Search

This paper presents LOLA, a massively multilingual large language model trained on more than 160 languages using a sparse Mixture-of-Experts Transformer architecture. Our architectural and implementation choices address the challenge of harnessing linguistic diversity while maintaining efficiency and avoiding the common pitfalls of multilinguality. Our analysis of the evaluation results shows competitive performance in natural language generation and understanding tasks. Additionally, we demonstrate how the learned expert-routing mechanism exploits implicit phylogenetic linguistic patterns to potentially alleviate the curse of multilinguality. We provide an in-depth look at the training process, an analysis of the datasets, and a balanced exploration of the model's strengths and limitations. As an open-source model, LOLA promotes reproducibility and serves as a robust foundation for future research. Our findings enable the development of compute-efficient multilingual models with strong, scalable performance across languages.

Two-dimensional (2D) semiconductors have emerged as leading candidates for the development of low-power and multifunctional computing applications, thanks to their qualities such as layer-dependent band gap tunability, high carrier mobility, and excellent electrostatic control. Here, we explore a pair of 2D semiconductors with broken-gap (Type III) band alignment and demonstrate a highly gate-tunable p-MoTe$_{2}$/n-SnS$_{2}$ heterojunction tunnel field-effect transistor with multifunctional behavior. Employing a dual-gated asymmetric device geometry, we unveil its functionality as both a forward and backward rectifying device. Consequently, we observe a highly gate-tunable negative differential resistance (NDR), with a gate-coupling efficiency of $\eta \simeq 0.5$ and a peak-to-valley ratio of $\sim$ 3 down to 150K. By employing density functional theory and exploring the density of states, we determine that interband tunneling within the valence bands is the cause of the observed NDR characteristics. The combination of band-to-band tunneling and gate controllability of NDR signal open the pathway for realizing gate-tunable 2D material-based neuromorphic and energy-efficient electronics., Comment: 22 pages, 5 figures

Let $\eta_{1}\ge \eta_{2}\ge\cdots\ge \eta_{n}$ be the eigenavalues of $\mathcal{ABS}$ matrix. In this paper, we characterize connected graphs with $\mathcal{ABS}$ eigenvalue $\eta_{n}>-1$. As a result, we determine all connected graphs with exactly two distinct $\mathcal{ABS}$ eigenvalues. We show that a connected bipartite graph has three distinct $\mathcal{ABS}$ eigenvalues if and only if it is a complete bipartite graph. Furthermore, we present some bounds for the $\mathcal{ABS}$ spectral radius (resp. $\mathcal{ABS}$ energy) and characterize extremal graphs. Also, we obtain a relation between $\mathcal{ABC}$ energy and $\mathcal{ABS}$ energy. Finally, the chemical importance of $\mathcal{ABS}$ energy is investigated and it shown that the $\mathcal{ABS}$ energy is useful in predicting certain properties of molecules.

Dental disease is a prevalent chronic condition associated with substantial financial burden, personal suffering, and increased risk of systemic diseases. Despite widespread recommendations for twice-daily tooth brushing, adherence to recommended oral self-care behaviors remains sub-optimal due to factors such as forgetfulness and disengagement. To address this, we developed Oralytics, a mHealth intervention system designed to complement clinician-delivered preventative care for marginalized individuals at risk for dental disease. Oralytics incorporates an online reinforcement learning algorithm to determine optimal times to deliver intervention prompts that encourage oral self-care behaviors. We have deployed Oralytics in a registered clinical trial. The deployment required careful design to manage challenges specific to the clinical trials setting in the U.S. In this paper, we (1) highlight key design decisions of the RL algorithm that address these challenges and (2) conduct a re-sampling analysis to evaluate algorithm design decisions. A second phase (randomized control trial) of Oralytics is planned to start in spring 2025.

Online AI decision-making algorithms are increasingly used by digital interventions to dynamically personalize treatment to individuals. These algorithms determine, in real-time, the delivery of treatment based on accruing data. The objective of this paper is to provide guidelines for enabling effective monitoring of online decision-making algorithms with the goal of (1) safeguarding individuals and (2) ensuring data quality. We elucidate guidelines and discuss our experience in monitoring online decision-making algorithms in two digital intervention clinical trials (Oralytics and MiWaves). Our guidelines include (1) developing fallback methods, pre-specified procedures executed when an issue occurs, and (2) identifying potential issues categorizing them by severity (red, yellow, and green). Across both trials, the monitoring systems detected real-time issues such as out-of-memory issues, database timeout, and failed communication with an external source. Fallback methods prevented participants from not receiving any treatment during the trial and also prevented the use of incorrect data in statistical analyses. These trials provide case studies for how health scientists can build monitoring systems for their digital intervention. Without these algorithm monitoring systems, critical issues would have gone undetected and unresolved. Instead, these monitoring systems safeguarded participants and ensured the quality of the resulting data for updating the intervention and facilitating scientific discovery. These monitoring guidelines and findings give digital intervention teams the confidence to include online decision-making algorithms in digital interventions.

Embeddings-as-a-Service (EaaS) is a service offered by large language model (LLM) developers to supply embeddings generated by LLMs. Previous research suggests that EaaS is prone to imitation attacks -- attacks that clone the underlying EaaS model by training another model on the queried embeddings. As a result, EaaS watermarks are introduced to protect the intellectual property of EaaS providers. In this paper, we first show that existing EaaS watermarks can be removed by paraphrasing when attackers clone the model. Subsequently, we propose a novel watermarking technique that involves linearly transforming the embeddings, and show that it is empirically and theoretically robust against paraphrasing., Comment: Work in Progress

Fast and accurate detection of light in the near-infrared (NIR) spectral range plays a crucial role in modern society, from alleviating speed and capacity bottlenecks in optical communications to enhancing the control and safety of autonomous vehicles through NIR imaging systems. Several technological platforms are currently under investigation to improve NIR photodetection, aiming to surpass the performance of established III-V semiconductor p-i-n (PIN) junction technology. These platforms include in situ-grown inorganic nanocrystals and nanowire arrays, as well as hybrid organic-inorganic materials such as graphene-perovskite heterostructures. However, challenges remain in nanocrystal and nanowire growth, large-area fabrication of high-quality 2D materials, and the fabrication of devices for practical applications. Here, we explore the potential for tailored semiconductor nanocrystals to enhance the responsivity of planar metal-semiconductor-metal (MSM) photodetectors. MSM technology offers ease of fabrication and fast response times compared to PIN detectors. We observe enhancement of the optical-to-electric conversion efficiency by up to a factor of ~2.5 through the application of plasmonically-active semiconductor nanorods and nanocrystals. We present a protocol for synthesizing and rapidly testing the performance of non-stoichiometric tungsten oxide (WO$_{3-x}$) nanorods and cesium-doped tungsten oxide (Cs$_y$WO$_{3-x}$) hexagonal nanoprisms prepared in colloidal suspensions and drop-cast onto photodetector surfaces. The results demonstrate the potential for a cost-effective and scalable method exploiting tailored nanocrystals to improve the performance of NIR optoelectronic devices.

Blockchain-based games have introduced novel economic models that blend traditional gaming with decentralized ownership and financial incentives, leading to the rapid emergence of the GameFi sector. However, despite their innovative appeal, these games face significant challenges, particularly in terms of market stability, player retention, and the sustainability of token value. This paper explores the evolution of blockchain games and identifies key shortcomings in current tokenomics models using entropy increase theory. We propose two new models - ServerFi, which emphasizes Privatization through Asset Synthesis, and a model focused on Continuous Rewards for High-Retention Players. These models are formalized into mathematical frameworks and validated through group behavior simulation experiments. Our findings indicate that the ServerFi is particularly effective in maintaining player engagement and ensuring the long-term viability of the gaming ecosystem, offering a promising direction for future blockchain game development.

Over the last few years, federated learning (FL) has emerged as a prominent method in machine learning, emphasizing privacy preservation by allowing multiple clients to collaboratively build a model while keeping their training data private. Despite this focus on privacy, FL models are susceptible to various attacks, including membership inference attacks (MIAs), posing a serious threat to data confidentiality. In a recent study, Rezaei \textit{et al.} revealed the existence of an accuracy-privacy trade-off in deep ensembles and proposed a few fusion strategies to overcome it. In this paper, we aim to explore the relationship between deep ensembles and FL. Specifically, we investigate whether confidence-based metrics derived from deep ensembles apply to FL and whether there is a trade-off between accuracy and privacy in FL with respect to MIA. Empirical investigations illustrate a lack of a non-monotonic correlation between the number of clients and the accuracy-privacy trade-off. By experimenting with different numbers of federated clients, datasets, and confidence-metric-based fusion strategies, we identify and analytically justify the clear existence of the accuracy-privacy trade-off.

The rapid growth in the use of Large Language Models (LLMs) and AI Agents as part of software development and deployment is revolutionizing the information technology landscape. While code generation receives significant attention, a higher-impact application lies in using AI agents for operational resilience of cloud services, which currently require significant human effort and domain knowledge. There is a growing interest in AI for IT Operations (AIOps) which aims to automate complex operational tasks, like fault localization and root cause analysis, thereby reducing human intervention and customer impact. However, achieving the vision of autonomous and self-healing clouds through AIOps is hampered by the lack of standardized frameworks for building, evaluating, and improving AIOps agents. This vision paper lays the groundwork for such a framework by first framing the requirements and then discussing design decisions that satisfy them. We also propose AIOpsLab, a prototype implementation leveraging agent-cloud-interface that orchestrates an application, injects real-time faults using chaos engineering, and interfaces with an agent to localize and resolve the faults. We report promising results and lay the groundwork to build a modular and robust framework for building, evaluating, and improving agents for autonomous clouds.

Microscopic interpretation of histopathology images underlies many important diagnostic and treatment decisions. While advances in vision-language modeling raise new opportunities for analysis of such images, the gigapixel-scale size of whole slide images (WSIs) introduces unique challenges. Additionally, pathology reports simultaneously highlight key findings from small regions while also aggregating interpretation across multiple slides, often making it difficult to create robust image-text pairs. As such, pathology reports remain a largely untapped source of supervision in computational pathology, with most efforts relying on region-of-interest annotations or self-supervision at the patch-level. In this work, we develop a vision-language model based on the BLIP-2 framework using WSIs paired with curated text from pathology reports. This enables applications utilizing a shared image-text embedding space, such as text or image retrieval for finding cases of interest, as well as integration of the WSI encoder with a frozen large language model (LLM) for WSI-based generative text capabilities such as report generation or AI-in-the-loop interactions. We utilize a de-identified dataset of over 350,000 WSIs and diagnostic text pairs, spanning a wide range of diagnoses, procedure types, and tissue types. We present pathologist evaluation of text generation and text retrieval using WSI embeddings, as well as results for WSI classification and workflow prioritization (slide-level triaging). Model-generated text for WSIs was rated by pathologists as accurate, without clinically significant error or omission, for 78% of WSIs on average. This work demonstrates exciting potential capabilities for language-aligned WSI embeddings., Comment: 9 main pages and 19 pages of supplemental material; 3 main tables, 3 main figures and 11 supplemental tables, 7 supplemental figures

Large Language Models(LLMs) have shown exceptional abilities, yet training these models can be quite challenging. There is a strong dependence on the quality of data and finding the best instruction tuning set. Further, the inherent limitations in training methods create substantial difficulties to train relatively smaller models with 7B and 13B parameters. In our research, we suggest an improved training method for these models by utilising knowledge from larger models, such as a mixture of experts (8x7B) architectures. The scale of these larger models allows them to capture a wide range of variations from data alone, making them effective teachers for smaller models. Moreover, we implement a novel post-training domain alignment phase that employs domain-specific expert models to boost domain-specific knowledge during training while preserving the model's ability to generalise. Fine-tuning Mistral 7B and 2x7B with our method surpasses the performance of state-of-the-art language models with more than 7B and 13B parameters: achieving up to $7.9$ in MT-Bench and $93.04\%$ on AlpacaEval., Comment: 7 pages, 4 figures

Dental disease is still one of the most common chronic diseases in the United States. While dental disease is preventable through healthy oral self-care behaviors (OSCB), this basic behavior is not consistently practiced. We have developed Oralytics, an online, reinforcement learning (RL) algorithm that optimizes the delivery of personalized intervention prompts to improve OSCB. In this paper, we offer a full overview of algorithm design decisions made using prior data, domain expertise, and experiments in a simulation test bed. The finalized RL algorithm was deployed in the Oralytics clinical trial, conducted from fall 2023 to summer 2024.

Matching a target spectrum with known spectra in a spectral library is a common method for material identification in hyperspectral imaging research. Hyperspectral spectra exhibit precise absorption features across different wavelength segments, and the unique shapes and positions of these absorptions create distinct spectral signatures for each material, aiding in their identification. Therefore, only the specific positions can be considered for material identification. This study introduces the Weighted Sum of Segmented Correlation method, which calculates correlation indices between various segments of a library and a test spectrum, and derives a matching index, favoring positive correlations and penalizing negative correlations using assigned weights. The effectiveness of this approach is evaluated for mineral identification in hyperspectral images from both Earth and Martian surfaces., Comment: Accepted in IEEE IGARSS 2024 conference

In health, most large language model (LLM) research has focused on clinical tasks. However, mobile and wearable devices, which are rarely integrated into such tasks, provide rich, longitudinal data for personal health monitoring. Here we present Personal Health Large Language Model (PH-LLM), fine-tuned from Gemini for understanding and reasoning over numerical time-series personal health data. We created and curated three datasets that test 1) production of personalized insights and recommendations from sleep patterns, physical activity, and physiological responses, 2) expert domain knowledge, and 3) prediction of self-reported sleep outcomes. For the first task we designed 857 case studies in collaboration with domain experts to assess real-world scenarios in sleep and fitness. Through comprehensive evaluation of domain-specific rubrics, we observed that Gemini Ultra 1.0 and PH-LLM are not statistically different from expert performance in fitness and, while experts remain superior for sleep, fine-tuning PH-LLM provided significant improvements in using relevant domain knowledge and personalizing information for sleep insights. We evaluated PH-LLM domain knowledge using multiple choice sleep medicine and fitness examinations. PH-LLM achieved 79% on sleep and 88% on fitness, exceeding average scores from a sample of human experts. Finally, we trained PH-LLM to predict self-reported sleep quality outcomes from textual and multimodal encoding representations of wearable data, and demonstrate that multimodal encoding is required to match performance of specialized discriminative models. Although further development and evaluation are necessary in the safety-critical personal health domain, these results demonstrate both the broad knowledge and capabilities of Gemini models and the benefit of contextualizing physiological data for personal health applications as done with PH-LLM., Comment: 72 pages

The quantum-classical hybrid variational quantum eigensolver (VQE) algorithm is among the most actively studied topics in atomic and molecular calculations on quantum computers, yet few studies address properties other than energies or account for relativistic effects. This work presents high-precision 18-qubit relativistic VQE simulations for calculating the permanent electric dipole moments (PDMs) of BeH to RaH molecules on traditional computers, and 6- and 12-qubit PDM computations for SrH on IonQ quantum devices. To achieve high precision on current noisy intermediate scale era quantum hardware, we apply various resource reduction methods, including Reinforcement Learning and causal flow preserving ZX-Calculus routines, along with error mitigation and post-selection techniques. Our approach reduces the two-qubit gate count in our 12-qubit circuit by 99.71%, with only a 2.35% trade-off in precision for PDM when evaluated classically within a suitably chosen active space. On the current generation IonQ Forte-I hardware, the error in PDM is -1.17% relative to classical calculations and only 1.21% compared to the unoptimized circuit.

We study the problem of learning under arbitrary distribution shift, where the learner is trained on a labeled set from one distribution but evaluated on a different, potentially adversarially generated test distribution. We focus on two frameworks: PQ learning [Goldwasser, A. Kalai, Y. Kalai, Montasser NeurIPS 2020], allowing abstention on adversarially generated parts of the test distribution, and TDS learning [Klivans, Stavropoulos, Vasilyan COLT 2024], permitting abstention on the entire test distribution if distribution shift is detected. All prior known algorithms either rely on learning primitives that are computationally hard even for simple function classes, or end up abstaining entirely even in the presence of a tiny amount of distribution shift. We address both these challenges for natural function classes, including intersections of halfspaces and decision trees, and standard training distributions, including Gaussians. For PQ learning, we give efficient learning algorithms, while for TDS learning, our algorithms can tolerate moderate amounts of distribution shift. At the core of our approach is an improved analysis of spectral outlier-removal techniques from learning with nasty noise. Our analysis can (1) handle arbitrarily large fraction of outliers, which is crucial for handling arbitrary distribution shifts, and (2) obtain stronger bounds on polynomial moments of the distribution after outlier removal, yielding new insights into polynomial regression under distribution shifts. Lastly, our techniques lead to novel results for tolerant testable learning [Rubinfeld and Vasilyan STOC 2023], and learning with nasty noise.

Popular parameter-efficient fine-tuning (PEFT) methods, such as LoRA and its variants, freeze pre-trained model weights \(W\) and inject learnable matrices \(\Delta W\). These \(\Delta W\) matrices are structured for efficient parameterization, often using techniques like low-rank approximations or scaling vectors. However, these methods typically show a performance gap compared to full fine-tuning. Although recent PEFT methods have narrowed this gap, they do so at the cost of additional learnable parameters. We propose SVFT, a simple approach that fundamentally differs from existing methods: the structure imposed on \(\Delta W\) depends on the specific weight matrix \(W\). Specifically, SVFT updates \(W\) as a sparse combination of outer products of its singular vectors, training only the coefficients (scales) of these sparse combinations. This approach allows fine-grained control over expressivity through the number of coefficients. Extensive experiments on language and vision benchmarks show that SVFT recovers up to 96% of full fine-tuning performance while training only 0.006 to 0.25% of parameters, outperforming existing methods that only recover up to 85% performance using 0.03 to 0.8% of the trainable parameter budget., Comment: 17 pages, 5 figures, 14 tables

Recent advances in robot skill learning have unlocked the potential to construct task-agnostic skill libraries, facilitating the seamless sequencing of multiple simple manipulation primitives (aka. skills) to tackle significantly more complex tasks. Nevertheless, determining the optimal sequence for independently learned skills remains an open problem, particularly when the objective is given solely in terms of the final geometric configuration rather than a symbolic goal. To address this challenge, we propose Logic-Skill Programming (LSP), an optimization-based approach that sequences independently learned skills to solve long-horizon tasks. We formulate a first-order extension of a mathematical program to optimize the overall cumulative reward of all skills within a plan, abstracted by the sum of value functions. To solve such programs, we leverage the use of tensor train factorization to construct the value function space, and rely on alternations between symbolic search and skill value optimization to find the appropriate skill skeleton and optimal subgoal sequence. Experimental results indicate that the obtained value functions provide a superior approximation of cumulative rewards compared to state-of-the-art reinforcement learning methods. Furthermore, we validate LSP in three manipulation domains, encompassing both prehensile and non-prehensile primitives. The results demonstrate its capability to identify the optimal solution over the full logic and geometric path. The real-robot experiments showcase the effectiveness of our approach to cope with contact uncertainty and external disturbances in the real world., Comment: In Proc. Robotics: Science and Systems (RSS), 2024

Many clinical tasks require an understanding of specialized data, such as medical images and genomics, which is not typically found in general-purpose large multimodal models. Building upon Gemini's multimodal models, we develop several models within the new Med-Gemini family that inherit core capabilities of Gemini and are optimized for medical use via fine-tuning with 2D and 3D radiology, histopathology, ophthalmology, dermatology and genomic data. Med-Gemini-2D sets a new standard for AI-based chest X-ray (CXR) report generation based on expert evaluation, exceeding previous best results across two separate datasets by an absolute margin of 1% and 12%, where 57% and 96% of AI reports on normal cases, and 43% and 65% on abnormal cases, are evaluated as "equivalent or better" than the original radiologists' reports. We demonstrate the first ever large multimodal model-based report generation for 3D computed tomography (CT) volumes using Med-Gemini-3D, with 53% of AI reports considered clinically acceptable, although additional research is needed to meet expert radiologist reporting quality. Beyond report generation, Med-Gemini-2D surpasses the previous best performance in CXR visual question answering (VQA) and performs well in CXR classification and radiology VQA, exceeding SoTA or baselines on 17 of 20 tasks. In histopathology, ophthalmology, and dermatology image classification, Med-Gemini-2D surpasses baselines across 18 out of 20 tasks and approaches task-specific model performance. Beyond imaging, Med-Gemini-Polygenic outperforms the standard linear polygenic risk score-based approach for disease risk prediction and generalizes to genetically correlated diseases for which it has never been trained. Although further development and evaluation are necessary in the safety-critical medical domain, our results highlight the potential of Med-Gemini across a wide range of medical tasks.

We measure the age-velocity relationship from the lag between ionized gas and stellar tangential speeds in ~500 nearby disk galaxies from MaNGA in SDSS-IV. Selected galaxies are kinematically axisymmetric. Velocity lags are asymmetric drift, seen in the Milky Way's (MW) solar neighborhood and other Local Group galaxies; their amplitude correlates with stellar population age. The trend is qualitatively consistent in rate (d(sigma)/dt) with a simple power-law model where sigma is proportional to t^b that explains the dynamical phase-space stratification in the solar neighborhood. The model is generalized based on disk dynamical times to other radii and other galaxies. We find in-plane radial stratification parameters sigma_(0,r} (dispersion of the youngest populations) in the range of 10-40 km/s and 0.2

Systems driven far from equilibrium often retain structural memories of their processing history. This memory has, in some cases, been shown to dramatically alter the material response. For example, work hardening in crystalline metals can alter the hardness, yield strength, and tensile strength to prevent catastrophic failure. Whether memory of processing history can be similarly exploited in flowing systems, where significantly larger changes in structure should be possible, remains poorly understood. Here, we demonstrate a promising route to embedding such useful memories. We build on work showing that exposing a sheared dense suspension to acoustic perturbations of different power allows for dramatically tuning the sheared suspension viscosity and underlying structure. We find that, for sufficiently dense suspensions, upon removing the acoustic perturbations, the suspension shear jams with shear stress contributions from the maximum compressive and maximum extensive axes that reflect the acoustic training. Because the contributions from these two orthogonal axes to the total shear stress are antagonistic, it is possible to tune the resulting suspension response in surprising ways. For example, we show that differently trained sheared suspensions exhibit: 1) different susceptibility to the same acoustic perturbation; 2) orders of magnitude changes in their instantaneous viscosities upon shear reversal; and 3) even a shear stress that increases in magnitude upon shear cessation. To further illustrate the power of this approach for controlling suspension properties, we demonstrate that flowing states well below the shear jamming threshold can be shear jammed via acoustic training. Collectively, our work paves the way for using acoustically induced memory in dense suspensions to generate rapidly and widely tunable materials., Comment: To be published in Physical Review X

Understanding the potential of generative AI (GenAI)-based attacks on the power grid is a fundamental challenge that must be addressed in order to protect the power grid by realizing and validating risk in new attack vectors. In this paper, a novel zero trust framework for a power grid supply chain (PGSC) is proposed. This framework facilitates early detection of potential GenAI-driven attack vectors (e.g., replay and protocol-type attacks), assessment of tail risk-based stability measures, and mitigation of such threats. First, a new zero trust system model of PGSC is designed and formulated as a zero-trust problem that seeks to guarantee for a stable PGSC by realizing and defending against GenAI-driven cyber attacks. Second, in which a domain-specific generative adversarial networks (GAN)-based attack generation mechanism is developed to create a new vulnerability cyberspace for further understanding that threat. Third, tail-based risk realization metrics are developed and implemented for quantifying the extreme risk of a potential attack while leveraging a trust measurement approach for continuous validation. Fourth, an ensemble learning-based bootstrap aggregation scheme is devised to detect the attacks that are generating synthetic identities with convincing user and distributed energy resources device profiles. Experimental results show the efficacy of the proposed zero trust framework that achieves an accuracy of 95.7% on attack vector generation, a risk measure of 9.61% for a 95% stable PGSC, and a 99% confidence in defense against GenAI-driven attack., Comment: Accepted article by IEEE International Conference on Communications (ICC 2024), Copyright 2024 IEEE

Health acoustic sounds such as coughs and breaths are known to contain useful health signals with significant potential for monitoring health and disease, yet are underexplored in the medical machine learning community. The existing deep learning systems for health acoustics are often narrowly trained and evaluated on a single task, which is limited by data and may hinder generalization to other tasks. To mitigate these gaps, we develop HeAR, a scalable self-supervised learning-based deep learning system using masked autoencoders trained on a large dataset of 313 million two-second long audio clips. Through linear probes, we establish HeAR as a state-of-the-art health audio embedding model on a benchmark of 33 health acoustic tasks across 6 datasets. By introducing this work, we hope to enable and accelerate further health acoustics research., Comment: 4 tables, 4 figures, 6 supplementary tables, 3 supplementary figures

Embedding as a Service (EaaS) has become a widely adopted solution, which offers feature extraction capabilities for addressing various downstream tasks in Natural Language Processing (NLP). Prior studies have shown that EaaS can be prone to model extraction attacks; nevertheless, this concern could be mitigated by adding backdoor watermarks to the text embeddings and subsequently verifying the attack models post-publication. Through the analysis of the recent watermarking strategy for EaaS, EmbMarker, we design a novel CSE (Clustering, Selection, Elimination) attack that removes the backdoor watermark while maintaining the high utility of embeddings, indicating that the previous watermarking approach can be breached. In response to this new threat, we propose a new protocol to make the removal of watermarks more challenging by incorporating multiple possible watermark directions. Our defense approach, WARDEN, notably increases the stealthiness of watermarks and has been empirically shown to be effective against CSE attack., Comment: Accepted to ACL2024 (Main Proceedings)

We present a simulation of various active learning strategies for the discovery of polymer solar cell donor/acceptor pairs using data extracted from the literature spanning $\sim$20 years by a natural language processing pipeline. While data-driven methods have been well established to discover novel materials faster than Edisonian trial-and-error approaches, their benefits have not been quantified for material discovery problems that can take decades. Our approach demonstrates a potential reduction in discovery time by approximately 75 %, equivalent to a 15 year acceleration in material innovation. Our pipeline enables us to extract data from greater than 3300 papers which is $\sim$5 times larger and therefore more diverse than similar data sets reported by others. We also trained machine learning models to predict the power conversion efficiency and used our model to identify promising donor-acceptor combinations that are as yet unreported. We thus demonstrate a pipeline that goes from published literature to extracted material property data which in turn is used to obtain data-driven insights. Our insights include active learning strategies that can be used to train strong predictive models of material properties or be robust to the initial material system used. This work provides a valuable framework for data-driven research in materials science.

We report the first-ever complete measurement of MAOStress material functions, which reveal that stress can be more fundamental than strain or strain rate for understanding linearity limits as a function of Deborah number. The material used is a canonical viscoelastic liquid with a single dominant relaxation time: Polyvinyl alcohol (PVA) polymer solution crosslinked with tetrahydroborate (Borax) solution. We outline experimental limit lines and their dependence on geometry and test conditions. These MAOStress measurements enable us to observe the frequency dependence of the weakly nonlinear deviation as a function of stress amplitude. The observed features of MAOStress material functions are distinctly simpler than MAOStrain, where the frequency dependence is much more dramatic. The strain-stiffening transient network model (SSTNM) was used to derive a model-informed normalization of the nonlinear material functions that accounts for their scaling with the linear material properties. Moreover, we compare the frequency-dependence of the critical stress, strain, and strain-rate for the linearity limit, which are rigorously computed from the MAOStress and MAOStrain material functions. While critical strain and strain-rate change by orders of magnitude throughout the Deborah number range, the critical stress changes by a factor of about two, showing that stress is a more fundamental measure of nonlinearity strength. This work extends the experimental accessibility of the weakly nonlinear regime to stress-controlled instruments and deformations, which reveal material physics beyond linear viscoelasticity but at conditions that are accessible to theory and detailed simulation.

Online reinforcement learning (RL) algorithms offer great potential for personalizing treatment for participants in clinical trials. However, deploying an online, autonomous algorithm in the high-stakes healthcare setting makes quality control and data quality especially difficult to achieve. This paper proposes algorithm fidelity as a critical requirement for deploying online RL algorithms in clinical trials. It emphasizes the responsibility of the algorithm to (1) safeguard participants and (2) preserve the scientific utility of the data for post-trial analyses. We also present a framework for pre-deployment planning and real-time monitoring to help algorithm developers and clinical researchers ensure algorithm fidelity. To illustrate our framework's practical application, we present real-world examples from the Oralytics clinical trial. Since Spring 2023, this trial successfully deployed an autonomous, online RL algorithm to personalize behavioral interventions for participants at risk for dental disease.

In order to circumvent statistical and computational hardness results in sequential decision-making, recent work has considered smoothed online learning, where the distribution of data at each time is assumed to have bounded likeliehood ratio with respect to a base measure when conditioned on the history. While previous works have demonstrated the benefits of smoothness, they have either assumed that the base measure is known to the learner or have presented computationally inefficient algorithms applying only in special cases. This work investigates the more general setting where the base measure is \emph{unknown} to the learner, focusing in particular on the performance of Empirical Risk Minimization (ERM) with square loss when the data are well-specified and smooth. We show that in this setting, ERM is able to achieve sublinear error whenever a class is learnable with iid data; in particular, ERM achieves error scaling as $\tilde O( \sqrt{\mathrm{comp}(\mathcal F)\cdot T} )$, where $\mathrm{comp}(\mathcal F)$ is the statistical complexity of learning $\mathcal F$ with iid data. In so doing, we prove a novel norm comparison bound for smoothed data that comprises the first sharp norm comparison for dependent data applying to arbitrary, nonlinear function classes. We complement these results with a lower bound indicating that our analysis of ERM is essentially tight, establishing a separation in the performance of ERM between smoothed and iid data.

This follow-on work investigates the effect of temperature on the frequency limit of 2-dimensional electron gas (2DEG) Hall-effect sensors., Comment: 4 pages

Due to statistical lower bounds on the learnability of many function classes under privacy constraints, there has been recent interest in leveraging public data to improve the performance of private learning algorithms. In this model, algorithms must always guarantee differential privacy with respect to the private samples while also ensuring learning guarantees when the private data distribution is sufficiently close to that of the public data. Previous work has demonstrated that when sufficient public, unlabelled data is available, private learning can be made statistically tractable, but the resulting algorithms have all been computationally inefficient. In this work, we present the first computationally efficient, algorithms to provably leverage public data to learn privately whenever a function class is learnable non-privately, where our notion of computational efficiency is with respect to the number of calls to an optimization oracle for the function class. In addition to this general result, we provide specialized algorithms with improved sample complexities in the special cases when the function class is convex or when the task is binary classification.

We have re-observed $\rm\sim$40 low-inclination, star-forming galaxies from the MaNGA survey ($\upsigma\sim65$~\kms) at $\sim$6.5 times higher spectral resolution ($\upsigma\sim10$~\kms) using the HexPak integral field unit on the WIYN 3.5m telescope. The aim of these observations is to calibrate MaNGA's instrumental resolution and to characterize turbulence in the warm interstellar medium and ionized galactic outflows. Here we report the results for the H$\rm\upalpha$ region observations as they pertain to the calibration of MaNGA's spectral resolution. Remarkably, we find that the previously-reported MaNGA line-spread-function (LSF) Gaussian width is systematically underestimated by only 1\%. The LSF increase modestly reduces the characteristic dispersion of HII regions-dominated spectra sampled at 1-2 kpc spatial scales from 23 to 20 km s$^{-1}$ in our sample, or a 25\% decrease in the random-motion kinetic energy. This commensurately lowers the dispersion zeropoint in the relation between line-width and star-formation rate surface-density in galaxies sampled on the same spatial scale. This modest zero-point shift does not appear to alter the power-law slope in the relation between line-width and star-formation rate surface-density. We also show that adopting a scheme whereby corrected line-widths are computed as the square root of the median of the difference in the squared measured line width and the squared LSF Gaussian avoids biases and allows for lower SNR data to be used reliably.

Consider the supervised learning setting where the goal is to learn to predict labels $\mathbf y$ given points $\mathbf x$ from a distribution. An \textit{omnipredictor} for a class $\mathcal L$ of loss functions and a class $\mathcal C$ of hypotheses is a predictor whose predictions incur less expected loss than the best hypothesis in $\mathcal C$ for every loss in $\mathcal L$. Since the work of [GKR+21] that introduced the notion, there has been a large body of work in the setting of binary labels where $\mathbf y \in \{0, 1\}$, but much less is known about the regression setting where $\mathbf y \in [0,1]$ can be continuous. Our main conceptual contribution is the notion of \textit{sufficient statistics} for loss minimization over a family of loss functions: these are a set of statistics about a distribution such that knowing them allows one to take actions that minimize the expected loss for any loss in the family. The notion of sufficient statistics relates directly to the approximate rank of the family of loss functions. Our key technical contribution is a bound of $O(1/\varepsilon^{2/3})$ on the $\epsilon$-approximate rank of convex, Lipschitz functions on the interval $[0,1]$, which we show is tight up to a factor of $\mathrm{polylog} (1/\epsilon)$. This yields improved runtimes for learning omnipredictors for the class of all convex, Lipschitz loss functions under weak learnability assumptions about the class $\mathcal C$. We also give efficient omnipredictors when the loss families have low-degree polynomial approximations, or arise from generalized linear models (GLMs). This translation from sufficient statistics to faster omnipredictors is made possible by lifting the technique of loss outcome indistinguishability introduced by [GKH+23] for Boolean labels to the regression setting.

Neural networks have been employed for a wide range of processing applications like image processing, motor control, object detection and many others. Living neural networks offer advantages of lower power consumption, faster processing, and biological realism. Optogenetics offers high spatial and temporal control over biological neurons and presents potential in training live neural networks. This work proposes a simulated living neural network trained indirectly by backpropagating STDP based algorithms using precision activation by optogenetics achieving accuracy comparable to traditional neural network training algorithms.