Combinatorial Set Research Articles

Combinatorial Civic Crowdfunding with Budgeted Agents: Welfare Optimality at Equilibrium and Optimal Deviation

Civic Crowdfunding (CC) uses the ``power of the crowd" to garner contributions towards public projects. As these projects are non-excludable, agents may prefer to ``free-ride," resulting in the project not being funded. Researchers introduce refunds for single project CC to incentivize agents to contribute, guaranteeing the project's funding. These funding guarantees are applicable only when agents have an unlimited budget. This paper focuses on a combinatorial setting, where multiple projects are available for CC and agents have a limited budget. We study specific conditions where funding can be guaranteed. Naturally, funding the optimal social welfare subset of projects is desirable when every available project cannot be funded due to budget restrictions. We prove the impossibility of achieving optimal welfare at equilibrium for any monotone refund scheme. Further, given the contributions of other agents, we prove that it is NP-Hard for an agent to determine its optimal strategy. That is, while profitable deviations may exist for agents instead of funding the optimal welfare subset, it is computationally hard for an agent to find its optimal deviation. Consequently, we study different heuristics agents can use to contribute to the projects in practice. We demonstrate the heuristics' performance as the average-case trade-off between the welfare obtained and an agent's utility through simulations.

Systematic Part Transfer by Extending a Modular Toolkit to Diverse Bacteria.

It is impractical to develop a new parts collection for every potential host organism. It is well-established that gene expression parts, like genes, are qualitatively transferable, but there is little quantitative information defining transferability. Here, we systematically quantified the behavior of a parts set across multiple hosts. To do this, we developed a broad host range (BHR) plasmid system compatible with the large, modular CIDAR parts collection for E.coli, which we named openCIDAR. This enabled testing of a library of DNA constructs across the Pseudomonadota─Escherichia coli, Pseudomonas putida, Cupriavidus necator, and Komagataeibacter nataicola. Part performance was evaluated with a standardized characterization procedure that quantified expression in terms of molecules of equivalent fluorescein (MEFL), an objective unit of measure. The results showed that the CIDAR parts enable graded gene expression across all organisms─meaning that the same parts can be used to program E.coli, P.putida, C.necator, and K.nataicola. Most parts had a similar expression trend across hosts, although each organism had a different average gene expression level. The variability is enough that to achieve the same MEFL in a different organism, a lookup table is required to translate a design from one host to another. To identify truly divergent parts, we applied linear regression to a combinatorial set of promoters and ribosome binding sites, finding that the promoter J23100 behaves very differently in K.nataicola than in the other hosts. Thus, it is now possible to evaluate any CIDAR compatible part in three other hosts of interest, and the diversity of these hosts implies that the collection will also be compatible with many other Proteobacteria (Pseudomonadota). Furthermore, this work defines an approach to generalize modular synthetic biology parts sets beyond a single host, implying that only a few parts sets may be needed to span the tree of life. This will accelerate current efforts to engineer diverse species for environmental, biotechnological, and health applications.

Combinatorial Generation Algorithms for Some Lattice Paths Using the Method Based on AND/OR Trees

Methods of combinatorial generation make it possible to develop algorithms for generating objects from a set of discrete structures with given parameters and properties. In this article, we demonstrate the possibilities of using the method based on AND/OR trees to obtain combinatorial generation algorithms for combinatorial sets of several well-known lattice paths (North-East lattice paths, Dyck paths, Delannoy paths, Schroder paths, and Motzkin paths). For each considered combinatorial set of lattice paths, we construct the corresponding AND/OR tree structure where the number of its variants is equal to the number of objects in the combinatorial set. Applying the constructed AND/OR tree structures, we have developed new algorithms for ranking and unranking their variants. The performed computational experiments have confirmed the obtained theoretical estimation of asymptotic computational complexity for the developed ranking and unranking algorithms.

ML Approach to Improve the Costs and Reliability of a Wireless Sensor Network.

Temperature-controlled closed-loop systems are vital to the transportation of produce. By maintaining specific transportation temperatures and adjusting to environmental factors, these systems delay decomposition. Wireless sensor networks (WSN) can be used to monitor the temperature levels at different locations within these transportation containers and provide feedback to these systems. However, there are a range of unique challenges in WSN implementations, such as the cost of the hardware, implementation difficulties, and the general ruggedness of the environment. This paper presents the novel results of a real-life application, where a sensor network was implemented to monitor the environmental temperatures at different locations inside commercial temperature-controlled shipping containers. The possibility of predicting one or more locations inside the container in the absence or breakdown of a logger placed in that location is explored using combinatorial input-output settings. A total of 1016 machine learning (ML) models are exhaustively trained, tested, and validated in search of the best model and the best combinations to produce a higher prediction result. The statistical correlations between different loggers and logger combinations are studied to identify a systematic approach to finding the optimal setting and placement of loggers under a cost constraint. Our findings suggest that even under different and incrementally higher cost constraints, one can use empirical approaches such as neural networks to predict temperature variations in a location with an absent or failed logger, within a margin of error comparable to the manufacturer-specified sensor accuracy. In fact, the median test accuracy is 1.02 degrees Fahrenheit when using only a single sensor to predict the remaining locations under the assumptions of critical system failure, and drops to as little as 0.8 and 0.65 degrees Fahrenheit when using one or three more sensors in the prediction algorithm. We also demonstrate that, by using correlation coefficients and time series similarity measurements, one can identify the optimal input-output pairs for the prediction algorithm reliably under most instances. For example, discrete time warping can be used to select the best location to place the sensors with a 92% match between the lowest prediction error and the highest similarity sensor with the rest of the group. The findings of this research can be used for power management in sensor batteries, especially for long transportation routes, by alternating standby modes where the temperature data for the OFF sensors are predicted by the ON sensors.

Impartial Hypergraph Games

We study two building games and two removing games played on a finite hypergraph. In each game two players take turns selecting vertices of the hypergraph until the set of jointly selected vertices satisfies a condition related to the edges of the hypergraph. The winner is the last player able to move. The building achievement game ends as soon as the set of selected vertices contains an edge. In the building avoidance game the players are not allowed to select a set that contains an edge. The removing achievement game ends as soon as the complement of the set of selected vertices no longer contains an edge. In the removing avoidance game the players are not allowed to select a set whose complement does not contain an edge. We develop some generic tools for finding the nim-value of these games and show that the nim-value can be an arbitrary nonnegative integer. The outcome of many of these games were previously determined for several special cases in algebraic and combinatorial settings. We provide several examples and show how our tools can be used to refine these results by finding nim-values.

Abstract 2716: Further development of microRNA-34a therapeutics to target castration-resistant prostate cancer

Abstract Prostate cancer (PCa) is heterogeneous and typically presents with multiple distinct cancer foci. Androgen receptor signaling inhibitors (ARSIs) are standard-of-care therapies for advanced PCa and can prolong patient survival by months. Treatment-naïve PCa harbors a population of stem-like cells (AR-/lo) functionally defined as PCa stem cells (PCSCs). During cancer progression or upon therapeutic treatment, non-CSCs may undergo cellular plasticity and be reprogrammed to PCSCs. As a result, PCSCs may become the predominant cell population in advanced (Gleason 9/10) and ARSI-resistant tumors, and the “root cause” for drug resistance. However, current ARSI-based treatment regimens primarily target AR+ CRPC cells or clones but largely ignore AR-/lo cells. microRNA-34a (miR-34a), a tumor-suppressive miRNA dysregulated in PCa, is a potent PCSC inhibitor by targeting cellular processes essential for CSC survival and functions. Here we aim to develop novel miR-34a therapeutics to target PCa cell heterogeneity and CRPC. miR-34a expression negatively correlates with the tumor stage in PCa samples. And miR-34a expression is significantly reduced in p53 mutated tumors. We explored two novel miR-34a delivery platforms. The first is packaging-free ligand-mediated conjugate utilizing 2- [3-(1,3-dicarboxy propyl) ureido] pentanedioic acid (DUPA) as a high-affinity binding ligand for prostate-specific membrane antigen (PSMA), which is upregulated in PCa. The second platform uses cationic lipid nanoparticles to package miR-34a (LNP-miR-34a). We show that during PCa progression and upon treatment, PSMA expression is upregulated and continues to be highly expressed in CRPC and PSMA expression level negatively correlates with overall survival in PCa. Moreover, DUPA-conjugate can specifically target PSMA-expressing PCa both in vitro and in vivo. DUPA-miR-34a significantly induces target gene silencing in vitro. On the other hand, LNP show good encapsulation efficiency for miR-34a, LNP-miR-34a significantly repressed cell proliferation by downregulating critical stemness factors in a dose-dependent manner. Meanwhile, we are pursuing chemical modifications of the miR-34a mimics aiming to improve serum stability and future in vivo applications. Both fully modified miR-34a (FM-miR-34a) and partially-modified-miR-34a (PM-miR-34a) exhibit improved serum stability without compromising the functionality. Our data suggest that DUPA-miR-34a represents a promising novel therapeutic for PCSC-enriched CRPC that continue to express PSMA, whereas LNP-miR-34a could be a systemic therapeutic for CRPC treatment in the combinatorial setting. miR-34a, as a direct target and a major mediator of p53 functions, may represent an effective replacement therapeutic in p53-mutated advanced PCa. Citation Format: Wen (Jess) Li, Xiaozhuo Liu, Ahmed M. Abdelaal, Sudarsan Kasireddy, Yafei Su, Yun Wu, Andrea L. Kasinski, Dean G. Tang. Further development of microRNA-34a therapeutics to target castration-resistant prostate cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2716.

Epistasis decreases with increasing antibiotic pressure but not temperature.

Predicting mutational effects is essential for the control of antibiotic resistance (ABR). Predictions are difficult when there are strong genotype-by-environment (G × E), gene-by-gene (G × G or epistatic) or gene-by-gene-by-environment (G × G × E) interactions. We quantified G × G × E effects in Escherichia coli across environmental gradients. We created intergenic fitness landscapes using gene knock-outs and single-nucleotide ABR mutations previously identified to vary in the extent of G × E effects in our environments of interest. Then, we measured competitive fitness across a complete combinatorial set of temperature and antibiotic dosage gradients. In this way, we assessed the predictability of 15 fitness landscapes across 12 different but related environments. We found G × G interactions and rugged fitness landscapes in the absence of antibiotic, but as antibiotic concentration increased, the fitness effects of ABR genotypes quickly overshadowed those of gene knock-outs, and the landscapes became smoother. Our work reiterates that some single mutants, like those conferring resistance or susceptibility to antibiotics, have consistent effects across genetic backgrounds in stressful environments. Thus, although epistasis may reduce the predictability of evolution in benign environments, evolution may be more predictable in adverse environments. This article is part of the theme issue 'Interdisciplinary approaches to predicting evolutionary biology'.

Evaluating the Sensitivity of Polarimetric Features Related to Rotation Domain and Mapping Chinese Fir AGB Using Quad-Polarimetric SAR Images

Unaffected by cloud cover and solar illumination, synthetic aperture radar (SAR) images coupled with quad-polarimetric techniques have significant potential for mapping forest aboveground biomass (AGB) in the mountains of southern China. To improve the accuracy of mapping forest AGB, it is necessary to accurately interpret and evaluate the sensitivity of polarimetric features related to polarimetric response in complex forests. In this study, several rotated polarimetric features were extracted from L-band quad-polarimetric ALOS PALSAR-2 images based on uniform polarimetric matrix rotation theory. In addition, the sensitivity of rotated polarimetric features with forest parameters was evaluated by the Pearson correlation coefficient, sensitivity index (SI), and saturation levels. Ultimately, the forest AGB was mapped with various combinatorial feature sets by a proposed feature selection method based on the sensitivity index. The results illustrated that rotated polarimetric features extracted from the rotational domain have higher sensitivity with various forest parameters and higher saturation levels for mapping forests than other traditional features. After using the proposed feature selection method and combinatorial feature sets, the rRMSE of mapped forest AGB ranged from 22.5% to 33.9% for two acquired images, and the best result was obtained from the combination of three types of polarimetric features (BC + C4 + Ro). It is also confirmed that different types of features extracted from quad-polarimetric SAR images have better compensation effects and the accuracy of mapped forest AGB is significantly improved.

A grammatical calculus for peaks and runs of permutations

We develop a nonstandard approach to exploring polynomials associated with peaks and runs of permutations. With the aid of a context-free grammar, or a set of substitution rules, one can perform a symbolic calculus, and the computation often becomes rather simple. From a grammar, it follows at once a system of ordinary differential equations for the generating functions. Utilizing a certain constant property, it is even possible to deduce a single equation for each generating function. To bring the grammar to a combinatorial setting, we find a labeling scheme for up-down runs of a permutation, which can be regarded as a refined property, or the differentiability in a certain sense, in contrast with the usual counting argument for the recurrence relation. The labeling scheme also exhibits how the substitution rules arise in the construction of the combinatorial structures. Consequently, polynomials on peaks and runs can be dealt with in two ways, combinatorially or grammatically. The grammar also serves as a guideline to build a bijection between permutations and increasing trees that maps the number of up-down runs to the number of nonroot vertices of even degree. This correspondence can be adapted to left peaks and exterior peaks, and the key step of the construction is called the reflection principle.

Цільова функція в комбінаторній оптимізації та її властивості

Introduction. Some properties of combinatorial optimization problems are described, which affect the regularity of changes in the values of the objective function regardless of the input data. It is shown that this regularity depends on the arrangement of combinatorial configurations, a certain structure of input information, the transposition of permutation elements and the symmetry of combinatorial sets (argument). In problems that are solved on permutations, and a subset of isomorphic combinatorial configurations, the class of the objective function is determined depending on their ordering and the structure of the input data. Formulation of the problem. There are works where, in combinatorial optimization, the change in the values of the objective function is investigated depending on the special structure of the input information. These studies are related to the selection of subclasses of solvable problems, and various structures are generalized for which the objective function changes in the same way. The task is to identify the parameters of problems of this class, in which the input data do not affect the pattern of changes in the values of the objective function. The proposed approach. To identify the parameters under which the input data do not affect the pattern of changes in the objective function values in combinatorial optimization, this pattern is analyzed from the arrangement of permutations, the transposition of its elements, and the structure of the input information. The change of the values of the objective function from the structure of its argument (the structure of combinatorial sets) is also investigated. Conclusion. Knowing the properties of combinatorial functions allows you to establish a change in the values of the target function depending on the transposition of the elements in the permutation, on the arrangement of combinatorial configurations. If the combinatorial set consists of one subset of isomorphic combinatorial configurations (permutations), then the set of values of the objective function follows the rules characteristic of permutations. If the set consists of subsets of isomorphic combinatorial configurations, then they can be ordered in such a way that the objective function for this ordering varies as piecewise monotonic functions regardless of the input data, and for the isomorphic subset it varies as in problems whose objective function argument is the permutation.

Weighted combinatorial Ricci flow and metrics defined by degenerate circle packings

Chow and Luo [1] in 2003 had shown that the combinatorial analogue of the Hamilton Ricci flow on surfaces under certain conditions converges to Thruston?s circle packing metric of constant curvature. The combinatorial setting includes weights defined for edges of a triangulation. Crucial assumption in the paper [1] was that the weights are nonnegative. Recently we have shown that same statement on convergence can be proved under weaker condition: some weights can be negative and should satisfy certain inequalities, [3]. Moreover, in [6] notions of degenerate circle packing and corresponding metric were introduced. In [6] theory of combinatorial Ricci flow for such metrics was developed, which includes Chow-Luo theory as a partial case for nondegenerate circle packing and nonnegative weights on edges. On the other hand, in [2] the combinatorial Yamabe flow was introduced and investigated. In [7, 8] we developed weighted modification of Yamabe flow. In this paper we merge ideas from these two theories and introduce weighted combinatorial Ricci flow on metrics defined by degenerate circle packings. We prove that under certain conditions for any initial metric the flow converges to a unique metric of constant curvature.

Virtual Receiver Matrix and Combinatory Analog Operations for Future Multifunction Reconfigurable Sensing and Communication Wireless Systems

This article proposes and presents a topological receiver architecture, named virtual receiver matrix (VRM), suitable for future smart multifunction wireless systems. This concept is devised and benefits from using combinatory analog operations with multiple distributed units in a receiver matrix or array. This mechanism of receiving data through spatially “floating” distributed virtual receiver channels offers an unprecedented solution of providing unparalleled degrees of freedom to implement multiple functions such as data reception, angle-of-arrival (AoA) detection, radar, and imaging operations among many others in a single receiver architecture. Interestingly, the total number of possible virtual receivers from different combinations of unit cells in a matrix is also significantly increased compared with a conventional “fixed” receiver array. Each virtual receiver, made of a combinatory set, depends on the characteristics of incoming signals and their illumination angle. A mathematical model is established and investigated for the design of unit cells. Although a prototype of choice is studied and designed for the fifth-generation (5G) wireless systems, it is anticipated that the VRM concept is applicable to the sixth-generation (6G) and future wireless systems for enhancing their functionality, capacity, agility, and speed. In this article, a multiport interferometric technique is used for each unit cell of VRM as a proof-of-concept; however, any other receiver type that uses the phase difference of incoming signals can be deployed as the unit cell for the realization of the VRM concept. First, the experimental results of the fabricated proof-of-concept prototype for various modulation schemes including QPSK, 16-quadrature amplitude modulation (QAM), and 32-QAM are shown with data rates up to 1 MS/s, having a maximum error vector magnitude (EVM) of 9%. Finally, a systematic scheme for 2-D AoA detection using a special combination of unit cells is proposed and demonstrated through the proposed VRM.

Group Testing Matrix Design for PCR Screening with Real-Valued Measurements.

Single-step nonadaptive group testing approaches for reducing the number of tests required to detect a small subset of positive samples from a larger set require solving two algorithmic problems. First, how to design the samples-to-tests measurement matrix, and second, how to decode the results of the tests to uncover positive samples. In this study, we focus on the first challenge. We introduce real-valued group testing, which matches the characteristics of existing PCR testing pipelines more closely than combinatorial group testing or compressed sensing settings. We show a set of conditions that allow measurement matrices to guarantee unambiguous decoding of positives in this new setting. For small matrix sizes, we also propose an algorithm for constructing matrices that meet the proposed condition. On simulated data sets, we show that the matrices resulting from the algorithm can successfully recover positive samples at higher positivity rates than matrices designed for combinatorial group testing setting. We use wet laboratory experiments involving SARS-CoV-2 nasopharyngeal swab samples to further validate the approach.

A Hyperparameter Adaptive Genetic Algorithm Based on DQN

The hyperparameters of the metaheuristic algorithm are difficult to determine when solving optimization problems. The existing methods mainly adjust hyperparameters through preset rules or traditional RL. The performance of the above methods is unsatisfactory and the generalization is poor. This work proposes a deep Q-learning network (DQN)-based dynamic setting framework for combinatorial hyperparameters, and applies it to a Genetic algorithm (GA) to improve its performance. By defining the four elements of the environment, state, action and reward required for learning strategy in advance, the parametrized strategy can be trained offline and different DQN models can be studied. Our method was compared with other algorithms and achieved the shortest path on 14 of 15 public TSP instances. Meanwhile, the test results on our simulation TSP validation dataset revealed that Category DQN achieved the best performance. This means the proposed method can effectively solve the problem of combinatorial hyperparameters setting, and bring more solving advantages to the GA.

A study on some combinatorial sets in partial semigroups

In this paper, we investigate the image and preimage of the important combinatorial sets such as central sets, [Formula: see text]-sets, and [Formula: see text]-sets which play an important role in the study of combinatorics under certain partial semigroup homomorphism. Using that we prove certain results that deal with the existence of [Formula: see text]-set which are not central in partial semigroup framework.

The functionality of aesthetic illusion and epistemic relativism in Grace Ogot’s The Strange Bride (1989)

Throughout a dimensional analysis of rationality and a factual relativism justification, The Strange Bride upholds a temporal perspective and a neo-experience of present within which Grace Ogot methodic metafiction, defines a differential relational model regarding the Be-, the There-ness and norms of belief. This vigorous realm of interaction impels the inflection of alethic modality to stand as a meromorphic function; then, the illusion of time and the spacetime dimension dive the cosmological time, the metaphysic of time, the linguistic conception of time and contemporary cultural theory in a combinatorial set of a pluricellular function and a multifactorial understanding. Through the temporal structure and the representation of the continuum of spacetime, Ogot circumscribes a connectionism transformation; therefore, the form and the content of the Be-ing, the relationalism between space-time continuum and the there-ness dimension of creative imagination become an aesthetic and functional performance. In this dynamic, the problematic of the Be-ing object from the quantitative nominal properties of time characterizes, in this model of narrative, the metaphysic of presence and the cultural theory of time.

Continuous Extensions On Euclidean Combinatorial Configurations

In this paper, we introduce a concept of the Euclidean combinatorial configuration as a mapping of a set of certain objects into a point of Euclidean space. We classify Euclidean combinatorial configurations sets based on their structure and constraints. The proposed typology forms the basis for studying continuous functional representations of combinatorial configurations. Special classes of functional extensions are introduced, their properties are described, and corresponding examples are given.

On Some Optimization Problems on Permutations

Numerous studies consider combinatorial optimization problems and their solution methods, since a large number of practical problems are described by means of combinatorial optimization models. Among these problems, the most prominent ones are function optimization problems on combinatorial configurations. Many of the studies mentioned above propose approaches and describe methods to solve combinatorial optimization problems for linear and fractionally linear functions on combinatorial sets such as permutations and arrangements. This work describes new approaches and methods to solve some maximization problems for linear, fractionally linear and quadratic functions on permutation set. Algorithms for solving these problems are given. For linear function, we provide a considerably easy method to find the permutation on which the function attains its maximum value. We describe the general algorithm to find fractionally linear function maximum. We consider the cases in which variables of numerator and denominator do not intersect, and when the numerator function and the denominator function have one common variable. We also describe the case in which the numerator function and the denominator function in total have incomplete variable list, i.e. when the total quantity of variables is less than the quantity of numbers in the permutation set. As for solving the problem of finding quadratic function maximum on permutation set, it turned out that there is no universal algorithm to solve this problem for any quadratic function. In this paper, we describe a method of finding maximum on permutation set for quadratic function consisting of two items. We provide examples of solving the considered optimization problems. Keywords: function, permutation, set, transposition, coefficients.

On the Connection of Matroids and Greedy Algorithms

Matroids are the combinatorial structure and Greedy algorithmic methods always produces optimal solutions for these mathematical models. A greedy method always selects the option that looks best at each step of process of finding optimal solution. In other words, it selects a choice which is optimal choice locally in such a strategy that this locally chosen option may direct to a solution that will be globally optimal. It is true that while selecting locally optimal solution at each stage, Greedy algorithms may not always yield optimal solutions [1-2], but if we can transform an unknown problem into matroid structure, then there must be a greedy algorithm that will always lead optimal solution for that unknown problem. The range of solutions provided by Greedy is large as compared to the applicability of the Matroid structure. In other words the problems that can be translated into Matroid structure is proper subset of set of all problems whether Greedy algorithm produces optimal solution. Matroid structure thus ensures the global optimal solution one can obtain with help of Greedy approach. We study various logarithmic and linear hierarchical based mathematical models from divergence sources to maximize our information for research purposes. We analyze the time complexity and provide constrains over the upper/lower bounds in correspondence with the optimal (maximum/minimum) solution. We try to establish the relationship between the maximization of information divergences, the optimal-likelihood theory, and classified sharing is instituted. We propose integration of unknown rough sets to matroids in this paper. Particularly, we devise methodically the upper and lower tightening bounds on rough matroids which may expand up to the generic combinatorial matroid structure. The relationships are established by the upper and lower tightening bounds approximations of generalized combinatorial rough sets based on different interdependent relation sets, respectively [4]. As we define the generalized lower/upper bounds for rough matroid, we define a new structure for lower/upper greedoid leading to generalization of the greedoid [5]. Additionally, based on the new established relation, the generalized rough set also provides a theory of poset matroid.

A Fast Algorithm for PAC Combinatorial Pure Exploration

We consider the problem of Combinatorial Pure Exploration (CPE), which deals with finding a combinatorial set of arms with a high reward, when the rewards of individual arms are unknown in advance and must be estimated using arm pulls. Previous algorithms for this problem, while obtaining sample complexity reductions in many cases, are highly computationally intensive, thus making them impractical even for mildly large problems. In this work, we propose a new CPE algorithm in the PAC setting, which is computationally light weight, and so can easily be applied to problems with tens of thousands of arms. This is achieved since the proposed algorithm requires a very small number of combinatorial oracle calls. The algorithm is based on successive acceptance of arms, along with elimination which is based on the combinatorial structure of the problem. We provide sample complexity guarantees for our algorithm, and demonstrate in experiments its usefulness on large problems, whereas previous algorithms are impractical to run on problems of even a few dozen arms. The code is provided at https://github.com/noabdavid/csale. The full version of this paper is available at https://arxiv.org/abs/2112.04197.