12023 research outputs found
Sort by
Animal Regeneration and its Loss: the Mouse as a Model of Limited Regeneration
Get PDFIn this dissertation, we explore animal regeneration through a comparative evolutionary-developmental framework. In Chapter 1, we review animal regeneration and its loss through examining broad developmental and physiological factors that correlate with regenerative ability. We also highlight the mouse as a model of regeneration loss, examining the limited regeneration of the digit tip and the heart. For each context, we discuss how these regenerative processes occur, the physiological and molecular factors involved, and previous attempts to induce or improve the regenerative response.
In Chapter 2, we explore the possibility of inducing regeneration in non-regenerating systems. Along with experiments in the jellyfish Aurelia coerulea, (formerly A. aurita sp. 1 strain) and the fruit fly Drosophila melanogaster, we find that supplementation with the amino acid L-leucine and sucrose induce appendage regeneration across these highly evolutionarily-diverged organisms. We discuss how this intervention highlights the conserved role of energetic parameters in regeneration, and how surpassing nutrient-based limitations may unlock regenerative responses in diverse contexts.
In Chapter 3, we characterize the derivatives of cardiac neural crest cells (CNCCs) in the hearts of neonatal mice at P1 and >P7. We confirm previous work on the diverse cardiac derivatives resulting from CNCCs, and provide additional evidence for CNCC-derived cardiomyocytes, a contribution still contested in current literature. We also demonstrate how CNCC derivatives form a distinct age-related developmental trajectory in the heart, and discuss how these changes may affect cardiac physiology and relate to the loss of neonatal heart regeneration.
Finally, in Chapter 4, we propose future directions based on the work carried out in Chapter 3. While CNCCs have explicitly been studied in the context of heart regeneration in zebrafish, their role in neonatal mouse heart regeneration has not been explored. First, we suggest further investigation into the differences between CNCC-derived and non-CNCC derived CMs, to see if their proliferative ability and molecular profile show different temporal dynamics over the course of embryonic and early postnatal stages. Then, to examine all CNCC-derived cell types in regenerating P1 and non-regenerating P8 hearts, we propose a single-nuclei RNA-sequencing experiment to better resolve questions about how the myocardial lineage interacts with nonmyocytes, and to capture the full extent of potential cardiomyocyte decline postnatally. Lastly, to understand how CNCC derivatives influence endogenous heart regeneration, we design a dual Cre and (r)tTA driver system in transgenic mice to perform a conditional ablation experiment of CNCCs in a cryoinjury model of neonatal heart regeneration.</p
A Novel Approach to Cardiac Health Assessment Using a Redesign of the Brachial Cuff Device
Get PDFCurrent diagnostic methodologies in cardiology face large tradeoffs between procedure invasiveness and diagnostic reliability, ultimately requiring individuals to undergo cardiac catheterization for accurate diagnosis. Given the current societal burden of cardiovascular disease, there is a need for translational medical devices that bridge the accuracy gap between invasive and non-invasive measurements in the assessment of heart health. This thesis focuses on the development and validation of a high-resolution cuff-based system for assessment of central cardiovascular health.
Traditional pressure cuffs suffer low signal resolution when applied to non-invasive pulse waveform acquisition. In the first section of this thesis, we develop a cuff-based device with a pneumatic filter for high fidelity pulse waveform acquisition. This work discusses the design and functionality of the cuff-device, and investigates the repeatability of the cuff-based measurement. Furthermore, the derived mathematical model of the pneumatic filter is shown to have an equivalent behavior to an electrical low-pass filter inclusive of a time constant and a frequency response curve.
The accuracy and reliability of the pulse waveform features from the cuff-device are evaluated with human study data. Firstly, an IRB study is performed at Caltech on a young and healthy population showing that the cuff-device data lies within a narrow distribution indicative of the healthy nature of the population. Secondly, data from a clinical trial collecting simultaneous invasive catheter, cuff, and ECG is analyzed. The first analysis compared waveform parameters from the cuff in sSBP hold pressure with simultaneous aortic catheter, showing strong correlations between the two measurement modalities for both magnitude and fluctuations thereof.
Lastly, this work investigated the relationship between cuff-based parameters and left ventricular functions. We introduced a cuff-based method for extraction of the pressure-sound waveform, a pressure based surrogate of heart sounds. The results from this analysis showed that the pressure-sound features correlate with the strength of the left ventricular isovolumetric contraction and relaxation. Other important results from this work demonstrated the correlations between the heart-lung interactions in the left ventricle and cuff parameters: breathing fluctuations proportionally affect LV pressures and cuff sSBP waveform parameters. Overall these results support the accuracy and reliability of a cuff-based device for central cardiovascular health assessments.</p
Non-Asymptotic Analysis of Single-Receiver Channels with Limited Feedback
Get PDFEmerging Internet of Things, machine-type communication, and ultra-reliable low-latency communication in 5G demand codes that operate at short blocklengths, have low error probability and low energy consumption, and can handle the random activity of a large number of communicating devices. Since many of the applications have a single central device, e.g., a base station, that resolves the communication and a varying number of users, these requirements on the code design motivate interest in the non-asymptotic analysis of codes in a variety of single-receiver channels. This thesis investigates three channel coding problems with the goals of understanding the fundamental limits of channel coding under stringent requirements on reliability, delay, and power, and proposes novel coding architectures that employ constrained feedback to attain those limits. In the first part, we consider point-to-point channels without feedback, and analyze the non-asymptotic limits in the moderate deviations regime in probability theory. The moderate deviations regime is suitable for accurately approximating the maximum achievable coding rate in the operational regimes of practical interest because it simultaneously considers high rates and low error probabilities. We propose a new quantity, channel skewness, which governs the fundamental limit at short blocklengths and low error probabilities. Our approximation is the tightest among the state-of-the-art approximations for most error probability and latency constraints of interest. In the second part, we investigate rateless channel coding with limited feedback. Here, rateless means that decoding can occur at multiple decoding times. In our code design, feedback is limited both in frequency and content; it is sparse, meaning that it is available only at a few instants throughout the communication epoch; and it is stop-feedback, meaning that the receiver informs the transmitters only about whether decoding has occurred rather than what symbols it has received. Our results demonstrate that sporadically sending a few bits is almost as efficient as sending feedback at every time instant. In the third part, we focus on rateless random access channel codes, where the number of active transmitters is unknown to both the transmitters and the receiver. Our rateless code design that reserves a decoding time for each possible number of active transmitters achieves the same first two terms in the asymptotic expansion of the achievable rate as codes where the transmitter activity is known a priori. This means that, remarkably, the random transmitter activity has almost no effect on achievable rates.
To obtain tight channel coding bounds, we analyze some non-asymptotic and asymptotic state-of-the-art bounds on the probability of the sum of independent and identical random variables, whose applications extend to source coding, hypothesis testing, and many others. In the scenarios where these tools are not directly applicable such as for the Gaussian channel, we propose new techniques to overcome that difficulty.</p
Transient Behavior of Granular Material
Get PDFThis PhD thesis focuses on the flows on granular materials, such as sand, glass beads, and powders, which are sheared at low speeds with gravity perpendicular to the flow direction. The study is conducted using a combination of experiments, simulations, and theory, with the goal of developing a unifying theory of granular materials that can be described by continuum models. The main objective is to understand how microscale physics propagate to macroscale phenomena and to address issues related to setting boundary conditions and predicting timescales from unsteady to steady states. This research primarily aims to investigate stress variations in granular materials as a function of shear rate, encompassing both steady and unsteady states. Additionally, the thesis examines the phenomena of wall force anomalies and vortex flows. In Couette cell experiments and vertical plane shear simulations, granular material demonstrates a downward flow near the vertical shearing wall and an upward flow adjacent to another static vertical wall. Interestingly, this vortex flow causes a change in the direction of vertical shear stress when wall shearing commences, contradicting the prevalent assumption that particles consistently apply a downward force on the vertical wall.
The study concludes with key findings, including the observation that normal and shear stresses on the shearing wall increase slowly after the initiation of shearing, and that steady-state values for these stresses are independent of the shearing speed within a certain range. The study also found that the height of particles near the shearing wall decreases gradually with the presence of vortex flow, and that the shear rate near the moving wall is initially high and decreases slowly to reach a steady state. Additionally, we used a non-local constitutive model and Boussinesq approximation to predict the downward flow that is driven by gravity and variations in the solid fraction near the shearing surface, as well as the decay profile of velocity in an infinitely wide box for the steady state.
Overall, this thesis contributes to our understanding of granular materials in the slow flow regime, providing insights into their behavior under shear. The non-local model accurately predicts the downward flow and velocity decay profile, indicating its potential as a valuable tool for future research.</p
Structural Insights into the Conformational Plasticity and Antibody Recognition of HIV-1 Env
Get PDFAcquired immunodeficiency syndrome (AIDS) and its causal agent, the human immunodeficiency virus 1 (HIV-1), remain a global public health concern since they were first identified in the early 1980s. Diligent research and gradual scientific advances have led to innovative strategies in HIV-1/AIDS prevention and treatment, transforming an obscure and deadly disease into a manageable condition with a normal life expectancy. Despite this progress, researchers have yet to develop a safe and effective vaccine against HIV-1. The work presented here describes a structural perspective related to the HIV-1 Envelope (Env) glycoprotein, the sole viral target of vaccines that seek to elicit neutralizing antibodies.
Env is the only viral protein on the surface of HIV-1 virions and is composed of a homotrimer of gp120/gp41 heterodimers. Env mediates entry into target cells by engaging the host receptor, CD4. CD4 triggers conformational changes in gp120, thereby enabling coreceptor recognition. Interactions with the host coreceptor trigger structural rearrangements in gp41 that facilitate fusion of host and viral membranes leading to infection. Our work builds upon our understanding of Env structural plasticity. First, we evaluated the conformational plasticity of soluble Env constructs using double electron-electron resonance (DEER) spectroscopy. This method measured distances between probes in Env subunits, allowing us to interpret the distribution in distances as Env flexibility. Our findings captured previously unseen nuances in static Env structures including gp41 elasticity and conformational heterogeneity associated with CD4-receptor binding. Although our work gave a new perspective on Env flexibility, it largely corroborated observations from static Env structures. Importantly, this suggested that soluble versions of Env, which serve as templates for immunogen design, retain favorable structural properties.
Informed with these insights in Env structure, we then sought to address a prevailing question related to receptor engagement: how many CD4 receptors are needed to induce gp120 conformation changes that lead to coreceptor binding followed by fusion? Prior work only characterized CD4-induced Env structural changes in Envs complexed with three soluble CD4 proteins. In our work, we designed and structurally characterized Envs bound to only one or two CD4 receptors. We found that Env engagement of one CD4 resulted in minor changes to the prefusion, closed Env conformation while Env bound to two CD4 molecules led to CD4-induced opening in the CD4-bound gp120s and a partially open conformation in the unliganded gp120.
Structural biology has also been leveraged to characterize the mechanism by which broadly neutralizing antibodies (bNAbs) recognize HIV-1 Env. We include an extensive review of how structural observations from antibodies bound to viral proteins contribute to our understanding of antibody-mediated viral neutralization. We also present a technical evaluation of bNAb binding assays that revealed how Env conformations can be unintentionally altered resulting in misleading antibody binding results and identified ideal methods to ensure reliable data.
Additionally, we report on projects related to bNAbs that target the CD4 binding site (CD4bs) epitope of Env. In the first, we characterized the inferred germline (iGL) precursor of BG24, a VRC01-class bNAb with features that make it a promising target for vaccine design. We solved four cryo-EM structures of BG24iGL constructs complexed with different Envs and provided insight on the mode of iGL accommodation. The second project centers around the IOMA-class of CD4bs bNAbs. We characterized features of IOMA-class bNAbs and measured how different features contribute to neutralization breadth and potency. Taken together, the conclusions from our work provide guidance for the next generation of structure-based, CD4bs-targeting immunogen design.</p
Optimal Design of Soft Responsive Actuators and Impact Resistant Structures
Get PDFThe rapid pace of development of new responsive and structural materials along with significant advances in synthesis techniques, which may incorporate multiple materials in complex architectures, provides an opportunity to design functional devices and structures of unprecedented performance. These include implantable medical devices, soft-robotic actuators, wearable haptic devices, mechanical protection, and energy storage or conversion devices. However, the full realization of the potential of these emerging techniques requires a robust, reliable, and systematic design approach. This thesis explores this through optimal design methods. By investigating pressing engineering problems which exploit these advances in materials and manufacturing, we develop optimal design methods to realize next-generation structures.
We begin by reviewing classical optimal design methods, the mathematical difficulties they raise, and the practical approaches of overcoming these difficulties. We introduce the canonical problem of compliance minimization of a linear elastic structure. After illustrating the intricacies of this seemingly simple problem, we detail contemporary methods used to address the underlying mathematical issues.
We then turn to extending these classical methods for emerging materials and technologies. We must incorporate optimal design with rich physical models, develop computational approaches for efficient numerics, and study mathematical regularization to obtain well-posed optimization problems. Additionally, care must be taken when selecting an application-tailored objective function which captures the desired behavior. Finally, we must also take into account manufacturing constraints in scenarios where the fabrication pathway affects the structural layout. We address these issues by exploring model optimal design problems. While these serve to ground the fundamental study, they are also relevant, pressing engineering problems.
The first application we consider is the design of responsive structures. Recent developments in material synthesis and 3D printing of anisotropic materials, such as liquid crystal elastomers (LCE), have facilitated the realization of structures with arbitrary morphology and tailored material orientation. These methods may also produce integrated structures of passive and active material. This creates a trade-off between stiffness and actuation flexibility when designing such structures. Thus, we turn to optimal design. This is complicated by anisotropic behavior and finite deformations, manufacturing constraints, and choice of objective function. Like many optimal design problems, the naive formulations are ill-posed giving rise to mesh dependence, lack of convergence, and other numerical deficiencies. So, starting with a simple setting using linear kinematics and working all the way to finite deformation, we develop a systematic mathematical theory that motivates, and then rigorously proves, an alternate well-posed formulation. We examine suitable objective functions, before studying a series of examples in both small and finite deformation. However, the manufacturing process constrains the design as extrusion-based 3D printing aligns nematic directors along the print path. We extended the formulation with these considerations to produce print-aware designs while also recovering the fabrication pathway. We demonstrate the formulation by designing and producing physical realizations of these actuators.
Next, we explore optimal design of impact resistant structures. The complex physics and numerous failure modes of structural impact creates challenges when designing for impact resistance. Here, we apply gradient-based topology optimization to the design of such structures. We start by constructing a variational model of an elastic-plastic material enriched with gradient phase-field damage, and present a novel method to accurately and efficiently compute its transient dynamic time evolution. Sensitivities over this trajectory are computed through the adjoint method, and we develop a numerical method to solve the resulting adjoint dynamical system. We demonstrate this formulation by studying the optimal design of 2D solid-void structures undergoing blast loading. Then, we explore the trade-offs between strength and toughness in the design of a spall-resistant structure composed of two materials of differing properties undergoing dynamic impact.
We conclude by summarizing the presented work and discuss the contribution towards the overarching goal of optimal design for emerging materials technologies. From our study, key issues have arose which must be addressed to further progress the field. We examine these and lay a pathway for future studies which will allow optimal design to tackle complicated, pressing engineering problems.</p
Computational Compensation for Model Imperfections in Photoacoustic Computed Tomography
Get PDFPhotoacoustic computed tomography (PACT) images biological tissues’ optical absorption through detection of photon-absorption-induced ultrasonic waves. Various systems have been proposed for PACT and they are described by different mathematical models to reconstruct from detected ultrasonic signals the photon-absorption-induced initial pressure, the main contrast in PACT. Accurate image reconstruction has high requirements for the system and the mathematical model, which is often imperfect in practice due to multiple factors, e.g., limited transducer bandwidth, finite transducer element size, sparse spatial sampling, partial-view detection, and tissue motion. The focus of this dissertation is on using computational methods to compensate for these model imperfections.
First, for a human breast imaging system based on a full-ring transducer array, we incorporate the limited transducer bandwidth into the model for spatiotemporal analysis to clarify the aliasing due to sparse spatial sampling and propose (1) two methods (radius-dependent spatiotemporal antialiasing and location-dependent spatiotemporal antialiasing) to mitigate these artifacts. Second, for an isotropic-resolution 3D PACT system formed by four arc arrays, we consider both the limited transducer bandwidth and the finite transducer element size and (2) compress the system matrix through singular value decomposition and fast Fourier transform for its efficient explicit expression. Enabled by this expression, we then propose (3) fast sparsely sampling functional imaging by incorporating a densely sampled prior image into the system matrix, which maintains the critical linearity while mitigating artifacts, and (4) intra-image nonrigid motion correction by incorporating the motion as subdomain translations into the system matrix and reconstructing the translations together with the image iteratively. Finally, for a single-shot 3D PACT system based on a single ultrasonic transducer, we propose (5) a fast implementation of the forward model by connecting traditional PACT with virtual detector responses through fast Fourier transform, and we iteratively reconstruct the image from signals with extremely compressed sensing and partial-view detection.
All these proposed methods enable image reconstruction or significantly improve image quality in numerical simulations, phantom experiments, and in vivo experiments. Although they are demonstrated only for certain PACT systems, they are directly applicable to other systems and can be extended to other tomographic imaging modalities such as X-ray computed tomography (X-ray CT) and magnetic resonance imaging (MRI).</p
Topological Invariants of Gapped Quantum Lattice Systems
Get PDFIn the first part of the thesis, a systematic way to construct topological invariants of gapped states of quantum lattices systems is proposed. It provides a generalization of the Berry phase and its equivariant analogue to systems with locality in arbitrary dimensions. For a smooth family of gapped ground states in d dimensions, it gives a closed (d + 2)-form on the parameter space which generalizes the curvature of the Berry connection. Its cohomology class is a topological invariant of the family. When the family is equivariant under the action of a compact Lie group G, topological invariants take values in the equivariant cohomology of the parameter space. These invariants unify and generalize the Hall conductance and the Thouless pump. We prove quantization properties of the invariants for low-dimensional invertible systems.
In the second part, we discuss the properties of the invariant associated with the Hall conductance for 2d lattice systems with U(1)-symmetry. We define anyonic states associated with the flux insertions and relate their statistics to this invariant. We also provide the construction of states realizing chiral topological order with a non-trivial value of this invariant. The construction is based on the data of a unitary regular vertex operator algebra.</p
Freeze Casting - from Battery Separators to Ceramic Scaffolds
Get PDFFreeze casting is a versatile pore-forming technique which allows tunability of pore structures including pore size, size distribution, morphology, and alignment in various material systems. It is that versatility that makes freeze casting a prospective candidate for fabrication of porous components used in a wide range of fields, ranging from biomaterials to supercapacitors. This work explores freeze casting as the processing route to fabricate battery separators and ceramic scaffolds.
The first part of this study assesses the feasibility of tape/freeze casting, a combination of tape casting and freeze casting, in fabricating battery separators for sodium-ion batteries. Poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP) is chosen as the separator material due to its chemical inertness in battery environments, and dioxane is selected as the solvent for PVDF-HFP owing to its dendritic crystal structure and the absence of demixing upon freezing the solution. PVDF-HFP membranes fabricated by a bi-directional tape/freeze casting with dioxane exhibit through-thickness, directionally aligned pore structures. Although PVDF-HFP is shown to surpass reference separators in electrolyte affinity and electrochemical performance, composite strategies are designed to provide enhanced mechanical and electrochemical properties.
Firstly, the effects of alumina, a reinforcing agent introduced via ball milling with dioxane to form suspensions prior to tape/freeze casting, are examined. Composite PVDF- HFP/Al2O3 membranes show similar microstructures to their polymer counterpart, with enhanced resistance to thermal shrinkage, elastic modulus, electrolyte uptake, and ionic conductivity. Moreover, coin cells made with composite membranes deliver better rate performance and cycling stability than those with polymer membranes and filter paper reference materials.
Secondly, an alternative route to incorporate inorganic reinforcing elements into PVDF- HFP membranes is found through a co-solvent process. Silica particles from a sol-gel reaction of tetraethoxysilane (TEOS) are introduced into PVDF-HFP membranes via a co- solvent method in conjunction with dimethyl sulfoxide (DMSO). The tape/freeze-cast PVDF-HFP membranes fabricated with DMSO alone exhibit directionally aligned pores, while a hierarchical pore morphology with circular pores on the aligned pore walls is observed in composite membranes fabricated with TEOS, and hence, silica additions.
Composite PVDF-HFP/SiO₂ membranes outperform their unreinforced polymer counterpart in terms of elastic modulus, thermal stability, electrolyte affinity, and ionic conductivity, along with capacity retention and cycling performance when assembled into coin cells.
The final portion of this study evaluates the capability of freeze casting for highly permeable ceramic scaffolds using a polymethylsiloxane preceramic polymer with tert- butyl alcohol (TBA), a solvent creating prismatic pores without side arms that affords high permeability. A double-sided freeze-casting configuration results in more controlled freezing of the polymer solutions in comparison with the conventional single-sided counterpart, and hence a more aligned pore structure is obtained. Further improvement in pore alignment accompanied by an eight-fold increase in water permeability is realized by templating the substrate in freeze-casting molds.</p
Reliable Controller Synthesis: Guarantees for Safety-Critical System Testing and Verification
Get PDFThe well-known quote by George Box states that "All models are wrong, but some are useful", and the controls and robotics communities alike have followed a similar paradigm to make significant theoretical and practical advances in the study of controllable systems to date. However, recent robotic system requirements include formal considerations for system safety, especially as we engineer systems that are required to work alongside us in our daily lives. As such, current research directions require analyses that consider these inaccurate system models, our inaccurate understanding of the environments in which these systems operate, and their combined effects on safe, effective system operation, e.g. the canonical autonomous driving problem in exceedingly difficult-to-model urban environments. Recently, this has led to burgeoning efforts in a formal study of controller verification. Specifically, verification denotes the process of determining whether a controller steers its system to exhibit desired behaviors despite the variety of environments the system might face during operation, e.g. whether the autonomous car's controller successfully drives the car to a destination without crashing into obstacles or pedestrians along the way. However, formalization of such a verification pipeline has proved difficult to date, especially since both the models we use for controller synthesis and our understanding of system environments are typically inaccurate.
As a result, this thesis describes our efforts in the development of a formal verification pipeline that addresses a few key challenges in traditional approaches to safety-critical system verification. The first contribution centers on difficult, reactive test synthesis. By test synthesis, we mean the construction of a (potentially difficult) environment in which we require the system under test to perform its objective, e.g. placement of parked cars around which an autonomous vehicle must park. Typically phrased as an optimization problem over the space of allowable environments, these tests are "static" insofar as they do not react to the system's choices made during the test. We posit that such reactivity could more accurately identify worst-case system behavior. As a result, we phrase reactive, maximally difficult test synthesis as a game-theoretic optimization problem, leveraging the same control theoretic tools that facilitate safety-critical controller synthesis - control barrier functions and signal temporal logic. We prove that our proposed synthesis technique is always solvable and always produces a realizable test environment. Finally, we showcase our results by synthesizing reactive tests for both single and multi-agent systems.
The second set of contributions centers on our efforts in uncertainty quantification. Due to un-modeled system and environmental aspects affecting system evolution in unpredictable ways, real-life systems need not realize the same paths every time. As such, typical analyses phrase verification as an optimization problem minimizing the expected value of a function over system trajectories with the expectation taken over this path variability, the distribution for which is assumed to be known. However, we posit that such an analysis should be risk-aware, i.e. account for this variability in a more principled fashion than an expectation-specific analysis, and should not assume apriori knowledge of the distribution corresponding to path variability, as it will be unknown in practice. To that end, we develop methods to bound a subset of risk measures for random variables whose distributions are unknown. This subset includes both Value-at-Risk and other, coherent risk measures heavily utilized in the controls and robotics communities. Simultaneously, we note that the same procedure can be applied to a wide class of non-convex optimization problems. In doing so, we develop a percentile-based optimization approach that rapidly identifies percentile solutions to optimization problems, i.e. a 90-th percentile solution is as good as 90% of solutions in the considered decision space.
The third set of contributions focuses on the application of the prior mathematical developments to facilitate both risk-aware safety-critical system verification and controller synthesis. We phrase risk-aware controller verification as a risk-measure identification problem and utilize the prior bounding results to provide an efficient, dimensionally-independent verification procedure. Then, we phrase risk-aware controller synthesis as an optimization problem maximizing the bound provided by our risk-aware verification method, and show that this problem is solvable by the percentile optimization methods mentioned prior. Finally, we lay the foundation for the utilization of the aforementioned mathematical developments in other aspects of controls and robotics and communities more broadly. We show how risk-measure bounding can augment models both offline and online to robustify safety-critical controllers, how percentile optimization can facilitate "optimal" input selection and guarantee generation for non-convex finite-time optimal controllers, and how multiple applications of the percentile approach can also bound the optimality gap of reported percentile solutions. We showcase all these results on hardware for multiple systems and highlight the data efficiency of our proposed approaches.</p