Search for: All records

The online list update problem is defined as follows: we are given a list of items and the cost to access any particular item is its position from the start of the list. A sequence of item accesses come online, and our goal is to dynamically reorder the list so that the aggregate access cost is small. We study the stochastic version of the problem where the items are accessed i.i.d. from an unknown distribution p. The study of the stochastic version goes back at least 60 years to McCabe. In this paper, we first consider the simple online algorithm which swaps an accessed item with the item right before it, unless it is at the very front. This algorithm is known as the Transposition rule. Wetheoretically analyze the stationary behavior of Transposition and prove that its performance is within 1 + o(1) factor of the optimal offline algorithm for access sequences sampled from heavy-tailed distributions, proving a conjecture of Rivest from 1976. While the stationary behavior of the Transposition rule is theoretically optimal in the aforemen tioned i.i.d setting, it can catastrophically fail under adversarial access sequences where only the last and second to last items are repeatedly accessed. A desirable outcome would be a policy that performs well under both circumstances. To achieve this, we use reinforcement learning to design an adaptive policy that performs well for both the i.i.d. setting and the above-mentioned adversarial access. Unsurprisingly, the learned policy appears to be an interpolation between Move-to-Front and Transposition with its behavior closer to Move-to-Front for adversarial access sequences and closer to Transposition for sequences sampled from heavy tailed distributions suggesting that the policy is adaptive and capable of responding to patterns in the access sequence. 

We study local filters for the Lipschitz property of real-valued functions f : V → [0,r], where the Lipschitz property is defined with respect to an arbitrary undirected graph G = (V, E ). We give nearly optimal local Lipschitz filters both with respect to ℓ1-distance and ℓ0-distance. Previous work only considered unbounded- range functions over [n]d. Jha and Raskhodnikova (SICOMP ‘13) gave an algorithm for such functions with lookup complexity exponential in d, which Awasthi et al. (ACM Trans. Comput. Theory) showed was necessary in this setting. We demonstrate that important applications of local Lipschitz filters can be accomplished with filters for functions whose range is bounded in [0,r]. For functions f : [n]d → [0,r], we achieve running time (dr log n )O (log r ) for the ℓ1-respecting filter and dO(r) polylog n for the ℓ0-respecting filter, thus circumventing the lower bound. Our local filters provide a novel Lipschitz extension that can be implemented locally. Furthermore, we show that our algorithms are nearly optimal in terms of the dependence on r for the domain {0,1}d, an important special case of the domain [n]d. In addition, our lower bound resolves an open question of Awasthi et al., removing one of the conditions necessary for their lower bound for general range. We prove our lower bound via a reduction from distribution-free Lipschitz testing and a new technique for proving hardness for adaptive algorithms. Finally, we provide two applications of our local filters to real-valued functions, with no restrictions on the range. In the first application, we use them in conjunction with the Laplace mechanism for differential privacy and noisy binary search to provide mechanisms for privately releasing outputs of black-box functions, even in the presence of malicious clients. In particular, our differentially private mechanism for arbitrary real-valued functions runs in time 2polylog min(r,nd ) and, for honest clients, has accuracy comparable to the Laplace mechanism for Lipschitz functions, up to a factor of O (log min(r,nd )). In the second application, we use our local filters to obtain the first nontrivial tolerant tester for the Lipschitz property. Our tester works for functions of the form f : {0,1}d → ℝ, makes queries, and has tolerance ratio 2.01. Our applications demonstrate that local filters for bounded-range functions can be applied to construct efficient algorithms for arbitrary real-valued functions. 

The study of space-bounded computation has drawn extensively from ideas and results in the field of communication complexity. Catalytic Computation (Buhrman, Cleve, Koucký, Loff and Speelman, STOC 2013) studies the power of bounded space augmented with a pre-filled hard drive that can be used non-destructively during the computation. Presently, many structural questions in this model remain open. Towards a better understanding of catalytic space, we define a model of catalytic communication complexity and prove new upper and lower bounds. In our model, Alice and Bob share a blackboard with a tiny number of free bits, and a larger section with an arbitrary initial configuration. They must jointly compute a function of their inputs, communicating only via the blackboard, and must always reset the blackboard to its initial configuration. We prove several upper and lower bounds: 1) We characterize the simplest nontrivial model, that of one bit of free space and three rounds, in terms of 𝔽₂ rank. In particular, we give natural problems that are solvable with a minimal-sized blackboard that require near-maximal (randomized) communication complexity, and vice versa. 2) We show that allowing constantly many free bits, as opposed to one, allows an exponential improvement on the size of the blackboard for natural problems. To do so, we connect the problem to existence questions in extremal graph theory. 3) We give tight connections between our model and standard notions of non-uniform catalytic computation. Using this connection, we show that with an arbitrary constant number of rounds and bits of free space, one can compute all functions in TC⁰. We view this model as a step toward understanding the value of filled space in computation. 

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 [GKKM'20], allowing abstention on adversarially generated parts of the test distribution, and TDS learning [KSV'23], 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 [RV'23], and learning with nasty noise. 

We study truthful mechanisms for approximating the Maximin-Share (MMS) allocation of agents with additive valuations for indivisible goods. Algorithmically, constant factor approximations exist for the problem for any number of agents. When adding incentives to the mix, a jarring result by Amanatidis, Birmpas, Christodoulou, and Markakis [EC 2017] shows that the best possible approximation for two agents and m items is ⌊m2⌋. We adopt a learning-augmented framework to investigate what is possible when some prediction on the input is given. For two agents, we give a truthful mechanism that takes agents' ordering over items as prediction. When the prediction is accurate, we give a 2-approximation to the MMS (consistency), and when the prediction is off, we still get an ⌈m2⌉-approximation to the MMS (robustness). We further show that the mechanism's performance degrades gracefully in the number of mistakes" in the prediction; i.e., we interpolate (up to constant factors) between the two extremes: when there are no mistakes, and when there is a maximum number of mistakes. We also show an impossibility result on the obtainable consistency for mechanisms with finite robustness. For the general case of n≥2 agents, we give a 2-approximation mechanism for accurate predictions, with relaxed fallback guarantees. Finally, we give experimental results which illustrate when different components of our framework, made to insure consistency and robustness, come into play. 

A fundamental notion of distance between train and test distributions from the field of domain adaptation is discrepancy distance. While in general hard to compute, here we provide the first set of provably efficient algorithms for testing localized discrepancy distance, where discrepancy is computed with respect to a fixed output classifier. These results imply a broad set of new, efficient learning algorithms in the recently introduced model of Testable Learning with Distribution Shift (TDS learning) due to Klivans et al. (2023).Our approach generalizes and improves all prior work on TDS learning: (1) we obtain universal learners that succeed simultaneously for large classes of test distributions, (2) achieve near-optimal error rates, and (3) give exponential improvements for constant depth circuits. Our methods further extend to semi-parametric settings and imply the first positive results for low-dimensional convex sets. Additionally, we separate learning and testing phases and obtain algorithms that run in fully polynomial time at test time. 

We consider the problem of hypothesis testing for discrete distributions. In the standard model, where we have sample access to an underlying distribution p, extensive research has established optimal bounds for uniformity testing, identity testing (goodness of fit), and closeness testing (equivalence or two-sample testing). We explore these problems in a setting where a predicted data distribution, possibly derived from historical data or predictive machine learning models, is available. We demonstrate that such a predictor can indeed reduce the number of samples required for all three property testing tasks. The reduction in sample complexity depends directly on the predictor’s quality, measured by its total variation distance from p. A key advantage of our algorithms is their adaptability to the precision of the prediction. Specifically, our algorithms can self-adjust their sample complexity based on the accuracy of the available prediction, operating without any prior knowledge of the estimation’s accuracy (i.e. they are consistent). Additionally, we never use more samples than the standard approaches require, even if the predictions provide no meaningful information (i.e. they are also robust). We provide lower bounds to indicate that the improvements in sample complexity achieved by our algorithms are information-theoretically optimal. Furthermore, experimental results show that the performance of our algorithms on real data significantly exceeds our worst-case guarantees for sample complexity, demonstrating the practicality of our approach. 

This paper revisits the study of two classical technical tools in theoretical computer science: Yao's trans-formation of distinguishers to next-bit predictors (FOCS 1982), and the “reconstruction paradigm” in pseudorandomness (e.g., as in Nisan and Wigderson, JCSS 1994). Recent works of Pyne, Raz, and Zhan (FOCS 2023) and Doron, Pyne, and Tell (STOC 2024) showed that both of these tools can be derandomized in the specific context of read-once branching programs (ROBPs), but left open the question of de randomizing them in more general settings. Our main contributions give appealing evidence that derandomization of the two tools is possible in general settings, show surprisingly strong consequences of such derandomization, and reveal several new settings where such derandomization is unconditionally possible for algorithms stronger than ROBPs (with useful consequences). Specifically: •We show that derandomizing these tools is equivalent to general derandomization. Specifically, we show that derandomizing distinguish - to- predict transformations is equivalent to prBPP=prP, and that derandomized reconstruction procedures (in a more general sense that we introduce) is equivalent to prBPP=prZPP. These statements hold even when scaled down to weak circuit classes and to algorithms that run in super-polynomial time. •Our main technical contributions are unconditional constructions of derandomized versions of Yao's transformation (or reductions of this task to other problems) for classes and for algorithms beyond ROBPs. Consequently, we deduce new results: A significant relaxation of the hypotheses required to derandomize the isolation lemma for logspace algorithms and deduce that NL=UL; and proofs that de-randomization necessitates targeted PRGs in catalytic logspace (unconditionally) and in logspace (conditionally). In addition, we introduce a natural subclass of prZPP that has been implicitly studied in recent works (Korten FOCS 2021, CCC 2022): The class of problems reducible to a problem called “Lossy Code”. We provide a structural characterization for this class in terms of derandomized reconstruction procedures, and show that this characterization is robust to several natural variations. Lastly, we present alternative proofs for classical results in the theory of pseudorandomness (such as two-sided derandomization reducing to one-sided), relying on the notion of deterministically transforming distinguishers to predictors as the main technical tool. 

We consider the question of orienting the edges in a graph G such that every vertex has bounded out-degree. For graphs of arboricity α, there is an orientation in which every vertex has out-degree at most α and, moreover, the best possible maximum out-degree of an orientation is at least α - 1. We are thus interested in algorithms that can achieve a maximum out-degree of close to α. A widely studied approach for this problem in the distributed algorithms setting is a "peeling algorithm" that provides an orientation with maximum out-degree α(2+ε) in a logarithmic number of iterations. We consider this problem in the local computation algorithm (LCA) model, which quickly answers queries of the form "What is the orientation of edge (u,v)?" by probing the input graph. When the peeling algorithm is executed in the LCA setting by applying standard techniques, e.g., the Parnas-Ron paradigm, it requires Ω(n) probes per query on an n-vertex graph. In the case where G has unbounded degree, we show that any LCA that orients its edges to yield maximum out-degree r must use Ω(√ n/r) probes to G per query in the worst case, even if G is known to be a forest (that is, α = 1). We also show several algorithms with sublinear probe complexity when G has unbounded degree. When G is a tree such that the maximum degree Δ of G is bounded, we demonstrate an algorithm that uses Δ n^{1-log_Δ r + o(1)} probes to G per query. To obtain this result, we develop an edge-coloring approach that ultimately yields a graph-shattering-like result. We also use this shattering-like approach to demonstrate an LCA which 4-colors any tree using sublinear probes per query. 

Interval scheduling is a basic problem in the theory of algorithms and a classical task in combinatorial optimization. We develop a set of techniques for partitioning and grouping jobs based on their starting and ending times, that enable us to view an instance of interval scheduling on many jobs as a union of multiple interval scheduling instances, each containing only a few jobs. Instantiating these techniques in dynamic and local settings of computation leads to several new results. For (1+ε)-approximation of job scheduling of n jobs on a single machine, we develop a fully dynamic algorithm with O(lognε) update and O(logn) query worst-case time. Further, we design a local computation algorithm that uses only O(logNε) queries when all jobs are length at least 1 and have starting/ending times within [0,N]. Our techniques are also applicable in a setting where jobs have rewards/weights. For this case we design a fully dynamic deterministic algorithm whose worst-case update and query time are poly(logn,1ε). Equivalently, this is the first algorithm that maintains a (1+ε)-approximation of the maximum independent set of a collection of weighted intervals in poly(logn,1ε) time updates/queries. This is an exponential improvement in 1/ε over the running time of a randomized algorithm of Henzinger, Neumann, and Wiese ~[SoCG, 2020], while also removing all dependence on the values of the jobs' starting/ending times and rewards, as well as removing the need for any randomness. We also extend our approaches for interval scheduling on a single machine to examine the setting with M machines.