critical to control hydrate nucleation and growth in the context of potential applications. Molecular dynamics is employed in this work combined with the mechanical definition of surface tension to assess the surface stresses that control some of the behavior at the interface. Ensuring careful sampling and simulation behavior, this work extracts meaningful results from molecular properties. We characterize the interfacial tension for sII methane/ethane hydrate and gas mixtures for different temperatures and pressures. We find that the surface tension trends positively with temperature in a balance of water-solid and water-gas interactions. The molecular dipole shows the complexities of water molecule behavior in small, compressed pre-melting layer that emerges as a quasi-liquid. These behaviors contribute to the developing knowledge base surrounding practical applications of this interface., Comment: 17 pages, 7 figures, 1 table

Massive quiescent galaxies in the young universe are expected to be quenched rapidly, but it is unclear whether they all experience starbursts before quenching and what physical mechanism drives rapid quenching. We study 16 massive quiescent galaxies ($\log(M_\star/M_\odot) > 10$) at $z\sim2$ selected from a representative sample of the Blue Jay survey. We reconstruct their star formation histories by fitting spectral energy distribution models to the JWST/NIRSpec $R\sim1000$ spectra. We find that massive quiescent galaxies can be split into three categories with roughly equal numbers of galaxies according to their SFHs: 1) Relatively old galaxies quenched at early epochs; 2) Galaxies that are rapidly and recently quenched after a flat or bursty formation history (depending on the assumed prior); 3) Galaxies that are rapidly and recently quenched after a major starburst. Most recently quenched galaxies show neutral gas outflows, probed by blueshifted $\rm Na\,I\,D$ absorption, and ionized gas emission, with line ratios consistent with active galactic nucleus (AGN) diagnostics. This suggests that AGN activity drives multi-phase gas outflows, leading to rapid quenching. By tracing back the SFHs of the entire sample, we predict the number density of massive quiescent galaxies at $z=4-6$: $n=3.0\pm1.4\times10^{-5}\,\rm Mpc^{-3}$. The two oldest massive quiescent galaxies in our sample appear to have extremely early formation and quenching ($z\gtrsim6$), possibly descendants of early post-starbursts at $z>3$. These galaxies still show neutral gas reservoirs and low-level star formation, consistent with weak H$\alpha$ emission, perhaps because the ejective AGN feedback that caused rapid quenching has weakened over time., Comment: Submitted to ApJ; 21 pages, 9 figures, 3 tables (excluding appendix)

Deep learning (DL) has recently emerged as a pivotal technology for enhancing magnetic resonance imaging (MRI), a critical tool in diagnostic radiology. This review paper provides a comprehensive overview of recent advances in DL for MRI reconstruction. It focuses on DL approaches and architectures designed to improve image quality, accelerate scans, and address data-related challenges. These include end-to-end neural networks, pre-trained networks, generative models, and self-supervised methods. The paper also discusses the role of DL in optimizing acquisition protocols, enhancing robustness against distribution shifts, and tackling subtle bias. Drawing on the extensive literature and practical insights, it outlines current successes, limitations, and future directions for leveraging DL in MRI reconstruction, while emphasizing the potential of DL to significantly impact clinical imaging practices.

How well do existing federated learning algorithms learn from client devices that return model updates with a significant time delay? Is it even possible to learn effectively from clients that report back minutes, hours, or days after being scheduled? We answer these questions by developing Monte Carlo simulations of client latency that are guided by real-world applications. We study synchronous optimization algorithms like FedAvg and FedAdam as well as the asynchronous FedBuff algorithm, and observe that all these existing approaches struggle to learn from severely delayed clients. To improve upon this situation, we experiment with modifications, including distillation regularization and exponential moving averages of model weights. Finally, we introduce two new algorithms, FARe-DUST and FeAST-on-MSG, based on distillation and averaging, respectively. Experiments with the EMNIST, CIFAR-100, and StackOverflow benchmark federated learning tasks demonstrate that our new algorithms outperform existing ones in terms of accuracy for straggler clients, while also providing better trade-offs between training time and total accuracy.

We study the problem of robust multivariate polynomial regression: let $p\colon\mathbb{R}^n\to\mathbb{R}$ be an unknown $n$-variate polynomial of degree at most $d$ in each variable. We are given as input a set of random samples $(\mathbf{x}_i,y_i) \in [-1,1]^n \times \mathbb{R}$ that are noisy versions of $(\mathbf{x}_i,p(\mathbf{x}_i))$. More precisely, each $\mathbf{x}_i$ is sampled independently from some distribution $\chi$ on $[-1,1]^n$, and for each $i$ independently, $y_i$ is arbitrary (i.e., an outlier) with probability at most $\rho < 1/2$, and otherwise satisfies $|y_i-p(\mathbf{x}_i)|\leq\sigma$. The goal is to output a polynomial $\hat{p}$, of degree at most $d$ in each variable, within an $\ell_\infty$-distance of at most $O(\sigma)$ from $p$. Kane, Karmalkar, and Price [FOCS'17] solved this problem for $n=1$. We generalize their results to the $n$-variate setting, showing an algorithm that achieves a sample complexity of $O_n(d^n\log d)$, where the hidden constant depends on $n$, if $\chi$ is the $n$-dimensional Chebyshev distribution. The sample complexity is $O_n(d^{2n}\log d)$, if the samples are drawn from the uniform distribution instead. The approximation error is guaranteed to be at most $O(\sigma)$, and the run-time depends on $\log(1/\sigma)$. In the setting where each $\mathbf{x}_i$ and $y_i$ are known up to $N$ bits of precision, the run-time's dependence on $N$ is linear. We also show that our sample complexities are optimal in terms of $d^n$. Furthermore, we show that it is possible to have the run-time be independent of $1/\sigma$, at the cost of a higher sample complexity.

We propose a sequential Monte Carlo (SMC) method to efficiently and accurately compute cut-Bayesian posterior quantities of interest, variations of standard Bayesian approaches constructed primarily to account for model misspecification. We prove finite sample concentration bounds for estimators derived from the proposed method along with a linear tempering extension and apply these results to a realistic setting where a computer model is misspecified. We then illustrate the SMC method for inference in a modular chemical reactor example that includes submodels for reaction kinetics, turbulence, mass transfer, and diffusion. The samples obtained are commensurate with a direct-sampling approach that consists of running multiple Markov chains, with computational efficiency gains using the SMC method. Overall, the SMC method presented yields a novel, rigorous approach to computing with cut-Bayesian posterior distributions.

Artificial intelligence (AI) assistants such as GitHub Copilot and ChatGPT, built on large language models like GPT-4, are revolutionizing how programming tasks are performed, raising questions about whether code is authored by generative AI models. Such questions are of particular interest to educators, who worry that these tools enable a new form of academic dishonesty, in which students submit AI generated code as their own work. Our research explores the viability of using code stylometry and machine learning to distinguish between GPT-4 generated and human-authored code. Our dataset comprises human-authored solutions from CodeChef and AI-authored solutions generated by GPT-4. Our classifier outperforms baselines, with an F1-score and AUC-ROC score of 0.91. A variant of our classifier that excludes gameable features (e.g., empty lines, whitespace) still performs well with an F1-score and AUC-ROC score of 0.89. We also evaluated our classifier with respect to the difficulty of the programming problem and found that there was almost no difference between easier and intermediate problems, and the classifier performed only slightly worse on harder problems. Our study shows that code stylometry is a promising approach for distinguishing between GPT-4 generated code and human-authored code., Comment: 13 pages, 5 figures, MSR Conference

To constrain the universe before recombination (380000 years after the Big Bang), we mostly rely on the measurements of the primordial abundances that indicate the first insight into the thermal history of the universe. The first production of the light elements is obtained by the Big Bang Nucleosynthesis (BBN). The production of the elements deuterium, helium-3, and helium-4 during BBN matches well the observations; however, the production of lithium (Lithium-7) based on the Standard Big Bang Nucleosynthesis (SBBN) is found to be higher by about a factor of three than the observed abundance from metal-poor halo stars. This so-called "Cosmological Lithium Problem" is still elusive and needs to be resolved. One important attempt to resolve this problem is to invoke a non-standard description of the SBBN to decrease the lithium abundance. In our previous work, we encountered a problem that the decrease in the Lithium-7 abundance requires an increase in the deuterium abundance to maximum values that are not accepted by observations. In the present work, a decrease in the lithium abundance could be achieved without maximizing the deuterium abundance by modifying the time-temperature relation in the range (0.43-0.91) GK during the nucleosynthesis process. This range is crucial to reducing the strong correlation between lithium and deuterium production. The main conclusion of the present work is that Lithium-7 abundance in the atmospheres of metal-poor stars cannot be analyzed without considering possible modifications to the primordial nucleosynthesis.

Motivation: RNA design aims to find at least one sequence that folds with the highest probability into a designated target structure, but some structures are undesignable in the sense that no sequence folds into them. Identifying undesignable structures is useful in delineating and understanding the limit of RNA designability, but has received little attention until recently. In addition, existing methods on undesignability are not scalable and not interpretable. Results: We introduce a novel graph representation and a new general algorithmic framework to efficiently identify undesignable motifs in a secondary structure. The proposed algorithm enumerates minimal motifs based on the loop-pair graph representation of a structure and establishes the undesignability of a motif by proposing rival substructure(s). Our work can also identify unique minimum undesignable motifs across different structures. Our implemented algorithms successfully identify 26 unique minimum undesignable motifs among 18 undesignable puzzles from the benchmark Eterna100. Additionally, our algorithm is so efficient that it scales to natural structures of 16S and 23S Ribosomal RNAs (about 1,500 and 3,000 nucleotides, resp.), and finds all of those structures in the widely used ArchiveII database to be undesignable, with 73 unique minimum undesignable motifs, under the standard Turner energy model in ViennaRNA.

Recent Large Language Models (LLMs) have demonstrated significant capabilities in generating code snippets directly from problem statements. This increasingly automated process mirrors traditional human-led software development, where code is often written in response to a requirement. Historically, Test-Driven Development (TDD) has proven its merit, requiring developers to write tests before the functional code, ensuring alignment with the initial problem statements. Applying TDD principles to LLM-based code generation offers one distinct benefit: it enables developers to verify the correctness of generated code against predefined tests. This paper investigates if and how TDD can be incorporated into AI-assisted code-generation processes. We experimentally evaluate our hypothesis that providing LLMs like GPT-4 and Llama 3 with tests in addition to the problem statements enhances code generation outcomes. We experimented with established function-level code generation benchmarks such as MBPP and HumanEval. Our results consistently demonstrate that including test cases leads to higher success in solving programming challenges. We assert that TDD is a promising paradigm for helping ensure that the code generated by LLMs effectively captures the requirements.

The Laser Interferometer Space Antenna (LISA) is the first scientific endeavour to detect and study gravitational waves from space. LISA will survey the sky for Gravitational Waves in the 0.1 mHz to 1 Hz frequency band which will enable the study of a vast number of objects ranging from Galactic binaries and stellar mass black holes in the Milky Way, to distant massive black-hole mergers and the expansion of the Universe. This definition study report, or Red Book, presents a summary of the very large body of work that has been undertaken on the LISA mission over the LISA definition phase., Comment: 155 pages, with executive summary and table of contents

End-to-End (E2E) unrolled optimization frameworks show promise for Magnetic Resonance (MR) image recovery, but suffer from high memory usage during training. In addition, these deterministic approaches do not offer opportunities for sampling from the posterior distribution. In this paper, we introduce a memory-efficient approach for E2E learning of the posterior distribution. We represent this distribution as the combination of a data-consistency-induced likelihood term and an energy model for the prior, parameterized by a Convolutional Neural Network (CNN). The CNN weights are learned from training data in an E2E fashion using maximum likelihood optimization. The learned model enables the recovery of images from undersampled measurements using the Maximum A Posteriori (MAP) optimization. In addition, the posterior model can be sampled to derive uncertainty maps about the reconstruction. Experiments on parallel MR image reconstruction show that our approach performs comparable to the memory-intensive E2E unrolled algorithm, performs better than its memory-efficient counterpart, and can provide uncertainty maps. Our framework paves the way towards MR image reconstruction in 3D and higher dimensions

Manual confirmation of static analysis reports is a daunting task. This is due to both the large number of warnings and the high density of false positives among them. Fuzzing techniques have been proposed to verify static analysis warnings. However, a major limitation is that fuzzing the whole project to reach all static analysis warnings is not feasible. This can take several days and exponential machine time to increase code coverage linearly. Therefore, we propose FuzzSlice, a novel framework that automatically prunes possible false positives among static analysis warnings. Unlike prior work that mostly focuses on confirming true positives among static analysis warnings, which requires end-to-end fuzzing, FuzzSlice focuses on ruling out potential false positives, which are the majority in static analysis reports. The key insight that we base our work on is that a warning that does not yield a crash when fuzzed at the function level in a given time budget is a possible false positive. To achieve this, FuzzSlice first aims to generate compilable code slices at the function level and then fuzzes these code slices instead of the entire binary. FuzzSlice is also unlikely to misclassify a true bug as a false positive because the crashing input can be reproduced by a fuzzer at the function level as well. We evaluate FuzzSlice on the Juliet synthetic dataset and real-world complex C projects. Our evaluation shows that the ground truth in the Juliet dataset had 864 false positives which were all detected by FuzzSlice. For the open-source repositories, we were able to get the developers from two of these open-source repositories to independently label these warnings. FuzzSlice automatically identifies 33 out of 53 false positives confirmed by developers in these two repositories. Thus FuzzSlice reduces false positives by 62.26% in the open-source repositories and by 100% in the Juliet dataset., Comment: The paper has been accepted for publication at ICSE 2024 (Research Track)

Despite the continued research and progress in building secure systems, Android applications continue to be ridden with vulnerabilities, necessitating effective detection methods. Current strategies involving static and dynamic analysis tools come with limitations like overwhelming number of false positives and limited scope of analysis which make either difficult to adopt. Over the past years, machine learning based approaches have been extensively explored for vulnerability detection, but its real-world applicability is constrained by data requirements and feature engineering challenges. Large Language Models (LLMs), with their vast parameters, have shown tremendous potential in understanding semnatics in human as well as programming languages. We dive into the efficacy of LLMs for detecting vulnerabilities in the context of Android security. We focus on building an AI-driven workflow to assist developers in identifying and rectifying vulnerabilities. Our experiments show that LLMs outperform our expectations in finding issues within applications correctly flagging insecure apps in 91.67% of cases in the Ghera benchmark. We use inferences from our experiments towards building a robust and actionable vulnerability detection system and demonstrate its effectiveness. Our experiments also shed light on how different various simple configurations can affect the True Positive (TP) and False Positive (FP) rates., Comment: This project report was presented as a part of the course CS858 at the University of Waterloo under the supervision of Prof. Yousra Aafer

The tasks of designing RNAs are discrete optimization problems, and several versions of these problems are NP-hard. As an alternative to commonly used local search methods, we formulate these problems as continuous optimization and develop a general framework for this optimization based on a generalization of classical partition function which we call "expected partition function". The basic idea is to start with a distribution over all possible candidate sequences, and extend the objective function from a sequence to a distribution. We then use gradient descent-based optimization methods to improve the extended objective function, and the distribution will gradually shrink towards a one-hot sequence (i.e., a single sequence). As a case study, we consider the important problem of mRNA design with wide applications in vaccines and therapeutics. While the recent work of LinearDesign can efficiently optimize mRNAs for minimum free energy (MFE), optimizing for ensemble free energy is much harder and likely intractable. Our approach can consistently improve over the LinearDesign solution in terms of ensemble free energy, with bigger improvements on longer sequences.